Another Anthropic Employee Resigns, Warning That AI Companies Are 'Gambling with Our Lives'

The debate over artificial intelligence safety intensified after former Anthropic employee Jacob Coxon announced his resignation on X, warning that the industry is racing toward self-improving superintelligence without adequate control measures. Coxon, a 27-year-old AI researcher, stated that developers are essentially gambling with public safety as systems advance rapidly.

Coxon's departure follows a pattern of internal dissent within major AI firms. Anthropic itself was founded by former OpenAI employees who left Sam Altman's startup over safety concerns. The apprehension is widespread, with nearly 1,400 AI company employees signing an open letter in July urging the US government to regulate the technology and curb the unchecked acceleration of Big Tech.

Adding to the warnings, OpenAI Chief Scientist Jakub Pachocki noted that AI capabilities are progressing faster than researchers' ability to monitor and control them. Evan Hubinger, a more senior Anthropic employee, also posted on X supporting Coxon's statements, admitting a genuine belief that AI could pose existential risks to humanity over the next decade due to a lack of concrete containment plans.

In response, an Anthropic spokesperson emphasized the company's commitment to building models with strong safeguards while acknowledging both the benefits and unprecedented risks of AI. Anthropic CEO Dario Amodei has previously cautioned that the rush to develop advanced systems could lead to catastrophic errors resulting in a loss of control.

Despite mounting warnings from researchers, federal regulation remains stagnant. The Trump administration has worked against state-level AI restrictions, and Congress has been hesitant to impose strict rules, often citing competition with China. Meanwhile, leading AI companies—including OpenAI, Anthropic, Google, and Meta—have engaged with the White House on voluntary review frameworks, though details largely remain confidential.

Both Sam Altman and Anthropic have recently expressed openness to pacing the development and release of powerful models to ensure society has adequate time to adapt to new technological capabilities.