AI & Models
Citrini Research models potential economic risks of agentic AI
Citrini Research published a scenario suggesting that agentic AI could cause significant economic destruction, including doubled unemployment and major stock market declines, within two years.
On Sunday, the analyst group Citrini Research published a report outlining a scenario where the integration of agentic AI—defined as AI systems capable of performing tasks autonomously—could bring on mass economic destruction over the next two years. The firm emphasizes that the report is more of a scenario than a prediction, illustrating a potential structural risk rather than a firm forecast. The scenario is framed as a report written from two years in the future, examining the gradual unspooling of the economy rather than existential, “Skynet-style” misalignment risks.
The core mechanism of this scenario is a negative feedback loop driven by corporate cost-cutting and automation. The model looks closely at the implications of integrating AI agents into the economy at large, specifically what happens when outside contractors are replaced by cheaper AI. As Citrini Research describes it, the cycle begins when “AI capabilities improved, companies needed fewer workers, white collar layoffs increased, displaced workers spent less, margin pressure pushed firms to invest more in AI, AI capabilities improved…” The analyst group notes that “It was a negative feedback loop with no natural brake…The system turned out to be one long daisy chain of correlated bets on white-collar productivity growth.”
In this hypothetical model, the economic consequences of this unspooling are severe. Within the two years of the scenario, unemployment has doubled, and the total value of the stock market has fallen by more than a third. This bear case is similar to the “Death of SaaS” scenario, but it goes further by implicating any business model that involves optimizing transactions between companies. While the report has caused significant discussion online and not all observers agree with its conclusions, identifying the specific point where the model fails remains difficult. For instance, while some question whether companies are ready to hand over purchasing decisions to autonomous AI agents, the scenario suggests that because many of these decisions are already handled by third-party contractors, the transition may be more plausible than it initially appears.
Why it matters
This report highlights a shift in AI risk discourse from existential “Skynet” fears to practical, structural economic risks, specifically regarding how agentic AI might replace human contractors and disrupt established business models.