In a dramatic escalation of concerns about artificial intelligence safety, a group of DeepMind researchers are planning a hunger strike for January 2025, demanding stronger safeguards against potential AGI (Artificial General Intelligence) risks. The researchers, who work at Google’s prestigious AI lab, are specifically calling for DeepMind CEO Demis Hassabis to implement more robust safety measures before pursuing advanced AI systems that could match or exceed human capabilities. This unprecedented action highlights the growing divide between AI development timelines and safety protocols within even the most respected research organizations.

The planned protest comes amid accelerating AI development across the tech industry, with DeepMind and other leading labs racing toward increasingly powerful models. The researchers’ concerns center on what they describe as inadequate safety measures for systems that could potentially reach AGI-level capabilities within the next few years. Their demands include implementing specific technical safeguards, establishing independent oversight mechanisms, and potentially slowing development until safety can be assured. While DeepMind has publicly committed to responsible AI development, these researchers believe current measures fall short of addressing the existential risks that advanced AI might pose.

This internal conflict at one of AI’s most influential organizations reflects broader tensions in the field between rapid innovation and responsible development. As AI capabilities continue to advance at a remarkable pace, the question of how to balance progress with safety has become increasingly urgent. The planned hunger strike represents perhaps the most dramatic manifestation yet of concerns that have been growing within the AI research community for years. How DeepMind and the broader industry respond to these escalating calls for caution could significantly shape the future trajectory of artificial intelligence development and governance.

Source: https://www.businessinsider.com/hunger-strike-deepmind-ai-threat-fears-agi-demis-hassabis-2025-9