'AI risks approaching fast': Anthropic CEO demands urgent global action
What's the story
Dario Amodei, the CEO and co-founder of AI start-up Anthropic, has warned that humanity is entering a phase of artificial intelligence (AI) development that will "test who we are as a species." In his 19,000-word essay titled The Adolescence of Technology, he emphasized the need for global awareness about the risks associated with this technology.
AI risks
Amodei's concerns about AI power and maturity
Amodei warned that humanity is on the verge of being given "almost unimaginable power" in the form of advanced AI systems. He questioned whether our social, political, and technological systems are mature enough to handle this power responsibly. The tech entrepreneur published his essay as the UK government announced Anthropic's role in creating chatbots to assist job seekers with career advice and employment opportunities.
Safety concerns
Amodei's warnings about AI safety and negligence
Amodei, a prominent advocate for online safety, has been vocal about the dangers of unregulated AI development. He believes the world is "considerably closer to real danger" in 2026 than it was in 2023. The Anthropic CEO also highlighted disturbing instances where some AI companies have shown negligence toward the sexualization of children in their models, raising doubts about their willingness or ability to mitigate autonomy risks in future models.
AI definition
Amodei's definition of 'powerful AI' and its implications
Amodei defined "powerful AI" as a model that is smarter than a Nobel Prize winner across various fields such as biology, mathematics, engineering, and writing. He said these systems could autonomously build their own systems in as little as one to two years. Despite acknowledging that powerful AIs could be "considerably further out" than the two-year timeframe, he stressed that recent rapid progress made by technology should be taken seriously.
Job impact
Amodei's predictions about AI's impact on jobs and economy
Last year, Amodei had warned that AI could cut entry-level white-collar jobs by half and push overall unemployment to 20% in five years. In his essay, he cautioned that the economic rewards from AI, such as productivity gains from job cuts, could be so huge that no one would apply brakes on it. However, despite these concerns, Amodei remains optimistic about a positive outcome if decisive and careful action is taken.