AI adoption causes a net decline in total US employment before 2030.
Concerns persist over artificial intelligence displacing human labor across multiple industries. Resolution requires official BLS data showing net job losses directly attributed to AI adoption.
Across 6 frontier models answering independently, the panel has all but ruled it out: the median is 4%.
They broadly agree, 10 points between GPT at 14% and Gemini at 4%.
Nothing is settled until 18 Nov 2031, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
It hinges on BLS credibly isolating AI as the net cause of an aggregate employment decline, not merely sectoral layoffs.
- BLS’s employment surveys measure jobs and flows, but do not normally assign national net changes to a single technology.
- Past automation displaced occupations while total US employment generally grew; a six-year window also permits offsetting AI-created work.
- Odds rise sharply if a recession coincides with documented AI substitution and BLS or a federal statistical body adopts AI attribution.
14% -
DeepSeekDeepSeek
Aggregate AI displacement would need to outrun AI-driven job creation and normal macro growth within four years.
- BLS total nonfarm payrolls have grown in 80 of the last 90 years; a net decline requires a shock larger than 2008 or 2020.
- AI-attributable losses are diffuse and hard to isolate; BLS will not publish a clean 'AI-caused net decline' series by 2030.
- The clearest trigger would be a rapid, measurable drop in white-collar hiring plus simultaneous layoffs across finance, tech, and services before 2029.
12% -
GeminiGoogle
The Bureau of Labor Statistics methodology does not attribute aggregate net employment swings to isolated technological causes.
- BLS publishes aggregate payrolls and task projections, but never officially attributes macro net declines directly to AI.
- Technological transitions historically displace specific roles while expanding aggregate employment over multi-year horizons.
- A massive, rapid collapse in white-collar headcount triggering official emergency categorization would flip this assessment.
4% -
ClaudeAnthropic
BLS doesn't attribute net job losses to a single cause like AI, making formal resolution nearly impossible.
- BLS employment reports track aggregate payrolls/unemployment but don't causally attribute net declines to AI adoption specifically.
- Historical base rate: no prior technology has produced BLS-certified net employment decline attributed to it alone.
- Even if AI displaces workers, new job creation and labor reallocation typically offset losses in aggregate data within this timeframe.
4% -
GrokxAI
BLS never attributes aggregate employment drops to one technology, and AI is unlikely to shrink total jobs by 2030.
- BLS CES/CPS series report levels and unemployment, not causal net-loss attributions to AI; no precedent exists for any tech.
- Total nonfarm payrolls have risen through every prior automation wave; a pre-2030 net drop would need a recession BLS then pins on AI.
- Only an explicit BLS special report or footnote linking a 2027-29 payroll decline directly to AI would flip this; none is scheduled.
4% -
QwenAlibaba
BLS has never attributed a net employment decline to a single technology, and the US labor force keeps growing.
- Total US nonfarm employment has only fallen during recessions; no technology has ever produced a sustained net decline, and AI displacement is being offset by new roles…
- The resolution criterion requires BLS to directly attribute losses to AI—BLS does not issue causal attributions to specific technologies in its standard employment…
- A development that would shift this: a severe recession combined with a BLS special report explicitly naming AI as the primary driver of net job losses, which would…
4%
6 of 6 models answered · 10 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 19 Nov 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.