How would AI actually ‘kill all humans’? Here are the top five most likely scenarios | Toby Walsh
From a dangerous new bioweapon to total societal breakdown, is a ‘superintelligent’ AI capable of wiping out humanity? Earlier this month, artificial intelligence researcher Jacob Coxon resigned from Anthropic after just four months. In an announcement on X, he stated : “The people building AI earnestly believe that it could kill us all by the end of the decade.” A senior member of Anthropic’s staff, Evan Hubinger, actually agreed with Coxon, adding he personally thinks the chance of this happening in the next decade is more than 10% . Continue reading...
Every model that read this
| Model | Provider | Stage | Score | Conf. | Latency | Prompt | When |
|---|---|---|---|---|---|---|---|
| Llama 3.3 70B | Meta | analysis | -50 | 70% | 4908ms | v1.0.0 / m1.0.1 | 2026-09-23 03:17 |
| GPT-4.1 mini | OpenAI | consensus | -70 | 85% | 3873ms | v1.0.0 / m1.0.1 | 2026-09-23 03:17 |
| Claude Sonnet 5 | Anthropic | consensus | -25 | 35% | 5072ms | v1.0.0 / m1.0.1 | 2026-09-23 03:18 |
| Gemini 2.5 Flash | consensus | -100 | 80% | 1627ms | v1.0.0 / m1.0.1 | 2026-09-23 03:18 | |
| Mistral Small 3.1 24B | Mistral AI | consensus | -70 | 80% | 5096ms | v1.0.0 / m1.0.1 | 2026-09-23 03:18 |
Speculative claims about AI risk
The claim that prominent AI researchers believe superintelligent AI could kill all humans within a decade points to substantial existential risk. The cited probability above 10% by a senior Anthropic staffer gives weight to this risk, although it remains uncertain. This implies a significant negative implication for humanity's survival from AI development risks within a medium-term horizon.
The piece is an opinion column speculating on extinction scenarios, citing one resignation and one researcher's subjective probability estimate rather than empirical evidence or incidents. The underlying claims are unverifiable predictions, so while the implication is extreme if true, evidentiary grounding is weak.
The article explicitly discusses scenarios where 'superintelligent' AI could 'wipe out humanity,' citing a researcher's belief that AI could 'kill us all by the end of the decade' with a 10% chance. This represents an extreme dystopian implication for human flourishing, despite being speculative.
The article discusses scenarios where AI could lead to human extinction. The claims are speculative and based on the beliefs of AI researchers. The scenarios are extreme and imply a high risk to humanity.
Evidence extracted
- Jacob Coxon believes AI could kill all humans by the end of the decade
- Evan Hubinger thinks the chance of AI killing all humans in the next decade is more than 10%
SOURCE The Guardian: Artificial Intelligence (tier 1)
↓
DOCUMENT 8e922fa4-dc10-474b-8443-70afb1eece73
https://theguardian.com/australia-news/commentisfree/2026/sep/23/how-would-ai-actually-kill-all-humans-here-are-the-top-five-most-likely-scenarios
↓
EVIDENCE 2 extracted excerpts
↓
MODEL RUN 5 runs, methodology 1.0.1
↓
SCORE -63 (Adverse)
↓
CONFIDENCE 70%