Key Takeaways
- • Wall Street institutional investors are expanding AI use for internal efficiency tasks such as research, document preparation, and idea generation, but are not using AI as the primary decision-maker in investment judgments, according to reports covering an October 27 Bloomberg-hosted event.
- • Executives from Citadel, Bridgewater Associates, and Man Group presented the view that current AI capability has not reached a level where it can serve as the main investment decision-maker.
- • Citadel restricts its investment teams from using AI beyond meeting preparation or idea testing and prohibits its use in documenting investment strategies.
- • Firms are devoting significant resources to cybersecurity and monitoring to contain AI risks, with Bridgewater stating it cannot be certain AI agents will not put the firm at risk.
- • Some hedge funds are privately building investment models by retraining AI on decades of market data, a effort that requires large computing capacity and drives exclusive cloud contracts and proprietary data centers for GPU access.
Summary
Wall Street institutional investors are broadening their use of artificial intelligence for internal tasks while stopping short of delegating investment decisions to it, according to reports summarizing a discussion hosted by Bloomberg on October 27. Executives from Citadel, Bridgewater Associates, and Man Group made the case that the technology has not yet reached a stage where it can act as the main decision-maker in investment judgments.
The reports describe two main approaches. The first, adopted by most banks, hedge funds, and private equity firms, uses large language models to improve productivity in research, document preparation, idea generation, and internal information sharing. The second, pursued privately by some hedge funds, involves retraining AI on decades of financial market data rather than human language to improve price prediction models. That work requires enormous computational power, leading some firms to negotiate exclusive contracts with major cloud providers or build their own data centers to secure GPUs.
Citadel's investment teams have built AI agents that replicate themselves to conduct independent market analysis, but the firm limits its investment teams, which use AI as assistants, to meeting preparation and idea testing, and bars its use in documenting investment strategies. Citadel's interim CTO Andrew Zeyani said the firm spends considerable time monitoring AI to ensure it does not exceed laboratory controls. Bridgewater's advanced AI lab is working to refine research and workflow processes across the organization. Aaron Linski, a partner at Bridgewater, said, "We cannot be certain that AI agents will not put Bridgewater at risk." Man Group is using AI to shift employees toward more interesting work and reduce tedious tasks.
The attendees agreed that AI holds no inherent advantage in financial investing. Greg Bond, CIO of Man Group, stressed that human judgment, collaboration, and critical thinking remain central to investment decisions and warned that people who hand all their work to AI are becoming its slaves. The participants argued that investment teams should rely on human interaction rather than AI alone, and one view held that in-person group discussion is preferable to PowerPoint-driven meetings. They noted that humans design AI systems and remain the driving force behind financial markets.
Note: this summary is based on a machine-translated Korean-language article from chosun.com and on web search snippets; the original article could not be retrieved in full.
The reports describe two main approaches. The first, adopted by most banks, hedge funds, and private equity firms, uses large language models to improve productivity in research, document preparation, idea generation, and internal information sharing. The second, pursued privately by some hedge funds, involves retraining AI on decades of financial market data rather than human language to improve price prediction models. That work requires enormous computational power, leading some firms to negotiate exclusive contracts with major cloud providers or build their own data centers to secure GPUs.
Citadel's investment teams have built AI agents that replicate themselves to conduct independent market analysis, but the firm limits its investment teams, which use AI as assistants, to meeting preparation and idea testing, and bars its use in documenting investment strategies. Citadel's interim CTO Andrew Zeyani said the firm spends considerable time monitoring AI to ensure it does not exceed laboratory controls. Bridgewater's advanced AI lab is working to refine research and workflow processes across the organization. Aaron Linski, a partner at Bridgewater, said, "We cannot be certain that AI agents will not put Bridgewater at risk." Man Group is using AI to shift employees toward more interesting work and reduce tedious tasks.
The attendees agreed that AI holds no inherent advantage in financial investing. Greg Bond, CIO of Man Group, stressed that human judgment, collaboration, and critical thinking remain central to investment decisions and warned that people who hand all their work to AI are becoming its slaves. The participants argued that investment teams should rely on human interaction rather than AI alone, and one view held that in-person group discussion is preferable to PowerPoint-driven meetings. They noted that humans design AI systems and remain the driving force behind financial markets.
Note: this summary is based on a machine-translated Korean-language article from chosun.com and on web search snippets; the original article could not be retrieved in full.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.