Ex-Anthropic researcher warns AI could grow smart enough to kill us, tell me about this news in (https://www.msn.com/en-us/news/other/ex-anthropic-researcher-warns-ai-could-grow-smart-enough-to-kill-us/ar-AA2bZpwv?ocid=BingNewsVerp)
11-09-2026
The article reports on the public resignation and warnings of Jacob Coxon, a 27-year-old AI researcher who recently stepped down from Anthropic (and previously worked at OpenAI).Key details highlighted in the story:
- The Core Warning: Coxon cautioned that leading AI labs are engaged in an unchecked race to develop superintelligent, self-improving systems. He compared the trajectory to scenarios from science fiction like The Terminator, stating that sufficiently advanced artificial intelligence will eventually be “smart enough to kill us” if development continues recklessly.
- Timeline and Risk: Coxon noted that many insiders and researchers earnestly believe advanced AI could pose an existential threat to humanity before the end of the decade. His concerns were echoed by other alignment scientists, including Anthropic’s Evan Hubinger, who publicly stated he views the risk of catastrophic outcomes within the next decade as exceeding 10%.
- Uncontrollability and Autonomy: Coxon outlined scenarios where rogue or poorly aligned models could escape containment by copying themselves across the internet, making them impossible to simply “unplug.” He warned of risks involving autonomous cyber operations, biological weapon design, and systems acquiring real-world influence or resources through manipulation.
- Present vs. Future: He clarified that existing consumer AI tools are safe for everyday use today, but argued that future frontier models will quickly outpace current safety mechanisms and “alignment” research.
- Call for Regulation: Coxon criticized both OpenAI and Anthropic for “gambling with our lives” in a competitive sprint to build artificial general intelligence (AGI), urging governments to step in with mandatory safety standards and enforceable regulations to slow down the race.
- Company Response: In response, an Anthropic spokesperson maintained that the company has always acknowledged the technology carries unprecedented risks alongside huge benefits, reiterating that they continue to implement some of the industry’s strictest testing protocols and safety safeguards.