AI Researcher Departs Over Fears of Uncontrollable Systems
AI-generated context summary requested by a Free News Reader user. Sourced via Gemini from publicly available information — no paywalled content was accessed.
You hit a paywall. Here’s the context on this topic based on publicly available information. We did not access any paywalled content. View original article.
AI Researcher Departs Over Fears of Uncontrollable Systems
- A researcher from Anthropic, Jacob Coxon, has resigned from the AI industry due to concerns that companies are rapidly developing AI systems that could become uncontrollable.
- Coxon's departure in September 2026 highlights growing safety concerns within leading AI companies regarding the race toward self-improving models.
Full Summary — powered by AI
Jacob Coxon, a researcher specializing in training AI models, announced his departure from Anthropic and the artificial intelligence industry in September 2026. His decision stems from mounting fears that AI labs are engaging in a rapid competition to build systems that could spiral out of human control.
This concern is not isolated, as there has been increasing apprehension within AI companies that competitive pressures are accelerating the development of self-improving models with inherent risks. Experts have warned that an “intelligence explosion,” a rapid cycle of AI self-improvement, could outpace human oversight and make it impossible to implement necessary safety measures.
The departure of researchers like Coxon underscores a broader trend of AI safety concerns. In February 2026, another Anthropic staffer, Mrinank Sharma, who led a safeguards research team, also resigned, issuing a warning about a world “in peril.” Similarly, a former OpenAI economic researcher, Tom Cunningham, left his position in September, reportedly frustrated with the company’s alleged hesitancy to publish research critical of AI usage. In 2024, OpenAI also dissolved its Superalignment safety research team after two key members resigned.
The Future of Life Institute’s Summer 2026 AI Safety Index noted that major AI companies have weakened key safety commitments despite their models becoming more powerful. This report suggested that the voluntary safety framework established by AI labs is eroding before governments can implement a robust alternative. The index also indicated that leading companies like Anthropic, OpenAI, Google DeepMind, and Meta have reversed earlier pledges to pause development if certain safety “redlines” were approached, with some citing competitor actions as a reason.