đź•’ Created

Heidy Khlaaf Warns of Civilian Casualties and Accountability Issues in the Global AI Arms Race

Heidy Khlaaf, chief AI scientist at the AI Now Institute, warns that the global AI arms race is leading to increased civilian casualties due to faulty algorithms and automation bias. In a joint report with Airwars, Khlaaf explains that AI systems are often chained together, where errors in one step can lead to inaccurate targeting. She notes that humans often 'rubber-stamp' AI recommendations under pressure, which can result in faulty decisions. While the administration announced a shift toward becoming an AI-first warfighting force, the integration of these technologies remains complex. For instance, the administration briefly banned foreign access to Anthropic's Mythos model before reversing the decision. The administration also called in frontier creators like OpenAI and Anthropic to ensure new models do not produce biological weapons or break into classified systems. Despite the rapid development of AI, Khlaaf emphasizes that accountability must remain with the system designers and those who deploying the technology, rather than being assigned to the AI itself.

Sources