ChatGPT has assigned a 25% probability that the United States will enter a recession within the next six months. The OpenAI model assessed that while risks remain present, a downturn is not the most likely outcome under current conditions.
The estimate comes amid ongoing debate among investors and economists over the trajectory of growth, employment, inflation, and interest rates.
The Drivers of This Development
In its assessment, ChatGPT placed the highest probability—55%—on a scenario in which the U.S. economy avoids recession and continues expanding at a slower pace through early 2027. It assigned a 20% chance to a soft landing with stronger growth, a 20% probability of a mild recession, and a 5% likelihood of a moderate or severe downturn. The model noted that recession risks could rise if unemployment climbs above 4.5%, payroll growth weakens, retail sales deteriorate, corporate defaults increase, or energy prices stay elevated. Recent data, including August’s addition of 162,000 jobs and a steady 4.1% unemployment rate, have so far avoided the sharp labor-market deterioration that often precedes downturns. For perspective, the 25% six-month probability sits above historical baseline levels in normal times but remains well below thresholds typically associated with an imminent recession.
It is important to note the fundamental difference between an AI-generated probabilistic assessment based on available economic indicators and formal forecasts from central banks or major institutions: the former reflects pattern recognition across data and narratives, while the latter often incorporates proprietary models, policy assumptions, and judgment calls.
Impact and Broader Context
The assessment adds another data point to the mixed signals surrounding the U.S. economic outlook. Strong recent employment figures have helped some institutions lower their recession probabilities, yet lingering concerns about consumer resilience, interest-rate effects, and external shocks keep uncertainty elevated. ChatGPT’s breakdown highlights that the base case remains continued, if slower, expansion rather than contraction. Markets and policymakers will continue to watch labor-market trends, spending data, and inflation closely for confirmation or shifts in the trajectory.
This development sparks important discussions about the role of large language models in economic forecasting and public interpretation of risk. Supporters see AI assessments as useful, rapidly updatable syntheses of complex information that can complement traditional analysis. Critics caution that such models can over- or under-weight recent data, lack true causal understanding, and should not be treated as authoritative predictions. Analysts observe that the relatively moderate 25% figure aligns with a broader recent decline in some institutional recession odds, even as vigilance remains warranted given elevated uncertainty.
Looking ahead, incoming economic reports and any changes in monetary policy expectations will test whether the current baseline of slower growth holds. This analysis is based on ChatGPT’s reported probabilistic assessment and contemporaneous economic data for accuracy and reliability. Recession probabilities remain subject to rapid revision as new information emerges.