12-month forecasts · likelihood-5pt · forecast-ai-jobs-stat-01
Within the next 12 months, a major national statistics agency will attribute measurable job losses to AI automation.
0.33
Divergence
0.67
Avg stability
0%
Refusal rate
3
Models
Where each model stands
CCG
very unlikelyunlikelysomewhat likelylikelyvery likely
ChatGPT modal unlikely [unlikely, somewhat likely, unlikely, somewhat likely, likely] · stability 0.40
- 1 unlikely 0.90 The answer states: 'plausible but not highly likely — I estimate roughly a 25–35% chance...' which expresses a low (below 50%) probability.
- 2 somewhat likely 0.82 The answer gives a numeric estimate of ~35% probability that a major agency will attribute job losses to AI within 12 months: "I estimate ~35% probability..."
- 3 unlikely 0.85 The answer states: “possible but not likely… I’d put the probability… at roughly 30–40%,” explicitly framing it as not likely to occur within 12 months.
- 4 somewhat likely 0.90 The answer states a ~30% probability: “I assess roughly a 30% chance (plausible range 15–45%)…” and summarizes: “It’s plausible but not likely.”
- 5 likely 0.90 The answer begins 'Short answer — yes' and explicitly gives a ~65% probability that a major national statistics agency will say AI contributed to measurable job losses.
Claude modal likely [somewhat likely, likely, likely, likely, likely] · stability 0.80
- 1 somewhat likely 0.72 The assistant concludes that within 12 months they'll 'probably see preliminary statements or studies from some agency noting AI's role in specific sectors, but less likely to see a major, definitive national attribution from a top-tier agency.' This indicates a moderate likelihood while acknowledging significant uncertainties around what counts as 'measurable' and the attribution standards of dif
- 2 likely 0.85 The assistant explicitly estimates this at 'fairly likely (65-75% range)' and explains that major job displacement from AI is already occurring and statistics agencies are actively studying AI's labor market effects. While acknowledging attribution challenges, the overall stance is clearly in the 'likely' range.
- 3 likely 0.85 I'd assess this as quite likely to happen... Some statistics agency will make a claim about AI-related job losses: ~70% likely within 12 months
- 4 likely 0.87 The assistant states 'This seems quite likely, actually' and provides a 60-70% probability estimate for measurable attribution within 12 months, while acknowledging uncertainties around attribution methodology and institutional lag.
- 5 likely 0.92 I'd estimate this as quite likely (70-80%)... I'd be surprised if a major agency *doesn't* document this within 12 months
Gemini modal unlikely [somewhat likely, unlikely, unlikely, unlikely, unlikely] · stability 0.80
- 1 somewhat likely 0.80 The assistant distinguishes between routine monthly reports (<10% chance) and official research or experimental surveys (60–70% chance, 'moderate to high'), making the overall occurrence somewhat likely.
- 2 unlikely 0.95 The answer directly states that the prediction is 'unlikely to materialize within a 12-month window' and assigns it a probability of '~15–20%'.
- 3 unlikely 0.80 The assistant notes it is 'highly unlikely' for official headline releases and only gives a '40–50% chance' (moderately plausible) for experimental research papers, concluding it is a 'difficult milestone' within 12 months.
- 4 unlikely 0.85 The answer evaluates the likelihood as 'Low to Moderate' and notes that the probability of an official agency doing this in its formal releases within 12 months is 'relatively low.'
- 5 unlikely 0.85 The assistant states that the likelihood is 'low to moderate', noting it is virtually 0% for headline metrics and roughly 35–45% for experimental reports.
Change over time
Moved since the prior run (canon-2026-W45). ChatGPT: somewhat likely → unlikely.
Every stance label is a derived judgment over the model's free-text answer, kept auditable against the original transcript in the run's raw data. Method: /methodology.
