Expert Insights
PERSPECTIVE ON A TIMELY POLICY ISSUE
September 2025
JIM MIGNANO, JONATHAN W. WELBURN
Artificial Intelligence
and Crypto in
Financial Services
Policy Primer
This publication has completed RAND’s research quality-assurance
process but was not professionally copyedited.
For more information on this publication, visit
About RAND
RAND is a research organization that develops solutions to public policy challenges to help make communities throughout the world
safer and more secure, healthier and more prosperous. RAND is nonprofit, nonpartisan, and committed to the public interest. To learn
more about RAND, visit .
Research Integrity
Our mission to help improve policy and decisionmaking through research and analysis is enabled through our core values of quality
and objectivity and our unwavering commitment to the highest level of integrity and ethical behavior. To help ensure our research
and analysis are rigorous, objective, and nonpartisan, we subject our research publications to a robust and exacting quality-assurance
process; avoid both the appearance and reality of financial and other conflicts of interest through staff training, project screening,
and a policy of mandatory disclosure; and pursue transparency in our research engagements through our commitment to the open
publication of our research findings and recommendations, disclosure of the source of funding of published research, and policies to
ensure intellectual independence. For more information, visit
RAND’s publications do not necessarily reflect the opinions of its research clients and sponsors.
Published by the RAND Corporation, Santa Monica, Calif.
© 2025 RAND Corporation
is a registered trademark.
Limited Print and Electronic Distribution Rights
This publication and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is
provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to its webpage
on is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research products for
commercial purposes. For information on reprint and reuse permissions, visit
PE-A3888-2
iii
About This Paper
As the world, and the United States in particular, confronts the challenges of economic
mobility, health and longevity, and new frontiers of technology and human well-being, have we
designed our social and economic systems to effectively address these convergences? RAND is
addressing these questions through an effort called Social and Economic Policy Rethink to
transform the approach to solving social and economic policy problems. The broader volume
encompasses seven reports that describe key social and economic stakes for policy on AI
adoption, examine how policymakers can account for rapid AI capabilities development and
adoption in and across sectors, and showcase how industry and policy leaders can responsibly
meet the challenges of an era of transformative AI. This report speaks to how different sectors
should respond to AI adoption, by highlighting the diverse landscape of cryptocurrencies, the
integration of AI in digital assets, the potential for disruption to financial services, and the need
for adaptive regulatory frameworks to manage systemic risks.
RAND Social and Economic Well-Being
RAND Social and Economic Well-Being is a division of RAND that seeks to actively
improve the health and social and economic well-being of populations and communities
throughout the world. This research was conducted in the Social and Behavioral Policy Program
within RAND Social and Economic Well-Being. The program focuses on such topics as risk
factors and prevention programs, social safety net programs and other social supports, poverty,
aging, disability, child and youth health and well-being, and quality of life, as well as other
policy concerns that are influenced by social and behavioral actions and systems that affect well-
being. For more information, email sbp@.
Funding
Funding for this research was provided by the contributions of the RAND Social and
Economic Policy Advisory Board; generous gifts by Frank M. Clark, Michael J. Critelli, William
A. Downe, Jihee Kim Huh and Peter Yun Huh, and the Donald M. James Family Foundation;
and income from the operation of RAND Social and Economic Well-Being.
Acknowledgments
We thank the entire Social and Economic Policy Rethink team for their insights throughout
this project. The paper’s peer reviewers, Drew Propson and Salil Gunashekar, provided
thoughtful and insightful comments on an earlier draft, which led to substantial improvements.
mailto:sbp@
iv
We extend our appreciation to the leadership of social and economic policy at RAND, Anita
Chandra, Peter Hussey and Heather Schwartz; to the RAND Social and Economic Policy board
for their support and advice; to the members of the Rethink financial services, health care, and
climate and energy working groups for their advice and scoping; to Lisa Coe for her careful
edits; and to the members of the Rethink team whose input helped guide this analysis
v
Summary
As the world, and the United States in particular, confronts the challenges of economic
mobility, health and longevity, and new frontiers of technology and human well-being, have we
designed our social and economic systems to effectively address these convergences? RAND is
addressing these questions through an effort across RAND’s social and economic policy research
divisions – Education and Labor, Health Care, and Social and Economic Wellbeing, which we
call the Social and Economic Policy ReThink. This perspective contributes to the first volume of
the Social and Economic Policy ReThink on the Adoption of Artificial Intelligence (AI) by
highlighting the diverse landscape of cryptocurrencies, the integration of AI in digital assets, the
potential for disruption to financial services, and the need for adaptive regulatory frameworks to
manage systemic risks.
Key Takeaways
The following are some key takeaways from the research:
• Digital assets are more than just Bitcoin, encompassing a diverse set of instruments—
including cryptocurrencies, stablecoins, utility tokens, tokenized assets, and non-fungible
tokens—that facilitate peer-to-peer exchanges without intermediaries and interact in
increasingly complex ways with traditional finance.
• Central bank digital currencies represent a separate category of publicly issued digital
assets. While few are operational, many countries are actively exploring their design and
implications for monetary policy, financial access, and system resilience.
• AI integration with digital assets is nascent, with deployments in proof-of-concept or
early development stages. Nonetheless, efforts are underway to apply AI to trading,
payments, lending, insurance, and compliance, especially where digital asset
infrastructure allows for automation and continuous operation.
• Autonomous agents equipped with cryptocurrency may challenge regulatory
assumptions, including identity-linked oversight frameworks. Their potential to act
independently in financial markets raises novel questions about accountability,
enforcement, and systemic impact.
• AI-cryptocurrency convergence could disrupt existing financial models, particularly
through disintermediation. Combining programmability, real-time data processing, and
distributed infrastructure, these systems may compete with banks, payment processors,
and other financial intermediaries.
• Systemic risks in this space may mirror those of traditional finance but with added
velocity and opacity, given the global, always-on nature of digital assets and the opacity
of some AI processes. Risk propagation could occur faster, with limited recourse or
visibility.
vi
• Policy responses will need to balance enabling experimentation while managing risk.
Early intervention will be key to shaping innovation trajectories without stifling
development.
vii
Contents
About This Paper ........................................................................................................................... iii
Summary .......................................................................................................................................... v
Contents ........................................................................................................................................ vii
Figure and Table .......................................................................................................................... viii
AI and Crypto in Financial Services: Policy Primer ....................................................................... 1
Technologies Underlying Crypto .............................................................................................................. 2
Taxonomy of Digital Assets ...................................................................................................................... 3
The Digital Asset Ecosystem .................................................................................................................... 4
AI and Crypto in Financial Services ......................................................................................................... 5
Crypto Deployments in AI ...................................................................................................................... 13
Analysis and Future Directions ............................................................................................................... 15
Policy Considerations .............................................................................................................................. 16
Glossary ......................................................................................................................................... 18
Abbreviations ................................................................................................................................ 22
References ..................................................................................................................................... 23
viii
Figure and Table
Figure
Figure 1. The Digital Asset Ecosystem ........................................................................................... 5
Table
Table 1. Taxonomy of Digital Assets .............................................................................................. 3
1
AI and Crypto in Financial Services: Policy Primer
Cryptocurrency (or “crypto”) is more than just Bitcoin. Crypto refers to private sector digital
assets that use cryptography and distributed ledger technology to process transactions without
needing The digital asset ecosystem that has emerged around crypto enables
peer-to-peer exchange over the Internet as well as novel ways of providing financial services.
While Bitcoin is a well-recognized and dominant cryptocurrency, there are thousands of
others. These serve a variety of purposes, such as facilitating faster transactions, supporting
distributed computer applications, enhancing user privacy, and enabling digital collectibles.
However, crypto has also facilitated a number of illicit activities including money laundering,
cybercrime, ransomware, narcotics trafficking, theft and fraud, human trafficking, terrorism and
proliferation Crypto can also refer to stablecoins, utility tokens, and non-fungible
tokens (NFTs).3
Central bank digital currencies (CBDCs) are another type of digital asset, distinct from
crypto. CBDCs are issued and regulated by central banks, serving as digital versions of national
currency. While few CBDCs have been deployed at scale, most central banks are actively
exploring their potential
By reducing reliance on banks and other financial institutions—a process known as
disintermediation—digital assets could lower costs for consumers, businesses, and governments
while expanding financial access to those underserved by traditional financial institutions. To
paraphrase advocates, digital assets could make payments as fast, easy, and inexpensive as
sending a text message to anyone in the world.
Skeptics suggest disintermediation presents risks. Traditional financial institutions provide
for fraud detection, dispute resolution, and risk management across markets and time horizons.
They also serve as key implementation points for contemporary financial regulations such as
anti-money laundering compliance, consumer protection, and crisis intervention. Removing
intermediaries could leave users to bear these risks while complicating regulatory
1 Financial Stability Oversight Council, Report on Digital Asset Financial Stability Risks and Regulation, 2022.
While “crypto” is commonly used to refer to cryptocurrency, in other domains it refers to cryptography. This primer
uses the term consistent with its common usage in finance, while recognizing this may not be universal.
2 . Department of Justice, The Role of Law Enforcement in Detecting, Investigating, and Prosecuting Criminal
Activity Related to Digital Assets, September 2022.
3 “Taxonomy of Digital Assets” (below) further discusses stablecoins, utility tokens, and NFTs, as well as CBDCs.
4 Atlantic Council, “Central Bank Digital Currency Tracker,” webpage, February 2025.
5 The term disintermediation can be somewhat misleading in the context of digital assets. Real-world crypto
implementations have largely introduced new intermediaries (see The Digital Asset Ecosystem below).
2
Experiments combining digital assets and artificial intelligence (AI) are proliferating. By
combining crypto’s distributed approach with AI’s rapid data analysis, the convergence could
transform financial services. AI’s potential to enhance efficiency and accessibility may be the
key to unlocking crypto’s full disruptive potential.
This primer focuses on the emerging intersection of AI and crypto in financial services as
presented by innovators. It aims to provide policymakers with a concise overview of potential
applications, emerging risks, and areas for regulatory attention. The primer is organized by
technology domain, beginning with foundational concepts before exploring specific use cases
across payments, trading, credit, insurance, and Some AI and crypto applications
and projects will fail, others might succeed. However, if AI and crypto become more intertwined,
policymakers will need to understand and prepare for potential impacts on consumers,
businesses, and governments.
Technologies Underlying Crypto
Efforts to create fast, inexpensive, and secure peer-to-peer electronic money date back to at
least the 1980s. These efforts were stymied by a so-called “double-spend problem,” where a
single unit of electronic money could be easily duplicated and spent more than
Bitcoin solved this issue by combining public key cryptography and distributed ledger
technology (DLT) to become the world’s first cryptocurrency in 2009. Public key cryptography
enables secure transactions through a pair of digital keys: a public key, which acts as an address
for receiving funds, and a private key, which is required to authorize spending. Cryptographic
hash functions ensure each transaction is securely validated and recorded in the crypto
DLT is a networked data system where records are shared and synchronized across many nodes,
making it nearly impossible to alter past transactions.
This dual-technology approach enables Bitcoin to operate as a digital token, representing
value that can be directly exchanged, and as an entry in a shared ledger that prevents double-
spending. By implementing digital tokens this way, Bitcoin established a system for transferring
value without intermediaries, laying the foundation for the broader digital asset ecosystem.
6 A glossary of technical and policy-relevant terms is also provided at the end to support clarity and consistency.
7 David Chaum, “Blind Signatures for Untraceable Payments,” Advances in Cryptology Proceedings of Crypto 82,
1983; David Andolfatto and Fernando M. Martin, “The Blockchain Revolution: Decoding Digital Currencies,”
Federal Reserve Bank of St. Louis Annual Report 2021, April 20, 2022. Chaum’s proposed cryptographic solution to
transaction anonymity laid early groundwork for peer-to-peer electronic money systems, including cryptocurrency.
Andolfatto and Martin explain how the double-spend problem impeded peer-to-peer electronic money prior to
blockchain solutions.
8 While quantum computing may affect the security of cryptographic hashes in the future, reversing them to expose
the original input is currently computationally infeasible. Cryptocurrency employs cryptographic hash functions to
securely link blocks of transactions together and to present complex puzzles that miners or validators must solve to
add new blocks to the system.
本报告来源于三个皮匠报告站（）,由用户Id:1079574下载,文档Id:908921,下载日期:2025-09-11
3
Bitcoin pioneered the implementation of DLT through blockchain—a secure, time-stamped
chain of data blocks that records every transaction. Each data block is cryptographically linked to
the previous one, creating an immutable transaction ledger. This approach prevents tampering,
and because the blockchain can be downloaded, provides a transparent record of
Blockchain protocols define the rules for transferring, validating, and storing data within the
network. For example, consensus mechanisms such as Proof of Work and Proof of Stake ensure
that participants agree on the ledger’s state without the need for a central authority.
Beyond these foundational elements, several innovations have further expanded digital asset
capabilities. Programmability—embedding code into digital assets through smart contracts—
allows for automated execution of pre-defined rules, supporting distributed applications and
services that can operate without human intervention. Tokenization is a method to represent real-
world or financial assets (., real estate, bonds, commodities) as digital tokens, enabling
fractional ownership and facilitating transferability.
Taxonomy of Digital Assets
Since Bitcoin’s launch in 2009, the universe of crypto and other digital assets has grown to
include a wide range of asset types. Table 1 categorizes key digital asset types, distinguishing
between private-sector cryptocurrencies, government-issued digital currencies, and other
blockchain-based assets.
Table 1. Taxonomy of Digital Assets
Digital Asset
Type
Description Examples
Cryptocurrencies Privately or non-government issued, blockchain-based
digital assets that use cryptography for transactions
without the need for an intermediary
Bitcoin, Litecoin, Monero, Dogecoin
Stablecoins Privately issued, DLT-based digital assets that attempt to
maintain value by pegging to other assets
USDC (“Circle”), USDT (“Tether”)
CBDCs Publicly issued digital assets regulated and backed by
central banks; includes wholesale CBDCs, restricted to
financial institutions for interbank settlement and
monetary policy operations, and retail CBDCs, available
to the general public
Digital Yuan / e-CNY (Chinese pilot),
digital euro (Euro area research and
experimentation), eNaira (Nigeria,
launched in 2021)
Utility Tokens Blockchain-based tokens designed to provide access to
a specific service or application
ETH (Ethereum gas fees), BNB
(Binance platform fees)
9 Another feature of many blockchain implementations is the application of open-source software development,
which supports community-driven adaptation and improvement of blockchain software. Key aspects include: open-
source code repositories (., GitHub) host crypto projects such as Bitcoin and Ethereum, allowing anyone to view
and download project software, run network nodes to participate in transaction validation, and propose code
changes; while anyone can suggest changes, network participants (., miners, validators, developers) must agree on
code changes through consensus governance mechanisms; if disagreements persist, developers can create a hard fork
(a permanent split resulting in a new blockchain) or a soft fork (a backward-compatible upgrade).
4
Tokenized
Assets
DLT-based representations of ownership of real-world or
financial assets (., real estate, stocks, commodities)
designed to enable fractional ownership and digital trade
Tokenized bonds, real estate tokens
NFTs Blockchain-based representations of ownership of digital
or physical items
CryptoPunks (pixel art characters),
Bored Ape Yacht Club (cartoon apes)
SOURCE: CBDC examples feature information from Atlantic Council, “Central Bank Digital Currency Tracker,”
webpage, February 2025; and European Central Bank, “Progress on the Preparation Phase of a Digital Euro –
Second Progress Report,” webpage, December 2, 2024.
NOTE: See the glossary at the end of this primer for sources defining constituent elements of this taxonomy.
This classification is important for understanding each asset’s function and regulatory
implications. Cryptocurrencies such as Bitcoin operate without central control, while stablecoins
aim to bridge the gap between crypto and traditional finance by maintaining price stability.
CBDCs are state-backed and designed to facilitate public policy.
Other asset types, such as utility tokens and NFTs, represent access rights or ownership
rather than functioning as currencies. Tokenized assets allow real-world and financial assets to
be traded in blockchain ecosystems, potentially increasing liquidity and market efficiency.
AI-driven systems might interact differently with various asset types, for example,
optimizing algorithmic trading for cryptocurrencies, managing risk in stablecoin-backed
financial products, or automating compliance for tokenized As AI integrates with
crypto, policymakers will need to distinguish between these asset categories to effectively assess
opportunities, risks, and regulatory needs.
The Digital Asset Ecosystem
The digital asset ecosystem is a dynamic, multi-layered network that supports a range of
digital assets and services outside traditional finance. It consists of multiple underlying
technologies and asset types that interact with each other and the broader financial system.
The taxonomy outlined earlier—cryptocurrencies, stablecoins, CBDCs, utility tokens,
tokenized assets, and NFTs—represents the different kinds of digital assets in this space.
The ecosystem interacts with the larger financial system in various ways. For example,
crypto exchanges enable users to trade into, between, and out of digital assets. Banks and
institutional investors are increasingly participating in the digital asset space, exploring
blockchain for settlement and clearance, or investing in crypto-based products.
Major actors in the ecosystem include:
10 Financial Stability Board, Artificial Intelligence and Machine Learning in Financial Services: Market
Developments and Financial Stability Implications, November 1, 2017; Financial Stability Board, The Financial
Stability Implications of Artificial Intelligence, November 14, 2024. The Financial Stability Board provides an
overview of AI applications across the financial sector, not limited to digital assets. These AI-driven innovations are
being widely explored and developed in traditional banking and finance, as well as in the digital assets space.
5
• Developers and open-source communities build and maintain the software and
protocols that run crypto networks.
• Miners and validators secure the networks and validate transactions.
• Exchanges and trading platforms facilitate the exchange of digital assets.
• Wallet providers enable users to store and manage their crypto securely.
• Institutional investors and banks engage with crypto for investment and technological
innovation.
• Payments companies that enable crypto integration with existing payment networks.
• Financial market infrastructures exploring or supporting blockchain-based clearing
and settlement systems.
• Regulators and policymakers create and enforce rules to protect consumers and
maintain market stability.
• Consumers and merchants use crypto for payments, remittances, and investment.
Figure 1 is designed to clarify how each component interacts within the digital asset
ecosystem and connects to the broader financial landscape.
Figure 1. The Digital Asset Ecosystem
AI and Crypto in Financial Services
AI deployments in crypto could accelerate innovation in financial AI-driven
solutions are being explored to optimize crypto payments, automate trading strategies, and
improve financial accessibility through blockchain-based credit and lending. AI could also
enhance tokenized asset markets, smart contracts, and blockchain-based insurance models, while
11 This primer uses the term AI to refer to systems that can perform tasks typically requiring human judgment. The
terms AI-driven, AI-powered, and AI-enabled are used interchangeably to identify applications of AI in crypto, often
involving machine learning unless otherwise specified.
Digital Asset
Infrastructure
Digital Asset Types
Service Platforms
Interface with
Traditional Finance
Underlying technologies (., public key
cryptography, DLT) and secondary
technologies (., programmability,
tokenization)
Crypto, stablecoins, CBDCs,
utility tokens, NFTs, tokenized assets
Crypto exchanges, wallet providers,
smart contract platforms
Banks, institutional investors,
regulatory bodies, payments companies,
financial market infrastructures
Consumers and merchants connect layers
Developers, open-source communities,
miners, and validators build, maintain, and
operate infrastructure
Users and Participants Layers Key Components
6
playing a role in regulatory technology by strengthening fraud detection and compliance
measures.
This section focuses on specific applications and projects at the intersection of AI and crypto,
rather than broader systemic, geopolitical, or environmental implications. These applications and
projects can have broad implications due to the global, 24/7 nature of digital assets. However,
matters such as cross-border regulatory coordination, jurisdictional spillovers, and externalities
including energy consumption are not addressed here and remain important areas for further
study.
Many of the applications and projects discussed build on existing uses of AI in financial
services, such as algorithmic trading, credit scoring, and fraud detection. The examples here
illustrate how such capabilities may evolve or expand through integration with digital asset
infrastructures.
A key limitation in this space is the lack of empirical evidence and real-world deployments to
support claims about AI’s implementation in crypto markets. Most applications and projects
remain in proof-of-concept or experimental stages, with their effectiveness unproven at scale.
Given the scarcity of peer-reviewed literature associated with the novelty of the AI-crypto
intersection, this primer draws on a rapid scan of available sources—including project websites
and blog-format content—while prioritizing authoritative industry and government publications
where
AI Deployments in Crypto Payments
Integrating AI into crypto payment systems presents aspirational goals to reduce transaction
frictions. Doing so could facilitate high-volume transactions and enhance accessibility for small-
dollar payments, particularly for unbanked consumers, as illustrated in the examples and use
cases below. However, these efforts face technical, regulatory, and adoption hurdles.
While crypto payments have shown considerable achievement in driving down transaction
fees, they still struggle to compete with traditional payment systems in terms of speed and
This is largely due to how blockchain consensus mechanisms require multiple
confirmations across a distributed network before finalizing For example, despite
12 This primer does not aim to endorse or validate specific applications, projects, techniques, or organizations.
Examples are included solely to illustrate emerging trends of possible policy relevance and were selected following
critical evaluation of publicly available documentation and relevance. Numerous other projects were excluded due to
insufficiently documented claims, lack of transparency, or otherwise poor evidentiary grounding.
13 Bank for International Settlements, Annual Economic Report 2022, June 2022; Bank for International
Settlements, The Crypto Ecosystem: Key Elements and Risks, July 2023.
14 Bank for International Settlements, July 2023.
7
being one of the fastest blockchains, Solana can execute over 2,000 transactions per second
(TPS), compared to Visa’s 65,000
AI could improve crypto payment efficiency, potentially surpassing traditional payment
systems. One application is AI-driven load balancing, which dynamically distributes transaction
verification across blockchain nodes to prevent bottlenecks and optimize network performance.
Rerouting traffic in this manner could help minimize latency and improve settlement
Additionally, machine learning models can be used to predict blockchain service fees and
congestion patterns, reducing
The integration of AI and crypto could be advantageous for very high-volume transactions,
such as those that occur within large companies, federal government operations, or healthcare
payment These sectors require systems that can handle vast numbers of transactions
without slowing down. Stablecoin networks might employ AI-driven load balancing and
predictive models for service fees and congestion to manage such heavy loads more efficiently.
While potential cost savings and processing improvements are uncertain, some stablecoin
transaction costs are already very competitive. For example, average transaction fees for
USDC—a leading stablecoin—range between and percent without the use of
Crypto payments could also benefit those whom traditional finance has underserved. Many
small businesses and unbanked individuals conduct transactions that are too small for
conventional banks to pursue or involve bank fees that are too costly to justify opening an
Ripple’s XRP cryptocurrency, for instance, facilitates peer-to-peer payments in under
five seconds, including cross-border remittances, with an average transaction fee of less than
$ Integrating AI into these systems might drive friction and transaction costs even lower.
15 Mustafa Bedawala and Arjuna Wijeyekoon, “A Deep Dive on Solana, a High Performance Blockchain Network,”
Visa Crypto Thought Leadership blog, undated.
16 European Telecommunications Standards Institute (ETSI), Permissioned Distributed Ledger (PDL); Artificial
Intelligence for Permissioned Distributed Ledger, ETSI GR PDL 032, , January 2025, p. 41.
17 Conall Butler and Martin Crane, “Blockchain Transaction Fee Forecasting: A Comparison of Machine Learning
Methods,” Mathematics, Vol. 11, No. 9, May 2023; Jiangqin Ma and Erfan Mahmoudinia, “Comprehensive
Modeling Approaches for Forecasting Bitcoin Transaction Fees: A Comparative Study,” arXiv,
arXiv:, February 2025.
18 Iñaki Aldasoro, Leonardo Gambacorta, Anton Korinek, Vatsala Shreeti, and Merlin Stein, “Intelligent Financial
System: How AI is Transforming Finance,” BIS Working Papers, No. 1194, June 2024. Aldasoro et al. detail how
AI applications could improve liquidity management and accelerate payment flows, which are important for
supporting high-volume transactions at scale, but are not exclusive to crypto.
19 Vuk Martin, “What Fees Are Cheaper: USDT or USDC?” Coincodex, webpage, January 6, 2025.
20 Asli Demirguc-Kunt, Klapper,Leora, Singer,Dorothe, and Ansar,Saniya, The Global Findex Database 2021:
Financial Inclusion, Digital Payments, and Resilience in the Age of COVID-19, World Bank Group, July 2022.
21 Caleb & Brown, “What is XRP? A Beginner’s Guide,” Caleb & Brown blog, January 21, 2025.
8
For example, Félix Pago uses natural language processing and stablecoins to enable fast, low-
cost cross-border remittances via
AI Deployments in Crypto Trading
As discussed in this section, crypto trading platforms are exploring the use of machine
learning and deep learning algorithms to analyze real-time market data, optimize strategies, and
execute trades autonomously. These efforts build on traditional algorithmic trading but introduce
greater adaptability and continuous learning, while exploiting crypto’s data-rich digital
infrastructure.
AI-Driven Trading Strategies
Unlike traditional algorithmic trading, which follows preset rules such as automatic portfolio
rebalancing, AI-driven systems can process continuous data streams and react to sudden shifts in
market This makes them well-suited to the 24/7 operations of digital asset markets.
While not exclusive to AI, Visa estimates that trading bots account for about $2 trillion, or 90
percent of monthly stablecoin transactions, highlighting the potential for deeper
Deploying AI could help traders respond in real time to sudden changes in sentiment,
liquidity fluctuations, and emerging arbitrage opportunities across different platforms. For
example, Amazon Managed Blockchain (AMB) Query APIs provide AI trading algorithms
access to real-time blockchain data, which can be combined with off-chain data to adjust trading
strategies and execute with AMB Crypto platforms are also beginning to broaden
access to algorithmic trading tools, which could make sophisticated crypto trading more widely
AI in Tokenized Asset Markets
AI could also play a role in digital asset tokenization, where real-world or financial assets
such as real estate, securities, and fine art are represented as blockchain-based
Tokenization is designed to improve liquidity by enabling fractional ownership and around-the-
clock trading.
22 CrossTech, “Félix,” webpage, undated.
23 Financial Stability Board, November 1, 2017
24 Cuy Sheffield, “Making Sense of Stablecoins,” Visa Perspectives blog, April 25, 2024.
25 Forrest Colyer, John Liu, and Michael Greenwald, “The Convergence of AI and Digital Assets: A New Dawn for
Financial Infrastructure,” AWS Web3 Blog, October 19, 2023.
26 Cryptohopper, “Algorithm Intelligence (.),” webpage, undated.
27 Iñaki Aldasoro, Sebastian Doerr, Leonardo Gambacorta, Rodney Garratt, and Priscilla Koo Wilkens, “The
Tokenisation Continuum,” BIS Bulletin, No. 72, April 11, 2023.
9
Incorporating AI into asset tokenization might enhance the efficiency and transparency of
financial markets. AI could automate and streamline the tokenization process by analyzing large
datasets to accurately assess asset values. It could also monitor and manage risks associated with
tokenized assets in real-time. AI’s ability to identify patterns and trends might improve
decisionmaking for investors and financial institutions, while optimizing investment strategies
based on individual risk tolerance and market conditions. Additionally, AI could enhance market
transparency and ensure best price execution by aggregating bid/ask quotes across global
tokenization
AI in Smart Contracts
Smart contracts execute predetermined actions when specific conditions are met, but they
lack the ability to adjust to changing circumstances. Integrating AI could enable smart contracts
to continuously self-adapt based on real-time For example, SingularityDAO’s Dynaset
combines “human intervention, algorithmic models, machine learning models, and artificial
intelligence infrastructures” to monitor and rebalance retail crypto
In blockchain-based finance, AI-powered smart contracts could help optimize liquidity
allocation by reacting to market conditions in real AI is also being explored to automate
smart contract arbitration by analyzing transaction records and digital
AI Deployments for Credit and Lending
Traditional credit scoring relies on centralized bureaus and historical financial data, which
can overlook individuals with limited-to-no conventional credit history. By contrast, blockchain-
based lending platforms can build real-time credit profiles combining conventional data with
ledger data such as transaction histories and wallet AI could help generate more
dynamic and comprehensive risk assessments by analyzing this data,34 while refining risk models
by incorporating real-time indicators, including transaction patterns, network activity, and even
external data such as social media sentiment or Internet of Things (IoT)-based financial behavior.
28 Merav Ozair, “The Future of Finance: AI Meets Tokenization,” Nasdaq, August 16, 2023.
29 Todd Kanaster, Andrew O’Neill, Rebecca Mun, Josh Stokesberry, Matta Uma Maheswara Reddy, and Ava Yang,
“Crypto and AI: Shaping the Future of the Internet,” S&P Global, October 1, 2024.
30 SingularityDAO, “DYNASET Terms & Conditions,” webpage, undated.
31 aelf, “AI-Powered Smart Contracts: The Impact of AI on Web3 Crypto Transactions,” aelf Blog, January 7, 2025.
32 Federico Ast, Jamilya Kamalova, and Yann Aouidef, “The Kleros Fellowship of Justice, 8th Generation:
Applications Open!” Kleros Blog, September 11, 2024.
33 Nizan Geslevich Packin and Yafit Lev-Aretz, “Crypto-Native Credit Score: Between Financial Inclusion and
Predatory Lending,” Cardozo Law Review, Vol. 45, No. 3, February 2024, pp. 845–898.
34 Zorka Jovanovic, Zhe Hou, Kamanashis Biswas, and Vallipuram Muthukkumarasamy, “Robust Integration of
Blockchain and Explainable Federated Learning for Automated Credit Scoring,” Computer Networks, Vol. 243,
April 2024.
10
Integrating AI with smart contracts could enable real-time loan adjustments based on market
conditions and borrower For example, AI-powered smart contracts might
dynamically modify interest rates based on credit risk and liquidity conditions, update collateral
requirements if asset values fluctuate, and trigger automated repayments or refinancing to
prevent
While these credit and lending innovations might improve efficiency and broaden financial
access, algorithmic bias remains a concern. High-profile failures—such as the Dutch tax
authority’s misuse of algorithms to identify suspected benefits fraud and persistent errors in
Michigan’s unemployment fraud detection system—illustrate how flawed automated systems
can produce systemic harm, particularly for vulnerable AI-driven credit and
lending systems may reinforce inequalities if they learn from biased datasets or fail to consider
broader socioeconomic
AI Deployments in Crypto for Insurance
AI deployments in crypto could have insurance applications through automated risk
assessments and claims processing. Similar to blockchain-based credit scoring, blockchain-based
insurance systems provide a wealth of risk assessment data that traditional approaches might
overlook, and that AI integrations could
AI might power blockchain-based underwriting by analyzing extensive datasets, including
historical claims, blockchain transactions, and real-time external data. Machine learning could be
used in blockchain-based insurance platforms to identify subtle risk patterns and to tailor policies
to individual risk profiles, allowing for more accurate
35 TechBullion, “Smart Contracts: Simplifying DeFi Loan Repayment Processes,” webpage, January 26, 2025.
36 See Akash Takyar, “AI in Loan Underwriting: Use Cases, Architecture, Technologies, Solution and
Implementation,” webpage, undated.
37 Melissa Heikkilä, “Dutch Scandal Serves as a Warning for Europe over Risks of Using Algorithms,” Politico
Europe, March 29, 2022; Jennifer Lord, “The Seven-Year Struggle to Hold an Out-of-Control Algorithm to
Account,” interview with Julia Angwin, Hello World, October 8, 2022.
38 Johnjerica Hodge, India Williams, Nicholas Gervasi, and Gabriella Weick, “Artificial Intelligence in Consumer
Lending: Addressing AI-Related Risks,” The Banking Law Journal, Vol. 141, No. 10, November–December 2024,
pp. 475–484.
39 Sukriti Bhattacharya, German Castignani, Leandro Masello, and Barry Sheehan, “AI Revolution in Insurance:
Bridging Research and Reality,” Frontiers in Artificial Intelligence, Vol. 8, April 2025.
40 See Kofi Immanuel Jones and Swati Sah, “The Implementation of Machine Learning in The Insurance Industry
with Big Data Analytics,” International Journal of Data Informatics and Intelligent Computing, Vol. 2, No. 2, June
2023, pp. 21–38.
11
AI-powered smart contracts might streamline claims verification and payouts, reducing
human intervention and accelerating resolution Such contracts could autonomously
assess claims based on objective data inputs, such as blockchain-verified
AI Deployments in Crypto RegTech
The crypto space has long been a testbed for regulatory technology (RegTech), a domain of
technological approaches to streamline the complexity and costs of compliance through
automation. In financial services, compliance processes such as Know Your Customer (KYC)
and anti-money laundering (AML) procedures can be time-consuming and labor intensive. AI-
powered RegTech might automate many of these functions while improving the detection of
illicit digital asset activities. Incorporating machine learning, biometric authentication, and
blockchain-based identity solutions could strengthen regulatory oversight and enhance security,
with early deployments in areas such as fraud detection, KYC automation, and verifiable
credentials explored below. However, as AI tools advance, so do the tactics of malicious actors,
making the resilience of AI-driven compliance systems an ongoing challenge for regulators and
the industry.
Real-Time Fraud Detection and AML Compliance
Historically, financial fraud detection and AML compliance were some of the earliest
applications of machine AI supports fraud detection and AML compliance by
continuously analyzing blockchain transactions for suspicious activity. Machine learning models
process vast amounts of transaction data in real time, identifying patterns indicative of money
laundering, illicit financing, or fraudulent behavior. These tools refine their accuracy over time
by learning from historical and current data, enabling financial institutions and regulators to flag
high-risk transactions before they escalate.
Progress with AI-powered fraud detection is already For example, blockchain
analytics firm Chainalysis reports that its recently acquired AI-powered fraud detection solution
41 Khyati Kapadiya, Usha Patel, Rajesh Gupta, Mohammad Dahman Alshehri, Sudeep Tanwar, Gulshan Sharma,
and Pitshou N. Bokoro, “Blockchain and AI-Empowered Healthcare Insurance Fraud Detection: An Analysis,
Architecture, and Future Prospects,” IEEE Access, Vol. 10, 2022, pp. 79,606–79,627.
42 Araddhana Arvind Deshmukh, Prabhakar Kandukuri, Janga Vijaykumar, Anna Shalini, S. Farhad, Elangovan
Muniyandy, and Yousef A. Baker El-Ebiary, “Event-based Smart Contracts for Automated Claims Processing and
Payouts in Smart Insurance,” International Journal of Advanced Computer Science & Applications, Vol. 15, No. 4,
2024.
43 For example, research on using neural networks for financial fraud detection dates as early as 1997. See Jarrod
West and Maumita Bhattacharya, “Intelligent Financial Fraud Detection: A Comprehensive Review,” Computers &
Security, Vol. 57, 2016, p. 49.
44 . Department of the Treasury, “Treasury Announces Enhanced Fraud Detection Process Using AI Recovers
$375M in Fiscal Year 2023,” press release, February 28, 2024.
12
“has helped top cryptocurrency exchanges decrease fraud by 60 percent, reduce scam-related
activities, and improve efficiency of manual operations.”45
AI-Driven Identity Verification and KYC
AI could help streamline certain KYC procedures, making compliance more efficient while
improving user security. KYC procedures include risk-based customer identity verification and
ongoing assessments meeting federal Customer Identification Program (CIP) and Customer Due
Diligence (CDD) CIP involves verifying a customer’s identity documentation
and CDD involves assessing a customer’s risk profile, particularly for money laundering and
terrorist financing, based on identity and financial AI applications in KYC include
document processing and biometric authentication for CIP, as well as predictive analytics for
CDD. However, integrating AI with KYC processes can perpetuate biases, leading to risk
assessments that improperly deny financial services, while digital identity solutions can pose
cybersecurity risks and privacy concerns given the extensive volume of personal data
AI-powered biometric authentication methods such as fingerprint scanning, facial
recognition, and iris scans replace traditional passwords and PINs, reducing reliance on easily
compromised For example, World Network, a crypto project founded by OpenAI
CEO Sam Altman, integrates iris scanning to verify user identities and distributes WLD crypto
tokens to verified AI can rapidly analyze biometric data and identity documents
against global databases, expediting KYC checks and ensuring regulatory compliance.
Automating these processes might also reduce onboarding time for new users.
Despite these advantages, AI-driven KYC systems face challenges, including deepfake
identities and synthetic fraud Malicious actors can also use AI to create highly
realistic fake identities, requiring equally advanced detection mechanisms to maintain trust and
45 Jonathan Levin, “Welcoming Fraud Detection Innovator Alterya to Chainalysis and Doubling Down on the
Prevention of Illicit Activity,” Chainalysis Blog, January 13, 2025.
46 Other regulatory requirements, such as beneficial ownership, suspicious activity, and currency transaction
reporting are not direct components of KYC but are integral to the broader AML framework and rely on KYC
processes.
47 For a brief overview, see Thomson Reuters, “5 Essential Steps for KYC/AML Onboarding and Compliance,”
Thomson Reuters Legal blog, June 24, 2024. For a deeper explanation of the regulatory requirements, see Federal
Financial Institutions Examination Council, “Customer Identification Program,” in BSA/AML Manual, February
2021, and Federal Financial Institutions Examination Council, “Customer Due Diligence — Overview,” in
BSA/AML Manual, May 5, 2018.
48 Financial Action Task Force, Opportunities and Challenges of New Technologies for AML/CFT, July 2021, pp.
42–43.
49 Alex Vasilchenko, “AI Biometric Authentication for Enterprise Security,” MobiDev Blog, January 31, 2025.
50 World Network, “A New Identity and Financial Network,” webpage, undated.
51 Heather Chen and Kathleen Magramo, “Finance Worker Pays Out $25 Million After Video Call with Deepfake
‘Chief Financial Officer’,” CNN, February 4, 2024.
13
As AI authentication systems evolve, ongoing improvements in verification methods
will be critical to countering increasingly sophisticated fraud tactics, while also addressing
longstanding concerns associated with biometric data use, including risks related to privacy, data
breaches, discrimination, and surveillance.
AI-Enhanced Verifiable Credentials
Blockchain-based verifiable credentials allow users to prove their identities or qualifications
without exposing additional personal Verifiable credentials digitally represent the same
information as physical credentials (., identity details, issuing authority), but employ digital
signatures and zero-knowledge proofs to protect sensitive information that otherwise does not
need to be A common example used to illustrate the value of this approach is a
person seeking to enter a venue where they only need to prove they are over 21 years old. Unlike
using conventional identification, verifiable credential proves their age without revealing any
other personal information. These immutable records can enhance trust and transparency in
financial transactions while reducing identity theft risks. AI might be used in blockchain-based
credential systems to support real-time validation of digital identities, ensuring that credentials
remain up to date and resistant to
Crypto Deployments in AI
Thus far, this primer has focused on how AI might be used to improve crypto and related
blockchain implementations. The intersection of AI and crypto could also involve crypto and
blockchain deployments to support AI. This section considers two such possibilities: (1)
equipping emerging autonomous agents with crypto tokens so they can independently transact
with other parties, and (2) blockchain-based computing for AI development.
Crypto-Equipped Autonomous Agents
AI-powered autonomous agents are in their infancy. They are designed to operate
independently, making decisions, executing complex tasks, and adapting to real-time data
without continuous human intervention. Unlike traditional bots, these agents can analyze
52 Muhammad Shahid Hanif, “Synthetic Identities: The Darker Side of Generative AI,” Forbes, May 29, 2024.
53 Verifiable credentials do not require blockchain but can be integrated with it.
54 Manu Sporny, Dave Longley, David Chadwick, and Ivan Herman, “Verifiable Credentials Data Model ,”
W3C Candidate Recommendation Draft, February 25, 2025.
55 Md Jobair Hossain Faruk, Jim Basney, and Jerry Q. Cheng, “Blockchain-Based Decentralized Verifiable
Credentials: Leveraging Smart Contracts for Privacy-Preserving Authentication Mechanisms to Enhance Data
Security in Scientific Data Access,” 2023 IEEE International Conference on Big Data (BigData) Proceedings,
December 2023, pp. 5,493–5,502.
14
markets, manage transactions, and regulate liquidity autonomously, allowing for rapid and
possibly novel financial
Crypto transactions provide a key mechanism for these agents, offering a low-cost,
borderless, and digital alternative to traditional payment systems. In fact, some analysts note that
current KYC and associated measures—which impose identity requirements to establish and
maintain accounts necessary to conduct digital commerce—make crypto uniquely suited for
autonomous agents to execute financial operations without human
Stablecoin firm Circle has demonstrated how AI agents can conduct transactions using digital
assets. In a recent proof of concept, multiple AI agents were automatically awarded USDC for
completing research tasks. These stablecoin payments were processed using Circle’s
Programmable Wallets, enabling the agents to receive compensation without human intervention.
Circle suggests future extensions could enable agents to spend USDC, “allowing them to pay for
services or data from one another, creating a dynamic marketplace of automated interactions.”58
The market impact of AI-driven agents is becoming tangible. For example, Terminal of
Truths, a semi-autonomous AI, gained a large online following and promoted the $GOAT crypto
token, driving its market capitalization to $950 This event highlights AI’s ability to
influence market sentiment, create liquidity, and introduce new financial dynamics, while also
raising concerns about stability and potential manipulation risks. As AI agents gain greater
financial autonomy, blockchain is likely to play an increasingly central role in AI-driven finance.
Blockchain-Based Distributed Computing for AI
AI development and execution require substantial computing power, traditionally provided
by centralized data Blockchain-based distributed computing networks could offer a
decentralized alternative, enabling participants to contribute idle computing resources in
exchange for crypto incentives. Two example projects are identified below. This model might
broaden access to advanced computing resources, allowing independent developers and startups
to participate in AI innovation.
Projects including Render Network and Bittensor are experimenting with this approach by
leveraging blockchain to coordinate graphics processing unit (GPU) power and AI training.
Render Network distributes GPU-intensive tasks for AI applications, allowing for faster and
56 Erik Pounds, “What is Agentic AI?” Nvidia Blog, October 22, 2024.
57 Binance Research, “Exploring the Future of AI Agents in Crypto,” Binance Blog, November 22, 2024; Tomer
Niv, “AI Agents Economy: Why Crypto May Hold the Key to Fund Management,” Forbes Digital Assets,
November 7, 2024.
58 Blessing Adesiji, “Enabling AI Agents with Blockchain,” Circle Blog, November 6, 2024.
59 Binance, 2024.
60 Konstantin F. Pilz, Yusuf Mahmood, and Lennart Heim, AI’s Power Requirements Under Exponential Growth:
Extrapolating AI Data Center Power Demand and Assessing Its Potential Impact on . Competitiveness, RAND
Corporation, RR-A3572-1, 2025.
15
more cost-effective rendering of high-resolution Bittensor facilitates decentralized AI
development by enabling participants to contribute computing power in exchange for crypto
rewards, fostering an open and scalable AI research
Analysis and Future Directions
The intersection of AI and digital assets could disrupt financial services—particularly
through crypto-equipped autonomous agents, frictionless payments, and advanced trading
systems—while posing systemic risks discussed in more detail below. However, aside from
some progress in AI-powered RegTech, most deployments remain experimental or aspirational.
As more deployments are launched, future research could examine case studies, test real-world
performance, and assess the unintended consequences of AI in crypto finance.
It is also important to note that many of the potential deployments discussed, such as real-
time trading analysis or automated payments, can be applied in traditional finance. What sets AI
in crypto apart is its foundation in DLT, which enables a 24/7 global market without
conventional intermediaries.
Crypto-Equipped Autonomous Agents
Crypto could enable autonomous agents to transact in ways the current financial system is
not equipped to manage. These agents could operate independent of human intervention, across
borders and beyond the reach of conventional oversight. Operating independently of human
oversight could sever the relationship between a person and financial activities, which may
challenge existing regulatory regimes that are ultimately linked to real-world identities. It also
raises questions about responsibility and liability for the financial activities of these agents, as
existing frameworks may struggle to determine accountability when transactions are severed
from real-world identities. As a result, regulatory regimes designed to combat money laundering,
prevent fraud, and ensure financial stability may need to adapt to address unique challenges and
opportunities posed by crypto-equipped autonomous agents.
Disintermediated Payments and Trading
AI applications in payments and trading—coupled with stablecoins, CBDCs, tokenized
assets, and smart contracts—could be particularly disruptive through disintermediation.
Stablecoins and CBDCs provide low transaction costs and fast processing times without the price
volatility associated with first-generation cryptocurrencies such as Bitcoin. Asset tokenization
and smart contracts enable fractional ownership, enhanced liquidity, and automation of complex
transactions. What remains to be seen is whether these digital asset innovations will be integrated
61 Render Network, “How It Works,” webpage, undated.
62 Opentensor, “Bittensor Paradigm,” webpage, undated.
16
into the existing regulated financial system or whether they will foster new financial models that
accelerate the rise of new entrants and reduce reliance on contemporary intermediaries.
Systemic Risk
AI deployments in crypto may share systemic vulnerabilities with similar applications in
traditional finance. For example, deep learning’s inherent characteristics (., nonlinearity, non-
determinism, high demand for data) can lead to uniform risk If many crypto
platforms adopt similar AI models, the resulting uniformity and network interconnectedness
might not only magnify individual risks but also lead to failures that propagate across platforms.
Furthermore, while both traditional finance and digital asset markets are experimenting with
AI-driven tools, the 24/7, borderless nature of digital assets means that systemic risks may spread
more quickly than in contemporary financial systems. In this context, the challenges identified in
deep learning could be even more pronounced in digital asset ecosystems.
While this primer focuses on the intersection of AI and crypto in financial services, several
broader issues—such as energy consumption, market speculation, and regulatory arbitrage—fall
outside its primary scope. Historical failures in the crypto ecosystem, such as the FTX collapse
and the failed Facebook Libra project, underscore risks that remain relevant to AI-driven
financial services. Understanding how these risks translate into AI-integrated digital assets
requires further study, particularly as the technology matures and regulatory frameworks evolve.
Policy Considerations
This primer explores the emerging intersection of AI and crypto in financial services as
presented by innovators. Many examples illustrated are in early stages or speculative in nature.
Further examination is necessary as the intersection evolves, particularly with respect to potential
drawbacks and policy Nonetheless, key policy considerations at the intersection of
AI and crypto are beginning to materialize, including:
• Recognizing that many AI applications in digital assets remain experimental or otherwise
obscure:
- regulatory sandboxes could enable controlled experimentation with AI-crypto
integrations, though their design may need to differ from traditional models. For
example, they may not focus on licensing outcomes but instead provide structured
environments for testing operational safeguards, risk controls, or disclosure practices.
63 See Gary Gensler and Lily Bailey, “Deep Learning and Financial Stability,” working paper, November 1, 2020.
64 Given the early stage of many AI-crypto projects and the lack of independently validated outcomes, this primer
relies on illustrative examples drawn from the most credible sources available, while acknowledging significant
evidence and transparency gaps. The examples discussed are included to surface emerging policy questions and
regulatory considerations that may require attention as the intersection of AI and crypto evolves.
17
- regulators may require firms to disclose the degree to which AI is integrated in their
digital asset systems and implement explainability techniques that help auditors and
market participants understand AI decisionmaking processes, similar to emerging
requirements for AI use in traditional financial institutions.
• Domestic and international regulatory gaps persist concerning both AI and crypto.
Special attention should be devoted to:
- monitoring for risks that could spill across platforms, sectors, or jurisdictions
- analyzing the impacts of displacing traditional financial intermediaries
- modernizing regulations to accommodate the emergence of autonomous
• Compliance and enforcement tools will continue to face a cat-and-mouse dynamic as bad
actors develop increasingly sophisticated methods to bypass detection; regulators,
enforcement agencies, and private sector partners will need to continuously update their
monitoring techniques to identify and address emerging illicit finance typologies.
Given the broad and cross-cutting nature of the risks and opportunities at the AI-crypto
intersection, multiple bodies may have overlapping or complementary roles in shaping policy
responses. While not an exhaustive list, relevant policymakers and regulators include financial
regulatory authorities (., Securities and Exchange Commission, Commodity Futures Trading
Commission), central banks (., Federal Reserve), enforcement agencies (., Financial
Crimes Enforcement Network, Federal Bureau of Investigation), international bodies (.,
Financial Action Task Force), data protection and privacy regulators (., Federal Trade
Commission), and technology and innovation offices (., White House Office of Science and
Technology Policy).
65 Although a detailed treatment is beyond this primer’s scope, these issues may also warrant coordinated
international responses to support regulatory harmonization and cross-border governance.
18
Glossary
This glossary provides definitions of technical and policy-relevant terms used throughout the
primer. Definitions may vary by source and context; those included here were selected for their
authoritativeness or developed for appropriateness within the scope and framing of this analysis.
anti-money laundering (AML): a collection of measures and regulations financial institutions
must follow to help prevent, detect, and report possible money laundering activities and other
forms of illicit finance.
autonomous agent: “a system situated within and a part of an environment that senses that
environment and acts on it, over time, in pursuit of its own agenda and so as to effect what it
senses in the future.”66
Bitcoin: the first implementation of blockchain technology and cryptocurrency. “Bitcoin”
(capitalized) refers to the system and protocol that enables peer-to-peer transactions using the
native “bitcoin” (lowercase) cryptocurrency.
blockchain: an implementation of distributed ledger technology that records metadata in a
cryptographically linked, append-only sequence of tamper-evident
central bank digital currency (CBDC): “a digital form of central bank money that is different
from balances in traditional reserve or settlement accounts.”68 Retail CBDC “generally refers
to a CBDC that is widely available to the public for day-to-day use in personal and
commercial transactions” whereas wholesale CBDC “generally refers to a CBDC with a
narrower use case, such as one designed primarily for large-value institutional payments and
not widely available to the general public.”69
66 Stan Franklin and Art Graesser, “Is It an Agent, or Just a Program? A Taxonomy for Autonomous Agents,”
Lecture Notes in Computer Science, Vol. 1193, August 1996, p. 25.
67 National Institute of Standards and Technology, “Blockchain,” Computer Security Resource Center Glossary,
webpage, undated.
68 Bank for International Settlements, Central Bank Digital Currencies, March 2018, p. 4.
69 Board of Governors of the Federal Reserve System, Money and Payments: The . Dollar in the Age of Digital
Transformation: Summary of Public Comments, April 2023, p. 5.
19
cryptocurrency: a type of digital asset “designed to work as a medium of payment or value
exchange” and implemented using digital ledger The terms “cryptocurrency”
or “crypto” are used here interchangeably with “cryptoassets” or “crypto-assets,” reflecting
global usage variation.
cryptographic hash function: a mathematical algorithm that takes any input and turns it into a
fixed-size sequence of numbers and
digital asset: an item of value that “exists only in digital form or that is the digital representation
of another asset.”72
disintermediation: the displacement of intermediaries from a process by direct transactions
between parties.
distributed ledger technology (DLT): a system that facilitates the use of shared ledgers across
multiple nodes, synchronized through a consensus mechanism; DLT ledgers are “designed to
be immutable, tamper-resistant, tamper-evident, and append-only, containing final and
definitive ledger records of confirmed and validated transactions.”73
intermediary: an entity that facilitates transactions between two or more parties.
intermediation: the process of involving intermediaries to facilitate transactions between two or
more parties.
Internet of Things (IoT): “builds out from today’s internet by creating a pervasive and self-
organising network of connected, identifiable and addressable physical objects enabling
application development in and across key vertical sectors through the use of embedded
chips.”74
Know Your Customer (KYC): a set of processes regulated institutions must implement to verify
the identity of customers in compliance with anti-money laundering regulations.
70 International Organization for Standardization, Blockchain and Distributed Ledger Technologies — Vocabulary,
ISO 22739:2024, Edition 2, 2024.
71 National Institute of Standards and Technology, “Cryptographic Hash Function,” Computer Security Resource
Center Glossary, webpage, undated.
72 International Organization for Standardization, 2024.
73 International Organization for Standardization, 2024.
74 Helen Rebecca Schindler, Jonathan Cave, Neil Robinson, Veronika Horvath, Petal Hackett, Salil Gunashekar,
Maarten Botterman, Simon Forge, and Hans Graux, Europe’s Policy Options for a Dynamic and Trustworthy
Development of the Internet of Things: SMART 2012/0053, RAND Corporation, RR-356-EC, 2013, p. xvii.
20
non-fungible token (NFT): a unique, transferable, and indivisible blockchain-based digital record
that represents a physical or digital asset; NFTs are created and managed by smart
Proof of Stake: a blockchain consensus mechanism that selects validators based on the amount of
network tokens they hold and are willing to stake as collateral; developed as an alternative to
the more energy-intensive Proof of Work consensus
Proof of Work: a blockchain consensus mechanism introduced by Bitcoin that requires
participants known as “miners” to solve complex cryptographic puzzles to validate
transactions and add new blocks to the chain, securing the network through computational
public key cryptography: a cryptographic method that uses a public key and a private key, which
enables secure communication between parties without needing to share a secret key in
advance; in a public key cryptography pair, the public key cannot be used to derive the
private key due to computational
regulatory technology (RegTech): “the application of various new technological solutions that
assist highly regulated industry stakeholders, including regulators, in setting, effectuating and
meeting regulatory governance, reporting, compliance and risk management obligations.”79
smart contract: a self-executing program on a blockchain that enforces and automates predefined
rules and
stablecoin: a digital asset designed to “maintain a stable value relative to a reference asset—
typically the . dollar.”81
75 National Institute of Standards and Technology, “Non-Fungible Token,” Computer Security Resource Center
Glossary, webpage, undated.
76 Sunny King and Scott Nadal, “PPCoin: Peer-to-Peer Crypto-Currency with Proof-of-Stake,” August 19, 2012.
77 Satoshi Nakamoto, “Bitcoin: A Peer-to-Peer Electronic Cash System,” 2008.
78 National Institute of Standards and Technology, “Public Key Cryptography (PKC),” Computer Security Resource
Center Glossary, webpage, undated.
79 World Economic Forum, Regulatory Technology for the 21st Century, March 2022, p. 4.
80 Vitalik Buterin, “Ethereum: A Next-Generation Smart Contract and Decentralized Application Platform,”
December 2014.
81 Pablo D. Azar, Garth Baughman, Francesca Carapella, Jacob Gerszten, Arazi Lubis, JP Perez-Sangimino, David
E. Rappoport, Chiara Scotti, Nathan Swem, Alexandros P. Vardoulakis, and Aurite Werman, “The Financial
Stability Implications of Digital Assets,” Economic Policy Review, Vol. 30, No. 2, November 2024, p. 15.
21
tokenization: “the process of recording claims on real or financial assets that exist on a traditional
ledger onto a programmable platform.”82
utility token: a digital asset that grants its owner access to specific goods or services, typically
provided by the token’s
verifiable credentials: tamper-resistant, blockchain-based digital certificates that prove identity
or qualifications without exposing unnecessary personal
82 Iñaki Aldasoro, Sebastian Doerr, Leonardo Gambacorta, Rodney Garratt, and Priscilla Koo Wilkens, April 11,
2023, p. 3.
83 International Organization for Standardization, 2024.
84 Manu Sporny, Dave Longley, David Chadwick, and Ivan Herman, “Verifiable Credentials Data Model ,”
W3C Candidate Recommendation Draft, February 25, 2025.
22
Abbreviations
AI artificial intelligence
AMB Amazon Managed Blockchain
AML anti-money laundering
CBDC central bank digital currency
CDD Customer Due Diligence
CIP Customer Identification Program
DLT distributed ledger technology
GPU graphics processing unit
IoT Internet of Things
KYC Know Your Customer
NFT non-fungible token
TPS transactions per second
23
References
Adesiji, Blessing, “Enabling AI Agents with Blockchain,” Circle Blog, November 6, 2024. As of
June 11, 2025:
aelf, “AI-Powered Smart Contracts: The Impact of AI on Web3 Crypto Transactions,” aelf Blog,
January 7, 2025. As of June 11, 2025:
Aldasoro, Iñaki, Sebastian Doerr, Leonardo Gambacorta, Rodney Garratt, and Priscilla Koo
Wilkens, “The Tokenisation Continuum,” BIS Bulletin, No. 72, April 11, 2023.
Aldasoro, Iñaki, Leonardo Gambacorta, Anton Korinek, Vatsala Shreeti, and Merlin Stein,
“Intelligent Financial System: How AI is Transforming Finance,” BIS Working Papers, No.
1194, June 2024.
Andolfatto, David, and Fernando M. Martin, “The Blockchain Revolution: Decoding Digital
Currencies,” Federal Reserve Bank of St. Louis Annual Report 2021, April 20, 2022. As of
June 11, 2025:
Ast, Federico, Jamilya Kamalova, and Yann Aouidef, “The Kleros Fellowship of Justice, 8th
Generation: Applications Open!” Kleros Blog, September 11, 2024. As of June 11:
Atlantic Council, “Central Bank Digital Currency Tracker,” webpage, February 2025. As of June
11, 2025:
Azar, Pablo D., Garth Baughman, Francesca Carapella, Jacob Gerszten, Arazi Lubis, JP Perez-
Sangimino, David E. Rappoport, Chiara Scotti, Nathan Swem, Alexandros P. Vardoulakis,
and Aurite Werman, “The Financial Stability Implications of Digital Assets,” Economic
Policy Review, Vol. 30, No. 2, November 2024, p. 15.
Bank for International Settlements, Central Bank Digital Currencies, March 2018.
Bank for International Settlements, Annual Economic Report 2022, June 2022.
Bank for International Settlements, The Crypto Ecosystem: Key Elements and Risks, July 2023.
Bedawala, Mustafa, and Arjuna Wijeyekoon, “A Deep Dive on Solana, a High Performance
Blockchain Network,” Visa Crypto Thought Leadership blog, undated. As of June 11, 2025:
24
Bhattacharya, Sukriti, German Castignani, Leandro Masello, and Barry Sheehan, “AI Revolution
in Insurance: Bridging Research and Reality,” Frontiers in Artificial Intelligence, Vol. 8,
April 2025.
Binance Research, “Exploring the Future of AI Agents in Crypto,” Binance Blog, November 22,
2024. As of June 11, 2025:
agents-in-crypto-4685351192205955466
Board of Governors of the Federal Reserve System, Money and Payments: The . Dollar in
the Age of Digital Transformation: Summary of Public Comments, April 2023.
Butler, Conall, and Martin Crane, “Blockchain Transaction Fee Forecasting: A Comparison of
Machine Learning Methods,” Mathematics, Vol. 11, No. 9, May 2023.
Caleb & Brown, “What is XRP? A Beginner’s Guide,” Caleb & Brown blog, January 21, 2025.
As of June 11, 2025:
Chaum, David, “Blind Signatures for Untraceable Payments,” Advances in Cryptology
Proceedings of Crypto 82, 1983.
Chen, Heather, and Kathleen Magramo, “Finance Worker Pays Out $25 Million After Video Call
with Deepfake ‘Chief Financial Officer’,” CNN, February 4, 2024. As of June 11, 2025:
Colyer, Forrest, John Liu, and Michael Greenwald, “The Convergence of AI and Digital Assets:
A New Dawn for Financial Infrastructure,” AWS Web3 Blog, October 19, 2023. As of June
11, 2025:
for-financial-infrastructure/
CrossTech, “Félix,” webpage, undated. As of June 11, 2025:
Cryptohopper, “Algorithm Intelligence (.),” webpage, undated. As of June 11, 2025:
Demirguc-Kunt, Asli, Klapper,Leora, Singer,Dorothe, and Ansar,Saniya, The Global Findex
Database 2021: Financial Inclusion, Digital Payments, and Resilience in the Age of COVID-
19, World Bank Group, July 2022.
Deshmukh, Araddhana Arvind, Prabhakar Kandukuri, Janga Vijaykumar, Anna Shalini, S.
Farhad, Elangovan Muniyandy, and Yousef A. Baker El-Ebiary, “Event-based Smart
Contracts for Automated Claims Processing and Payouts in Smart Insurance,” International
Journal of Advanced Computer Science & Applications, Vol. 15, No. 4, 2024.
25
European Central Bank, “Progress on the Preparation Phase of a Digital Euro – Second Progress
Report,” webpage, December 2, 2024. As of June 11, 2025:
European Telecommunications Standards Institute (ETSI), Permissioned Distributed Ledger
(PDL); Artificial Intelligence for Permissioned Distributed Ledger, ETSI GR PDL 032,
, January 2025, p. 41. As of June 11, 2025:
Faruk, Md Jobair Hossain, Jim Basney, and Jerry Q. Cheng, “Blockchain-Based Decentralized
Verifiable Credentials: Leveraging Smart Contracts for Privacy-Preserving Authentication
Mechanisms to Enhance Data Security in Scientific Data Access,” 2023 IEEE International
Conference on Big Data (BigData) Proceedings, December 2023, pp. 5,493–5,502.
Federal Financial Institutions Examination Council, “Customer Due Diligence — Overview,” in
BSA/AML Manual, May 5, 2018. As of June 11, 2025:
Federal Financial Institutions Examination Council, “Customer Identification Program,” in
BSA/AML Manual, February 2021. As of June 11, 2025:
Financial Action Task Force, Opportunities and Challenges of New Technologies for AML/CFT,
July 2021, pp. 42–43. As of June 11, 2025:
Financial Stability Board, Artificial Intelligence and Machine Learning in Financial Services:
Market Developments and Financial Stability Implications, November 1, 2017.
Financial Stability Board, The Financial Stability Implications of Artificial Intelligence,
November 14, 2024.
Financial Stability Oversight Council, Report on Digital Asset Financial Stability Risks and
Regulation, 2022. As of June 11, 2025:
Franklin, Stan, and Art Graesser, “Is It an Agent, or Just a Program? A Taxonomy for
Autonomous Agents,” Lecture Notes in Computer Science, Vol. 1193, August 1996.
Gensler, Gary, and Lily Bailey, “Deep Learning and Financial Stability,” working paper,
November 1, 2020. As of June 11, 2025:
Geslevich Packin, Nizan, and Yafit Lev-Aretz, “Crypto-Native Credit Score: Between Financial
Inclusion and Predatory Lending,” Cardozo Law Review, Vol. 45, No. 3, February 2024, pp.
845–898.
26
Hanif, Muhammad Shahid, “Synthetic Identities: The Darker Side of Generative AI,” Forbes,
May 29, 2024. As of June 11, 2025:
darker-side-of-generative-ai/
Heikkilä, Melissa, “Dutch Scandal Serves as a Warning for Europe over Risks of Using
Algorithms,” Politico Europe, March 29, 2022. As of June 11, 2025:
using-algorithms/
Hodge, Johnjerica, India Williams, Nicholas Gervasi, and Gabriella Weick, “Artificial
Intelligence in Consumer Lending: Addressing AI-Related Risks,” The Banking Law
Journal, Vol. 141, No. 10, November–December 2024, pp. 475–484.
Immanuel Jones, Kofi, and Swati Sah, “The Implementation of Machine Learning in The
Insurance Industry with Big Data Analytics,” International Journal of Data Informatics and
Intelligent Computing, Vol. 2, No. 2, June 2023, pp. 21–38.
International Organization for Standardization, Blockchain and Distributed Ledger Technologies
— Vocabulary, ISO 22739:2024, Edition 2, 2024. As of June 11, 2025:
Jovanovic, Zorka, Zhe Hou, Kamanashis Biswas, and Vallipuram Muthukkumarasamy, “Robust
Integration of Blockchain and Explainable Federated Learning for Automated Credit
Scoring,” Computer Networks, Vol. 243, April 2024.
Kanaster, Todd, Andrew O’Neill, Rebecca Mun, Josh Stokesberry, Matta Uma Maheswara
Reddy, and Ava Yang, “Crypto and AI: Shaping the Future of the Internet,” S&P Global,
October 1, 2024. As of June 11, 2025:
future-of-the-internet
Kapadiya, Khyati, Usha Patel, Rajesh Gupta, Mohammad Dahman Alshehri, Sudeep Tanwar,
Gulshan Sharma, and Pitshou N. Bokoro, “Blockchain and AI-Empowered Healthcare
Insurance Fraud Detection: An Analysis, Architecture, and Future Prospects,” IEEE Access,
Vol. 10, 2022, pp. 79,606–79,627.
King, Sunny, and Scott Nadal, “PPCoin: Peer-to-Peer Crypto-Currency with Proof-of-Stake,”
August 19, 2012.
Levin, Jonathan, “Welcoming Fraud Detection Innovator Alterya to Chainalysis and Doubling
Down on the Prevention of Illicit Activity,” Chainalysis Blog, January 13, 2025. As of June
11, 2025:
27
Lord, Jennifer, “The Seven-Year Struggle to Hold an Out-of-Control Algorithm to Account,”
interview with Julia Angwin, Hello World, October 8, 2022. As of June 11, 2025:
control-algorithm-to-account
Ma, Jiangqin, and Erfan Mahmoudinia, “Comprehensive Modeling Approaches for Forecasting
Bitcoin Transaction Fees: A Comparative Study,” arXiv, arXiv:, February
2025. As of June 11, 2025:
Martin, Vuk, “What Fees Are Cheaper: USDT or USDC?” Coincodex, webpage, January 6,
2025. As of June 11, 2025:
Nakamoto, Satoshi, “Bitcoin: A Peer-to-Peer Electronic Cash System,” 2008.
National Institute of Standards and Technology, “Blockchain,” Computer Security Resource
Center Glossary, webpage, undated. As of June 11, 2025:
National Institute of Standards and Technology, “Cryptographic Hash Function,” Computer
Security Resource Center Glossary, webpage, undated. As of June 11, 2025:
National Institute of Standards and Technology, “Non-Fungible Token,” Computer Security
Resource Center Glossary, webpage, undated. As of June 11, 2025:
National Institute of Standards and Technology, “Public Key Cryptography (PKC),” Computer
Security Resource Center Glossary, webpage, undated. As of June 11, 2025:
Niv, Tomer, “AI Agents Economy: Why Crypto May Hold the Key to Fund Management,”
Forbes Digital Assets, November 7, 2024. As of June 11, 2025:
hold-the-key-to-fund-management/
Opentensor, “Bittensor Paradigm,” webpage, undated. As of June 11, 2025:
Ozair, Merav, “The Future of Finance: AI Meets Tokenization,” Nasdaq, August 16, 2023. As of
June 11, 2025:
28
Pilz, Konstantin F., Yusuf Mahmood, and Lennart Heim, AI’s Power Requirements Under
Exponential Growth: Extrapolating AI Data Center Power Demand and Assessing Its
Potential Impact on . Competitiveness, RAND Corporation, RR-A3572-1, 2025. As of
June 11, 2025:
Pounds, Erik, “What is Agentic AI?” Nvidia Blog, October 22, 2024. As of June 11, 2025:
Render Network, “How It Works,” webpage, undated. As of June 11, 2025:
Schindler, Helen Rebecca, Jonathan Cave, Neil Robinson, Veronika Horvath, Petal Hackett, Salil
Gunashekar, Maarten Botterman, Simon Forge, and Hans Graux, Europe’s Policy Options
for a Dynamic and Trustworthy Development of the Internet of Things: SMART 2012/0053,
RAND Corporation, RR-356-EC, 2013. As of June 11, 2025:
Sheffield, Cuy, “Making Sense of Stablecoins,” Visa Perspectives blog, April 25, 2024. As of
June 11, 2025:
SingularityDAO, “DYNASET Terms & Conditions,” webpage, undated. As of June 11, 2025:
Sporny, Manu, Dave Longley, David Chadwick, and Ivan Herman, “Verifiable Credentials Data
Model ,” W3C Candidate Recommendation Draft, February 25, 2025. As of June 11,
2025:
Takyar, Akash, “AI in Loan Underwriting: Use Cases, Architecture, Technologies, Solution and
Implementation,” webpage, undated. As of June 11, 2025:
TechBullion, “Smart Contracts: Simplifying DeFi Loan Repayment Processes,” webpage,
January 26, 2025. As of June 11, 2025:
Thomson Reuters, “5 Essential Steps for KYC/AML Onboarding and Compliance,” Thomson
Reuters Legal blog, June 24, 2024. As of June 11, 2025:
compliance/
29
. Department of Justice, The Role of Law Enforcement in Detecting, Investigating, and
Prosecuting Criminal Activity Related to Digital Assets, September 2022. As of June 11,
2024:
launches-nationwide-network
. Department of the Treasury, “Treasury Announces Enhanced Fraud Detection Process
Using AI Recovers $375M in Fiscal Year 2023,” press release, February 28, 2024. As of
June 11, 2025:
Vasilchenko, Alex, “AI Biometric Authentication for Enterprise Security,” MobiDev Blog,
January 31, 2025. As of June 11, 2025:
West, Jarrod, and Maumita Bhattacharya, “Intelligent Financial Fraud Detection: A
Comprehensive Review,” Computers & Security, Vol. 57, 2016, pp. 47–66.
World Economic Forum, Regulatory Technology for the 21st Century, March 2022.
World Network, “A New Identity and Financial Network,” webpage, undated. As of June 11,
2025:
