AI’s Reverberations across Finance

Source details

Sales of software, hardware, and services for AI systems will climb 29 percent this year to $166 billion and top $400 billion in 2027, according to International Data Corp.
Financial sector spending will more than double to $97 billion in 2027, with a 29 percent compound annual growth rate—the fastest of five major industries.
In June 2023, JPMorgan Chase & Co. had 3,600 AI help-wanted postings (Evident Insights Ltd.).
The debut of OpenAI’s ChatGPT in November 2022 quickly topped 100 million users to become the fastest-growing application in internet history.
Nearly half of hedge funds use ChatGPT professionally; more than two-thirds of those use it to write marketing text or summarize reports or documents (BNP Paribas survey of funds with $250 billion in total assets).

Investment and asset management:
- Amundi SA (with €2 trillion ($2.1 trillion) under management and more than 100 million clients) is building AI infrastructure for macroeconomic and market research and using AI-based tools to customize portfolios and gauge real-time client sentiment.
- AI applications include robo-advising tools and portfolio customization based on client risk preferences.
Banking and large financial firms:
- JPMorgan spends more than $15 billion a year on technology, deploys almost a fifth of its approximately 300,000 employees in tech, and has an AI research group employing 200. AI supports prospecting, marketing, risk management, fraud prevention, payment processing, and money movement systems.
Hedge funds and trading:
- Longstanding pioneers in cutting-edge tech, hedge funds are embracing generative AI for operational and marketing tasks.

Central bank applications:
- Brazil’s central bank built a prototype robot to download and categorize consumer complaints using machine learning.
- The Reserve Bank of India hired McKinsey and Accenture to help deploy AI and related analytics in supervision.
- The Bank of Canada built a machine learning tool to detect anomalies in regulatory submissions; automated daily runs free up staff for follow-up.
- The European Central Bank automates classification of data from 10 million business and government entities, scrapes websites to track product prices in real time, and uses AI to help supervisors parse news, reports, and filings. The ECB is exploring large language AI models to help write code, test software, and make public communications easier to understand.
Standard-setting and findings:
- The Basel Committee on Banking Supervision found that AI can make lending more efficient in credit decisions and in thwarting money laundering, while noting risks such as opaque model outcomes, potential for bias, and greater cyber risks.
- The BIS Innovation Hub’s Project Aurora showed neural networks can help detect money laundering by identifying patterns and anomalies that traditional methods can’t.

Herding and macroprudential implications:
- US Securities and Exchange Commission Chair Gary Gensler warned AI could heighten financial fragility and “promote herding—with individual actors making similar decisions because they are getting the same signal from a base model or data aggregator.” He is charged with protecting a $46 trillion stock market that makes up two-fifths of the world total.
- Gensler and coauthor Lily Bailey (2020) cautioned that regulations rooted in earlier eras “are likely to fall short in addressing the systemic risks posed by broad adoption of deep learning in finance.”
Model limitations and crisis amplification:
- AI models trained on past data may fail in unprecedented situations and can amplify negative feedback loops, contributing to systemic risks and possible crisis exacerbation (term “polycrisis” referenced).
- Opaque AI applications create new systemic risks and may quickly amplify adverse dynamics during shocks.

Data quality and unstructured data:
- Cleaning and making exponentially growing data intelligible is a key issue, especially for unstructured data; AI can help humans make important distinctions.
Human–AI interaction:
- Experts emphasize AI as an augmentation tool: AI “cannot replace the brain,” and wholly AI-driven processes could be dangerous. Interpretation, understanding, and human checks remain essential.
Talent and competitive pressures:
- There is intense demand for AI talent in finance; Evident Insights founder Alexandra Mousavizadeh described it as “a war for talent.”
- Three of the top 10 cities in Evident’s talent index are in India.

Benefits:
- AI offers potential for better asset protection, market prediction, efficiency in credit decisions, fraud and money-laundering detection, customer service personalization, and supervisory automation.
Risks requiring policy action:
- Opaque models, bias, cyber risks, potential for herding and amplified systemic shocks, and limits in unprecedented scenarios.
- Supervisory and macroprudential frameworks will need new thinking to assess systemwide risks and distinguish responsible from irresponsible innovation.
Research and monitoring priorities:
- Continue testing AI applications in supervision and financial-stability analysis.
- Improve data cleaning and governance for unstructured data.
- Strengthen human oversight, ethical compliance, and model interpretability.

Jeff Kearns, Finance & Development, December 2023.

Ai’S Reverberations Across Finance
- Ai’S Reverberations Across Finance (Markdown version)
- Ai’S Reverberations Across Finance (PDF)