About | HeinOnline Law Journal Library | HeinOnline Law Journal Library | HeinOnline

LOG IN
LOG IN


1 [1] (July 25, 2025)

handle is hein.crs/aiugei0001 and id is 1 raw text is: Congressional Research Servie Inforrning the legislative debate since 1914 S July 25, 2025 Artificial Intelligence and Derivatives Markets: Policy Issues Al Use and Derivatives: Background The use of artificial intelligence (AI) in derivatives markets has ballooned in recent decades, creating new policy issues as well as amplifying existing ones. As of 2023, 99% of leading financial services firms in derivatives markets reported that they had deployed Al in some capacity. Another survey of Al usage more broadly by financial firms found that, as of 2024, 89% of respondents were already using generative Al, although primarily for internal (not client-facing) usage. Generative AI refers to Al models-in particular those that use machine learning (ML) and are trained on large volumes of data-that are able to generate new content, such as computer code, text, or videos. Most firms using ML or other forms of predictive Al were using them for risk management, fraud detection and prevention, operations, and compliance purposes. An in-depth study of Al in the derivatives markets in 2024 by the Technology Advisory Committee of the Commodity Futures Trading Commission (CFTC) noted that the question of how Al models are used throughout the financial services sector is highly relevant for the CFTC and for the derivatives markets it oversees. Al could potentially automate processes in derivatives trading, such as risk management; surveillance; fraud detection; and the identification, execution, and back-testing of trading strategies. Academics note that increased Al use has led to greater efficiencies in areas such as back-office processing and trade execution. More recently, generative Al has enabled investment firms to process large quantities of unstructured data to enhance their analytic trading tools. The growing use of Al also raises new risks and a number of questions for congressional oversight of the CFTC and derivatives regulation. Potential policy issues include how to ensure strong cybersecurity and other protections against third-party risks from services provided by outside information technology firms. Other questions involve how to assure that generative Al does not lead to market manipulation and, more broadly, how to ensure market stability and transparency amid faster trade execution, including by Al models. On January 25, 2024, the CFTC issued a Request for Comment on the Use of Artificial Intelligence in CFTC-Regulated Markets, raising issues discussed below and others. No further guidance has been issued to date. Third-Party Risks The CFTC broadly regards third-party risk as the potential for harm that arises from reliance on outside parties (third parties) to perform services or activities on behalf of a registered entity, such as a swap dealer or futures commission merchant. A 2024 Institute of International Finance survey of Al usage found that 94% of financial firms responding anticipated their use of third-party AI/ML solutions to increase in the next 12 months. Third-party risk may encompass operational, financial, cybersecurity, or regulatory concerns stemming from the use of third-party vendors or service providers. Concentration and Cybersecurity Risks A May 2025 study by the Government Accountability Office (GAO) warned that financial instability may arise from reliance on a concentrated group of third-party Al service providers (e.g., cloud providers, data providers, and technology providers). A failure at one of these provider companies may impact an outsized set of financial companies, increasing systemic risk for the sector. In June 2025, CFTC Commissioner Kristin Johnson called cybersecurity risks a growing concern that could be amplified by concentration risk. In one significant cybersecurity breach in January 2023, ION Cleared Derivatives, a provider of back-office services for many global futures commission merchants used in clearing and settlement for a large number of global transactions, triggered a ripple effect of disruptions across markets. In February 2025, Bybit, a cryptocurrency exchange, lost nearly $1.5 billion from a hacking incident-one of the single largest losses from a crypto exchange. The hack appeared to stem from a third-party-provided critical infrastructure system. Although hacking and cybersecurity remain long-standing risks, the increased use of Al means cyber risk can potentially endanger more key financial processes. Trading Risks The CFTC study also identified certain trading risks, particularly from generative Al, which may interest policymakers and Congress through its oversight role. Challenges from High-Speed Trading One of the most notable effects of the adoption of algorithmic trading strategies has been to increase the speed of reactions to information. The CFTC study flagged that high-speed algorithmic trading, in cases where humans have been out of the loop, has at times resulted in market disruptions and market instability. It cites the example in August 2012 of Knight Capital Group-then a registered broker-dealer and formerly one of the largest traders of U.S. equities-which deployed a faulty trading algorithm which, though not AI-powered, had consequences that demonstrate more broadly the necessity of human oversight in automated decision-making. The algorithm mistakenly placed approximately $7 billion in orders across more than 150 stocks in less than an hour, ultimately causing $460 million in losses to Knight Capital. ,ittps://crsreports~congress.gov