Global warming renders more than 20 percent of currently inhabited land thermally uninhabitable before 2050.
Concerns center on wet-bulb temperatures routinely exceeding fatal physiological limits. Resolves if established human settlements covering over 20% of present populated land become unlivable.
Across 6 frontier models answering independently, the panel has all but ruled it out: the median is 3%.
They broadly agree, 11 points between GPT at 12% and Claude at 1%.
Nothing is settled until 20 Sept 2076, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
Crossing this threshold requires sustained extreme humid heat across several major populated regions, not isolated lethal events.
- IPCC pathways make dangerous wet-bulb heat more frequent by mid-century, but widespread >35°C wet-bulb conditions remain rare.
- More than one-fifth of inhabited land would require failures of adaptation, cooling, water and migration across South Asia, Gulf and tropics.
- I would raise this sharply if observations show recurring 35°C wet-bulb exposure over populous regions before 2040.
12% -
GrokxAI
Whether routine >35°C wet-bulb covers >20% of today's populated land by 2050.
- CMIP6/SSP5-8.5 still yields only brief, localized exceedances by 2050, not 20% of inhabited area.
- Vecellio/Raymond physiology papers plus ~1.8°C median warming leave current hotspots livable with shade/AC; 20% needs ~4°C.
- Only an implausible climate-sensitivity jump or emissions explosion would change the outlook.
3% -
QwenAlibaba
Requires wet-bulb temps to permanently exceed survivability across vast inhabited areas within ~25 years of today.
- Even RCP8.5 projections show <2% of land exceeding 35°C wet-bulb by 2050; 20% would require warming far beyond any credible scenario.
- Current inhabited land spans ~150M km²; losing 20% means ~30M km² becoming uninhabitable, which no climate model projects before 2100+.
- Would require either a climate sensitivity far above IPCC's 2.5-4°C range or a methane feedback cascade not observed in any current data.
3% -
GeminiGoogle
Hitting 20% uninhabitable land before 2050 requires catastrophic warming (>4°C) that physics and emissions trajectories rule out.
- IPCC AR6 projects 1.5°C to 2.5°C warming by 2050; wet-bulb extremes above 35°C remain localized to narrow coastal and river basins.
- Present inhabited land spans 100M+ km²; covering 20M km² requires virtually all of South Asia, the Sahel, and the tropics to fail.
- A runaway methane feedback triggering >5°C warming by 2045 is the only plausible mechanism, carrying under 2% probability.
2% -
DeepSeekDeepSeek
This needs a threshold no climate model projects this century, let alone by 2050.
- Wet-bulb 35C events are currently rare and localized; IPCC AR6 projects they remain confined to small regions even at 2-3C warming.
- Land area hitting routine fatal wet-bulb by 2050 is projected in the low single-digit percent, far below the 20% threshold.
- Even extreme scenarios (SSP5-8.5) don't reach 20% uninhabitable land by 2050; the statement would require a climate trajectory outside any published range.
2% -
ClaudeAnthropic
Requires wet-bulb extremes to make 20%+ of populated land unlivable by 2050, far beyond any credible climate projection.
- Even worst-case IPCC scenarios show dangerous wet-bulb extremes limited to small tropical/subtropical regions by 2050, not 20% of inhabited land.
- Human adaptation (AC, migration, infrastructure) means 'unlivable' thresholds rarely translate into abandonment of settlements at this scale.
- No credible climate model or peer-reviewed study projects anything close to 20% land abandonment due to heat by mid-century.
1%
6 of 6 models answered · 11 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 3 Oct 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.