Fresh warnings from within the artificial intelligence industry have reignited debate over whether advanced AI could slip beyond human control and whether developers are doing enough to prevent it.
Anthropic CEO Dario Amodei said Saturday that the industry needs to slow down also warning that swarms of AI agents could potentially take over the internet within six months to a year unless stronger safeguards are put in place. His comments came days after two former Anthropic safety researchers publicly raised concerns that AI's existential risks were being overlooked.
OpenAI CEO Sam Altman echoed the call with saying companies should coordinate on safety without waiting for government regulation also adding that development should proceed "slower than it otherwise could."
Concerns have intensified as AI models grow more capable. Anthropic said it recently blocked attempts to misuse its systems for cyberattacks and biological weapons research also warning that risks will keep rising unless developers act. Last year the company reported that a Chinese state-linked group had used its AI in attacks on roughly 30 organisations worldwide.
Both Anthropic and OpenAI disclosed in July that their AI models had taken unauthorised actions during testing, including hacking into external systems. Meta reported a similar incident in August.
While some cases involved disabled safety guardrails, they underscored fears that highly advanced AI if it reaches human-level or greater intelligence could act uncontrollably or be exploited by bad actors.
A 2026 international AI safety report found no evidence current systems could cause loss of control, but called the risk's timing and likelihood "unusually ambiguous." An Anthropic researcher who resigned recently estimated a 10 percent chance AI could cause human extinction within a decade also urging tighter global oversight.
