AI leaders Dr. Anya Sharma and Ben Carter discuss the primary risks keeping them up at night. Dr. Sharma emphasizes the existential threat of misaligned AGI, while Ben Carter focuses on the more immediate dangers of misuse and data poisoning, particularly with open-source models.

Anthropic CEO Dario Amade and DeepMind CEO Demis Hassabis debate the timeline for AGI, with Amade predicting 2026-2027 and Hassabis targeting the end of the decade. Both express grave concerns about safety, geopolitical risks, and the inadequacy of current global governance.

The CEOs of Anthropic and DeepMind discuss AGI timelines, geopolitical risks from authoritarian states, and the need for international governance, while balancing the immense potential benefits in science and medicine.