An AI researcher, Jacob Coxon, resigns from Anthropic PBC and publicly calls on colleagues to rethink their work, saying the race to build advanced AI systems is unsafe. In his departure statement, he warns that the pursuit of potentially self-improving technology creates existential risks, using the phrase “gambling with our lives.” Several outlets report that he also criticizes the broader industry trend toward faster, more powerful deployments.
Across reports, Coxon is described as having previously worked at OpenAI before joining Anthropic, and he has spent years involved in pretraining AI models, a stage where systems learn from large datasets. While Financial Times and others frame the resignation as part of growing internal unease at AI labs, TechCrunch and Wired add more context about the “alignment” and safety discussions inside Anthropic and the pace of progress.
Outlets differ mainly in emphasis: some focus on the personal departure and the warning to other staffers, while others highlight Silicon Valley’s wider competitive pressures. Bloomberg and the South China Morning Post stress that Coxon directs concerns not only at Anthropic, but also at OpenAI, arguing both companies focus too heavily on beating rivals rather than safety.
