AI Safety Researchers Voice Extinction-Level Concerns in New Interview Series
A nonprofit research organization called Palisade Research has published a series of video interviews with AI researchers currently and formerly employed at major AI labs, including OpenAI, Google DeepMind, and Anthropic. The interviews, hosted on the platform frominside.ai, feature roughly a dozen researchers discussing the potential risks posed by advanced artificial intelligence.
Among the contributors, Geoffrey Irving—a former employee at both OpenAI and Google DeepMind—stated that he estimates roughly a 50 percent probability that AI could drive human extinction. Neel Nanda, a current research scientist at Google DeepMind, offered a more conservative but still significant estimate, suggesting at least a 10 percent chance that advanced AI could cause human extinction.
The researchers' concerns center on the development of superintelligent systems—AI that surpasses human cognitive capabilities across most domains. Several interviewees emphasized that such systems could behave in unpredictable ways if not properly aligned with human values and intentions.
This public discussion of existential risk occurs against a backdrop of ongoing debate within the AI community about how to balance rapid advancement with safety considerations. While some labs have published safety frameworks and committed to responsible development practices, critics argue that the pace of deployment continues to outstrip the development of robust safeguards.
Palisade Research describes itself as a nonprofit focused on studying both the capabilities and motivations of AI systems. The organization said it compiled the interviews to bring transparency to how researchers at leading companies perceive the risks associated with their work.