A former Anthropic researcher told the BBC that staff working on artificial intelligence are "genuinely frightened" for humanity’s future, highlighting an uneasy mood inside the sector.

Industry insiders voice alarm

The researcher, who asked to remain anonymous, said the fear stems from rapid advances in large‑language models that are outpacing existing safety measures. According to the interview, engineers and scientists report sleepless nights worrying that the technology could be misused or evolve beyond human control.

Anthropic, founded in 2020 by former OpenAI executives, has long marketed itself as a safety‑first AI lab. Yet the internal unease described mirrors concerns voiced by other firms in recent months.

AI research lab with computer screens displaying code and model visualisations

Calls for a development pause

Anthropic’s chief executive, Dario Amodei, has publicly urged a slowdown in AI development, citing "serious" risks to society. In a statement released earlier this week, he warned that unchecked progress could amplify existential threats, echoing calls from OpenAI’s Sam Altman and other industry leaders.

The CEO’s appeal is part of a growing chorus urging policymakers to consider moratoria or tighter oversight. The researcher said that within Anthropic, the call for restraint has moved from abstract debate to a palpable sense of urgency among engineers.

Regulatory backdrop

Governments worldwide are responding to similar alarms. The European Union’s AI Act, slated for adoption later this year, seeks to classify high‑risk systems and impose strict conformity assessments. In the United States, the White House has convened an AI safety summit, inviting CEOs, academics and civil‑society groups to discuss risk mitigation.

These policy moves, however, are still in early stages. The researcher cautioned that without industry‑wide consensus on safety standards, “the gap between capability and control will only widen.”

Implications for the future

If the anxiety expressed by Anthropic staff proves widespread, it could reshape investment flows. Venture capitalists have already begun scrutinising funding rounds for firms that cannot demonstrate robust alignment protocols.

For the public, the revelations underscore why debates over AI ethics are moving from academic journals to boardrooms and legislative chambers. As the technology’s influence spreads—from medical diagnostics to automated decision‑making—“the stakes are no longer theoretical,” the former researcher warned.

Anthropic’s internal sentiment, combined with its CEO’s slowdown plea, may prompt other developers to adopt precautionary measures, potentially slowing the pace of breakthrough releases in the coming months.