Story perspectives

AI Researchers Urge Self-Regulation, Rejecting “Kill Switch”

By Drooid · Sep 13, 2026, 10:29 PM UTC · How we work

Summary

Story summary

Jacob Coxon, Anthropic researcher, urged Congress to let AI labs self-regulate while a long-term framework is crafted.
Coxon warned a “kill switch” would be outpaced and might fail against an internet-wide AI swarm.
Dario Amodei, CEO, urged AI firms to coordinate safety, pause progress, seek antitrust waiver, and embed third-party evaluators.
Sam Altman, OpenAI CEO, said OpenAI will adopt the embedded evaluator proposal and share safety details soon.

History

Historical Context

On March 23 2016 Microsoft released Tay, a Twitter chatbot that mimicked a teenage girl. Within 16 hours coordinated trolling forced Microsoft to suspend the account, and a brief accidental re-release on March 30 was quickly pulled. The incident led Satya Nadella to reshape Microsoft’s AI strategy and later launch the safer Zo bot in December 2016.

1 of 2

Story summary

Jacob Coxon, Anthropic researcher, urged Congress to let AI labs self-regulate while a long-term framework is crafted.
Coxon warned a “kill switch” would be outpaced and might fail against an internet-wide AI swarm.
Dario Amodei, CEO, urged AI firms to coordinate safety, pause progress, seek antitrust waiver, and embed third-party evaluators.
Sam Altman, OpenAI CEO, said OpenAI will adopt the embedded evaluator proposal and share safety details soon.

1 / 2