At FactSet, AI made analysts' reports richer but their forecasts less accurate, so investors may be right to doubt them.
Why do investors react weakly to AI-assisted analyst reports?
This explores why markets might discount analyst reports written with AI help, and whether the cause is the reports themselves, the analysts producing them, or how readers judge AI involvement.
This explores why investors might give less weight to analyst reports produced with AI assistance. First, a limit: the collection has no study that directly measures how investors or stock prices respond to these reports. It does have strong evidence on the likely cause, though, and that evidence points away from the obvious explanation. The closest study is FactSet's rollout of generative AI to its analysts Does AI-enriched analyst reports improve forecast accuracy?. Reports became noticeably richer, citing 26% more sources and covering 24% more ground, yet the analysts' forecasts got less accurate. The telling detail is that a machine-learning model given the same inputs lost no accuracy. The information wasn't worse. The analysts were overloaded. If investors trust these reports less, they may be responding sensibly to a real drop in forecast quality rather than to a bias against AI.
That changes the question from "why don't investors trust AI?" to "why does more material produce worse judgment?" Several notes in the collection describe how this happens. Polished AI output can stand in for thinking, borrowing the old assumption that professional-looking work reflects expert judgment Does polished AI output trick audiences into trusting it?. A related idea is 'cognitive surrender': checking is costly and fluent text feels trustworthy, so people stop verifying what the AI hands them When do users stop checking whether AI output is actually backed?. A longer report can hide a weaker forecast, and experienced readers of analyst research may be marking it down for exactly that reason.
A second explanation involves perception rather than quality. When AI use is visible, people downgrade the work, but only slightly. Readers rated an identical news article lower when it carried an AI disclosure, by less than 0.15 points on a 7-point scale Does disclosing AI assistance make readers trust articles less?. Workers expect to be seen as less competent and less diligent when they use AI, so they disclose it less Do people fear judgment when they use AI at work?. LinkedIn's CEO says the same reputational cost slows adoption of AI post-writing, because readers notice and call out machine-written prose Why aren't LinkedIn users adopting AI post-writing tools?. This penalty is real but small. It probably adds to a weak reaction more than it explains one.
What you might not have expected: AI may help forecasting more when it supports the analyst's judgment than when it piles on material. Learning to Guide has the machine point out which parts of the input matter instead of making the decision. In that work, this removed anchoring bias (over-relying on whatever the AI suggests first) and kept responsibility with the human Can AI guidance reduce anchoring bias better than AI decisions?. In fields like venture investing, where human experts beat chance only modestly, LLMs can already outperform them on their own Can language models beat human venture capital experts?. So the weak point may not be AI in finance at all, but a setup where AI adds volume and a human has to absorb it.
Sources 8 notes
FactSet's GenAI integration increased report richness by 26% in sources and 24% in coverage, but forecast accuracy declined under heavier processing demands. A machine-learning benchmark on identical inputs showed no accuracy loss, suggesting human cognitive constraints rather than information quality drove the decline.
Generative AI produces visually sophisticated outputs without underlying judgment, leveraging the historical heuristic that professional-looking work signals expert thinking. This substitution is especially risky for less experienced workers who lack domain knowledge to evaluate substance beyond form.
Users systematically accept AI outputs without verification because checking is costly and fluent output builds false confidence. This receiver-side surrender—measured in studies showing 80% unchallenged adoption—is what enables inflationary token systems to function at scale.
Both human raters (n=1,970) and LLM raters (n=2,520) scored an identical news article lower when it included an AI disclosure statement, but the penalty was small—less than 0.15 points on a 7-point scale.
Across four experiments with 4,439 participants, people using AI expected others to judge them as less competent and diligent, and reported lower willingness to disclose AI use to managers and colleagues. The gap suggests a social cost that users foresee and act on.
Show all 8 sources
LinkedIn's CEO attributes lower-than-expected adoption of AI post-writing to reputational risk: posts are publicly attributed and readers who detect machine-generated prose call it out, reducing the poster's economic opportunity. This contrasts with private AI use in drafting and email.
Learning to Guide eliminates anchoring bias and unassisted hard cases by having machines supply interpretive guidance rather than autonomous decisions, keeping responsibility with humans while improving their judgment through enhanced perception.
VCBench shows several LLMs exceed human baselines in founder-success prediction, with DeepSeek-V3 achieving 6× market-index precision. In sparse-signal forecasting where experts only modestly beat chance, even raw LLM capability suffices to clear the human bar.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- Evidence of a social evaluation penalty for using AI
- GenAI as a Power Persuader: How Professionals Get Persuasion Bombed When They Attempt to Validate LLMs
- What Influences Readers' and Writers' Perceived Necessity of AI Disclosure?
- Being honest about using AI at work makes people trust you less, research finds
- Understanding Reader Perception Shifts upon Disclosure of AI Authorship
- Humans learn to prefer trustworthy AI over human partners
- Can AI Do Strategy?
- AI-Powered (Finance) Scholarship