This semi-monthly column highlights news, government documents, NGO/IGO papers, conferences, industry white papers and reports, academic papers and speeches, and central bank actions on the subject of AI’s fast paced impact on the banking and finance sectors. The chronological links provided are to the primary sources, and as available, indicate links to alternate free versions.
NEWS:
The rise of the AI investment banker, Financial Times, April 30, 2025. The AI chatbot coming for analysts’ jobs Junior bankers are notoriously overworked. A new start-up is hoping to change that, but it might also put the 20-somethings out of a job in the process. Rogo, an artificial intelligence start-up developed by former Lazard analyst Gabriel Stengel, aims to automate some of the grunt work that fills the days of junior investment bankers. The chatbot that Rogo has developed can assess a company’s market position and pull basic valuation comparisons. It’s already been deployed at investment banks Moelis and Nomura, as well as investment firms Tiger Global and GTCR. Stengel said he’d spend days as an analyst at Lazard triangulating research reports and filings with the Securities and Exchange Commission to calculate a “peak sales” valuation ratio. Rogo can now do that in minutes. In its latest funding round, the start-up raised $50mn from a group of investors led by Thrive Capital, valuing it at $350mn, up from $80mn four years ago. The world’s biggest banks and trading firms are already spending billions of dollars on AI applications. At JPMorgan Chase, employees have access to an in-house large language model, while private capital firms have their own AI models to assess buyouts. “We’re training reasoning models that think like investors and investment bankers,” said Stengel. “You run these big experiments to see, hey, can we be as thoughtful as a partner at Tiger Global? Can we be as thoughtful as Blair Effron at Centerview?” So, is it time to say goodbye to the humble analyst? The verdict’s still out. Some bankers think AI tools will eventually allow Wall Street banks to cut the number of entry-level positions. Others think they’ll free up banks to work on more deals, meaning more hands needed on deck. For now, though, MDs will still need someone to call at 3am when they discover that their pitchbook uses Times New Roman instead of Garamond.
AI Is Transforming Document Handling In Heavily Regulated Industries, Forbes May 1, 2025. Secure and regulatory-compliant document exchange is the basis of multiple sectors, from finance to healthcare. Banks and their clients need to exchange official documents between each other to manage their relationship and to execute transactions, and physicians need to share patient data between each other across healthcare systems. These areas, like banking, finance and healthcare, are heavily regulated to protect the confidentiality of personal data. In the U.S., for example, patient healthcare data is stringently protected by the Health Insurance Portability and Accountability Act (HIPAA). While these regulations are in place for good reasons, they put pressure on companies, organizations and individuals to maintain compliance and operational efficiency. It becomes difficult to verify documents or to integrate data into electronic records systems in a speedy way. Continuing to manage document exchanges manually, or using older technologies like fax machines, creates liabilities for firms and organizations. Not only is data at greater risk of being breached by bad actors due to failed delivery, interception and unauthorized access, but sloppy data collection and constraints in scaling data collection can lead to fines, as regulatory audits can flag discrepancies.
The CFO Imperative How Finance Leaders Are Staying Ahead In A Volatile World. Brex, April 28, 2025. The finance function has always been a guardian of stability, but today, stability itself is a moving target. We are operating in a world where the pace of change has outgrown traditional planning cycles. New tariffs reshuffle global supply chains overnight. AI moves faster than regulatory frameworks can catch up. Economic volatility, typically cyclical, now feels perpetual. In this environment, finance leaders are being judged less on how well they predict the future and more on how well they adapt their plans. At Brex, we wanted to understand how finance leaders are meeting this moment. To do so, we surveyed 500 senior finance executives across global enterprises at the beginning of the year. Then, when the tariff shock hit in April, we went back to the field to see how fast sentiment and strategies were evolving.
What we found wasn’t hesitation or fear; it was swift, deliberate reprioritization.
- First, economic volatility is rewriting the playbook.
Strong early-in-the-year optimism around growth, IPOs, and M&As eroded sharply post-tariffs, with growth positivity alone dropping nearly 20 points. Expansion plans have been replaced with calls for stronger risk management frameworks, and leaders are delaying major moves not out of caution, but out of strategic patience. They’re quickly recognizing that moving fast doesn’t always mean moving first. - Second, the imperative for speed is universal.Whether it’s payment operations, reconciliation processes, or decision-making itself, finance leaders overwhelmingly agree: speed is no longer a competitive advantage; it’s a baseline requirement for survival. Those who can’t move capital, data, and decisions quickly across global operations will be left behind.
- Third, the finance stack is being pressure-tested.
Complexity has become a tax on growth, and leaders are aggressively consolidating vendors and systems as they demand breadth and depth from fewer partners. With speed a top driver of success, there’s no margin left for fragmented, brittle financial systems that slow teams down. - Fourth, AI is forcing a reckoning.
Ninety-four percent of finance leaders are prioritizing AI skill-building for their teams, yet nearly three-quarters cite significant challenges with adoption. The pressure to demonstrate ROI is real and immediate. Finance leaders are realizing that AI isn’t just about efficiency gains; it’s about redefining the core capabilities of the finance team for a new era. - Finally, the role of the finance leader is expanding whether we like it or not. Finance is no longer a back-office function.
It’s a strategic nerve center deeply connected to sales, risk, product development, and global operations. With this broader mandate, however, comes a larger question: are we willing to empower smarter, decentralized decision-making across the organization? Or will we cling to outdated control models and become the bottlenecks we fear?
Anthropic Economic Index: AI’s Impact on Software Development. April 28, 2025. Jobs that involve computer programming are a small sector of the modern economy, but an influential one. The past couple of years have seen them changed dramatically by the introduction of AI systems that can assist with—and automate—significant amounts of coding work. In our previous Economic Index research, we found very disproportionate use of Claude by US workers in computer-related occupations: that is, there were many more conversations with Claude about computer-related tasks than one would predict from the number of people working in relevant jobs. It’s the same in the educational context: Computer Science degrees—which involve large amounts of coding—show highly disproportionate AI use. To understand these changes in more detail, we conducted an analysis of 500,000 coding-related interactions across Claude.ai (the “default” way that most people interact with Claude) and Claude Code (our new specialist coding “agent” that can independently accomplish chains of complex tasks using a variety of digital tools).
We found three key patterns:
- The coding agent is used for more automation. 79% of conversations on Claude Code were identified as “automation”—where AI directly performs tasks—rather than “augmentation,” where AI collaborates with and enhances human capabilities (21%). In contrast, only 49% of Claude.ai conversations were classified as automation. This might imply that as AI agents become more commonplace, and as more agentic AI products are built, we should expect more automation of tasks.
- Coders commonly use AI to build user-facing apps. Web-development languages such as JavaScript and HTML were the most common programming languages used in our dataset, and user interface and user experience tasks were among the top coding uses. This suggests that jobs that center on making simple applications and user interfaces may face disruption from AI systems sooner than those focused purely on backend work.
- Startups are the main early adopters of Claude Code, while enterprises lag behind. In a preliminary analysis, we estimated that 33% of conversations on Claude Code served startup-related work, compared to only 13% identified as enterprise-relevant applications. The adoption gap suggests a divide between nimbler organizations using cutting-edge AI tools, and traditional enterprises.
From Anthropic: AI’s Impact on Software Development “In our previous Economic Index research, we found very disproportionate use of Claude by US workers in computer-related occupations: that is, there were many more conversations with Claude about computer-related tasks than one would predict from the number of people working in relevant jobs. It’s the same in the educational context: Computer Science degrees—which involve large amounts of coding—show highly disproportionate AI use.” [more inside]
60% of AI agents work in IT departments – here’s what they do every day. ZDNet. April 30, 2025. Everyone wants an AI agent – but for what? Here’s how they’re being used today.
AI tools mostly fumble basic financial tasks, study finds, Washington Post, April 23, 2025. There’s no shortage of tech leaders predicting that AI will replace humans, fulfilling even complex tasks with speed and accuracy. A new independent study offers grounds to question the hype. A test of 22 general-purpose AI models from OpenAI, Anthropic, x.AI, Meta, Google and other leading players in artificial intelligence found that all scored less than 50 percent accuracy, on average, for simple tasks required of entry-level financial analysts. “The level of BS we see out there is absurd,” said Rayan Krishnan, the chief executive of Vals AI, which conducted the study. Although the latest AI models score well on public benchmarks measuring math or coding skills, the questions for those tests are widely circulated online and have likely become part of the data that AI systems are trained on, Krishnan said…
GPT 4.1, 4.1 Mini, and 4.1 Nano evaluated on all benchmarks! We just evaluated GPT 4.1, GPT 4.1 Mini, and GPT 4.1 Nano on all benchmarks!
- GPT 4.1 delivers impressive results with a 75.5% average accuracy across benchmarks.
- Impressive performance on proprietary benchmarks! GPT 4.1 is now the leader on CorpFin (71.2%), and shows strong performance on CaseLaw (85.8%, 4/53), and MMLU Pro (80.5%, 6/33).
- GPT 4.1 Nano and GPT 4.1 Mini bring AI to time-sensitive applications with an outstanding latency of only 3.62s and 6.60s respectively while still achieving 59.1% and 75.1% average accuracy.
- Compact but capable! Despite its size, GPT 4.1 Mini performs admirably on Math500 (88.8%, 10/36) and MGSM (87.9%, 20/34).
- Size versus performance tradeoff: The smaller models do show lower performance on some complex tasks, with GPT 4.1 Nano ranking near the bottom on MMLU Pro (62.3%, 30/33) and MGSM (69.8%, 32/34).
PAPERS:
NBER
Ethnographic Records, Folklore, and AI, Working Paper 33700. DOI 10.3386/w33700. Issue Date In this Handbook chapter, I examine how integrating ethnographic and folklore records has shaped research on culture and economics in the 21st century. Advances in text analysis techniques and the incorporation of historical and satellite data have transformed the field. I explore how George Peter Murdock’s ethnographic contributions and Yuri Berezkin’s seminal folklore motif index have been utilized to shed light on the roots of comparative development. I conclude by proposing a methodology for leveraging Large Language Models to extract cultural insights from folklore motifs, demonstrating how ancestral narratives can complement ethnographic records and offer valuable perspectives on societal norms and the historical forces shaping economic behavior today.
How Good is AI at Twisting Arms? Experiments in Debt Collection. Working Paper 33669. DOI 10.3386/w33669. Issue Date How good is AI at persuading humans to perform costly actions? We study calls made to get delinquent consumer borrowers to repay. Regression discontinuity and a randomized experiment reveal that AI is substantially less effective than human callers. Replacing AI with humans six days into delinquency closes much of the gap. But borrowers initially contacted by AI have repaid 1% less of the initial late payment one year later and are more likely to miss subsequent payments than borrowers who were always called by humans. AI’s lesser ability to extract promises that feel binding may contribute to the performance gap.
NGO/IGOs
IMF – The Global Impact of AI: Mind the Gap. By Eugenio M Cerutti, Antonio I Garcia Pascual, Yosuke Kido, Longji Li, Giovanni Melina, Marina Mendes Tavares, Philippe Wingender. April 11, 2025. This paper examines the uneven global impact of AI, highlighting how its effects will be a function of (i) countries’ sectoral exposure to AI, (ii) their preparedness to integrate these technologies into their economies, and (iii) their access to essential data and technologies. We feed these three aspects into a multi-sector dynamic general equilibrium model of the global economy and show that AI will exacerbate cross-country income inequality, disproportionately benefiting advanced economies. Indeed, the estimated growth impact in advanced economies could be more than double that in low-income countries. While improvements in AI preparedness and access can mitigate these disparities, they are unlikely to fully offset them. Moreover, the AI-driven productivity gains could reduce the traditional role of exchange rate adjustments due to AI’s large impact in the non-tradable sector—a mechanism akin to an inverse Balassa-Samuelson effect.
Measuring Human Leadership Skills with AI Agents. Working Paper 33662. DOI 10.3386/w33662. Issue Date We show that leadership skill with artificially intelligent (AI) agents predicts leadership skill with human groups. In a large pre-registered lab experiment, human leaders worked with AI agents to solve problems. Their performance on this “AI leadership test” was strongly correlated (ρ=0.81) with their causal impact as leaders of human teams, which we estimate by repeatedly randomly assigning leaders to groups of human followers and measuring team performance. Successful leaders of both humans and AI agents ask more questions and engage in more conversational turn-taking; they score higher on measures of social intelligence, fluid intelligence, and decision-making skill, but do not differ in gender, age, ethnicity or education. Our findings indicate that AI agents can be effective proxies for human participants in social experiments, which greatly simplifies the measurement of leadership and teamwork skills.
AI and Productivity in Europe. April 4, 2025. Misch, Florian ; Park, Ben ; Pizzinelli, Carlo ; Sher, Galen. The discussion on Artificial Intelligence (AI) often centers around its impact on productivity, but macroeconomic evidence for Europe remains scarce. Using the Acemoglu (2024) approach we simulate the medium-term impact of AI adoption on total factor productivity for 31 European countries. We compile many scenarios by pooling evidence on which tasks will be automatable in the near term, using reduced-form regressions to predict AI adoption across Europe, and considering relevant regulation that restricts AI use heterogeneously across tasks, occupations and sectors. We find that the medium-term productivity gains for Europe as a whole are likely to be modest, at around 1 percent cumulatively over five years. While economcially still moderate, these gains are still larger than estimates by Acemoglu (2024) for the US. They vary widely across scenarios and countries and are sustantially larger in countries with higher incomes. Furthermore, we show that national and EU regulations around occupation-level requirements, AI safety, and data privacy combined could reduce Europe’s productivity gains by over 30 percent if AI exposure were 50 percent lower in tasks, occupations and sectors affected by regulation.
AI Adoption and Inequality – April 4, 2025. Rockall, J Emma ; Mendes Tavares, Marina ; Pizzinelli, Carlo. There are competing narratives about artificial intelligence’s impact on inequality. Some argue AI will exacerbate economic disparities, while others suggest it could reduce inequality by primarily disrupting high-income jobs. Using household microdata and a calibrated task-based model, we show these narratives reflect different channels through which AI affects the economy. Unlike previous waves of automation that increased both wage and wealth inequality, AI could reduce wage inequality through the displacement of high-income workers. However, two factors may counter this effect: these workers’ tasks appear highly complementary with AI, potentially increasing their productivity, and they are better positioned to benefit from higher capital returns. When firms can choose how much AI to adopt, the wealth inequality effect is particularly pronounced, as the potential cost savings from automating high-wage tasks drive significantly higher adoption rates. Models that ignore this adoption decision risk understating the trade-off policymakers face between inequality and efficiency.
IMF-IOSCO Online Conference on Market-Based Finance. April 21, 2025. Capital markets are becoming more important for financial stability as they grow and change. The second IMF-IOSCO conference focuses on how artificial intelligence (AI) is transforming these markets and the shift from traditional mutual funds to ETFs.