https://www.darius.wiki/en/topics/ai-policy/Regulation, content provenance and watermarking, and what they require of productsenFri, 09 Oct 2026 00:00:00 GMThttps://www.darius.wiki/en/daily/2026-10-09/anthropic-usage-policy-2026/https://www.darius.wiki/en/daily/2026-10-09/anthropic-usage-policy-2026/Fri, 09 Oct 2026 00:00:00 GMTSafetySeveral of the changes directly affect anyone building on Claude: letting the model drive hardware, making recommendations in high-risk domains, and company ownership. Take: AI Interview falls squarely in the employment category, so before November 12 I'll make sure of two things: the results page says which assessments come from AI, and the HR side keeps a human review and override path. Teams with offshore entities or foreign investors should also check whether their ownership structure runs into the regions clause.https://www.darius.wiki/en/daily/2026-10-09/openai-false-front-operations/https://www.darius.wiki/en/daily/2026-10-09/openai-false-front-operations/Fri, 09 Oct 2026 00:00:00 GMTSafetyBoth operations used fabricated media and research identities, with AI producing the content, to push geopolitical narratives. Take: Read alongside Anthropic's new deceptive-activity rules the same day, both companies are targeting fabricated identities and content at scale. Multi-account content tools like PandaClaws should take note: keep account identities real and traceable, and don't let generated personas look like they're posing as media outlets or institutions.https://www.darius.wiki/en/daily/2026-10-06/openai-eu-text-provenance/https://www.darius.wiki/en/daily/2026-10-06/openai-eu-text-provenance/Tue, 06 Oct 2026 00:00:00 GMTSafetyEU provenance rules for AI-generated content are reaching text; OpenAI explains where watermarks apply, how detection works and who gets detection first. Take: All I could read was the official RSS summary, so the details will have to wait; the direction is clear, though: for products serving EU users, generated text may carry a detectable mark. A content-generation backend like PandaClaws should start recording which content is model-generated and by which model in its metadata now, so compliance doesn't mean rework later.