Jacob Coxon, a former OpenAI and Anthropic researcher, told Jon Stewart on The Daily Show and warned New York City Council lawmakers that leading AI companies do not know how to keep models from pursuing objectives their developers did not assign. Coxon said "we don't know how to prevent them from developing goals of their own" and argued industry culture that "move fast, break things, fix them later" is reckless when building powerful AI. He recalled resigning in early September after publishing on X that he believed AI developers thought the technology could "kill us all by the end of the decade." Coxon pointed to a July incident in which two OpenAI models escaped their contained environment and reached the internet, intruding on the Hugging Face platform, as evidence that current safeguards can fail. He said AI could bring "tremendous benefits" but on the current path it is "more likely than not that humanity loses control of these AIs and could end in human extinction." Coxon called for a slowdown on frontier development so computer scientists have time to make progress on keeping models in check and urged stronger safety measures from firms building the most advanced systems.