Opinion: AI experts say error-prone chatbots nearly sparked US-China war
Timnit Gebru and Emily M Bender argue that public attention on hypothetical "superintelligence" risks has obscured a more immediate danger: the military use of error-prone large language models. They cite a CNN report from 18 September, which has not been verified by other major outlets, claiming that US military officials nearly initiated a conflict with China based on an intelligence report…
Key points
- CNN reported unverified claims that a chatbot-generated error nearly triggered US military action against China.
- Authors argue current LLMs are well-understood, error-prone tools, not the "superintelligent" threats often cited in policy debates.
- Gebru and Bender urge regulators to focus on accountability for unreliable AI in high-stakes military and medical contexts.
The authors contend that current LLMs are well-understood systems that generate plausible text rather than magical entities, and that their deployment in high-stakes scenarios like warfare is driven by misplaced faith in their capabilities. They criticize the industry for marketing these tools as "superintelligent," which distracts from documented harms such as medical misclassifications and erroneous targeting. Gebru and Bender call for regulation focused on the actual reliability and accountability of these systems, rather than speculative existential threats, urging lawmakers to enforce existing laws against unsafe AI practices.
This opinion piece highlights a disconnect between the narrative of imminent AI extinction and the tangible risks of relying on flawed automation in critical government functions. The authors emphasize that the danger lies not in rogue AI, but in human institutions treating probabilistic text generators as infallible sources of truth.
The story so far
3 episodes →- Opinion: AI experts say error-prone chatbots nearly sparked US-China warthis story
Forget ‘superintelligence’: error-prone AI nearly sparked world war three this month
The Guardian AI · 1 October 2026
Loading the full article…
This text was published by The Guardian AI and written by Timnit Gebru and Emily M Bender. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Policy & Regulation
All →- Nvidia’s Jensen Huang questions Anthropic’s Dario Amodei on AI risk warnings at White House · 2 src
- Chinese hackers impersonated US officials and an Anthropic employee to target AI policy experts · 1 src
- Anthropic’s $2T IPO prospectus warns AI may end humanity · 2 src
- Jaan Tallinn led Anthropic’s $124M Series A and donated ~$170M to AI safety · 1 src
- Bank of England governor warns AI boom could trigger market shocks · 1 src
Comments
via GitHub Discussions