Anthropic weakens its safety pledge in the wake of the Pentagon's pressure campaign (www.engadget.com)

Anthropic, the maker of the Claude AI chatbot, has announced a significant change to its safety protocols under pressure from the US Pentagon, which is pushing for unrestricted military access to its technology. The company has modified its Responsible Scaling Policy (RSP) to lower safety guardrails, abandoning its commitment to halt training new AI models unless specific safety measures are in place. Instead, Anthropic will now employ "Risk Reports" and "Frontier Safety Roadmaps" to offer public transparency, marking a shift from strict safety red lines to a more relative approach to safety. This change reflects Anthropic’s belief that without modifying its commitments, the company risks falling behind in the competitive AI landscape. This development holds major implications for the AI/ML community, as it raises concerns about the ethical responsibilities of AI companies amid growing military and commercial pressures. Critics warn that the decision to ease safety requirements could lead to a "frog-boiling" effect, where gradual safety compromises could lead to catastrophic risks. As competition intensifies, many worry that companies like Anthropic may prioritize rapid development over crucial safety measures, potentially jeopardizing societal well-being. The Pentagon's pressure and threats of invoking the Defense Production Act further highlight the precarious balance between innovation, ethical considerations, and national security in the rapidly evolving AI field.