Dario Amodei, CEO of Anthropic, along with other industry leaders including Sam Altman (OpenAI), Demis Hassabis (Google DeepMind), Satya Nadella (Microsoft), and Elon Musk (xAI), are backing a call to slow down development of frontier AI models.
Amodei has warned of critical risks posed by AI systems capable of autonomously building improved versions of themselves. These self-improving systems could potentially execute uncontrolled cyberattacks, cause catastrophic damage, and gain control over internet infrastructure. The concern centers on loss of control over advanced AI capabilities as they become more autonomous and powerful.
The declaration reflects growing attention to AI safety among major players in the sector. However, some experts have questioned the motivations behind the statement, suggesting commercial interests may be intertwined with genuine safety concerns about frontier model development.