Police hugging AI so tight, soon we'll all be pre-crime suspects (responsiblestatecraft.org)

A recent announcement from Anthropic's researcher Evan Hubinger alarmingly suggested that "Superintelligent AI" could pose a significant risk to humanity, sparking concerns about future AI capabilities. However, more pressing issues revolve around the current integration of AI technologies within the U.S. military and intelligence frameworks. Anthropic, amidst a contentious relationship with the Department of Defense, highlighted its AI model Claude's role in military targeting systems and offensive cyber operations, revealing a troubling connection between AI firms and state surveillance practices. The implications of this partnership raise ethical considerations about privacy, surveillance, and pre-crime policing. The Department of Homeland Security has solicited AI tools capable of predictive modeling for identifying potential domestic threats, expanding its surveillance capabilities considerably. There are concerns regarding how vast amounts of data collected on citizens could be misused, especially in light of previous practices like the Trailblazer program and the controversial intelligence assessments linked to political beliefs. This growing reliance on AI technologies such as Claude and Palantir’s tools in state operations underscores a critical juncture for the AI/ML community, where the balance between innovation and ethical responsibility must be carefully navigated.