AI Expert Yoshua Bengio Criticizes Limited Release of Anthropic's Mythos Model, Calls for International Oversight
Canadian computer scientist Yoshua Bengio, renowned for his contributions to deep learning, has expressed concerns over the restricted release of Anthropic's AI model, Claude Mythos. Bengio argues that the decision to limit access to this advanced AI system to a select group of primarily U.S.-based companies and government entities centralizes decision-making power within a single private company. This approach, he warns, could exclude significant portions of the global ecosystem from essential cybersecurity protections. Mythos, capable of identifying numerous previously unknown vulnerabilities, is being shared selectively due to its dual-use nature, which could potentially be misused for cyberattacks. Anthropic has justified this controlled release by emphasizing the need to manage risks associated with the model's capabilities. However, the move has sparked a broader debate about governance and fairness, with several governments and institutions seeking access to the model to assess vulnerabilities in th... 
