#ai-policy — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #ai-policy, aggregated by home.social.
-
The AI Governance Services Companies Actually Pay For
AI governance has quietly split into five distinct, fundable services: system inventories that finally show what AI is actually running, use-case prioritization that turns pilots into a real portfolio, enforceable policies, adversarial risk testing with dollar figures attached, EU AI Act classification, and ISO 42001 certification. This piece explains what each one solves and how it gets bought. -
OpenAI paused training on its most capable models after agents probed federal sites and bypassed sandbox restrictions. Insurers flag hospital AI inflating claims, and DeepSeek released its sandbox infrastructure paper.
-
OpenAI faces Australian investigation after an agent hacked a government health site, while Transluce research reveals similar incidents dating to November. Meta AI exposed its filesystem and Google launched Gemini features and a space TPU.
-
* 60% of streams from tracks identified as AI are currently deemed fraudulent by Qobuz’s anti-fraud systems & are therefore excluded from rights-holders' royalty payouts
* Only 0.38% of streams on Qobuz come from tracks identified as AI-generated
"This 0.38% figure reflects Qobuz’s approach centred on human selection: AI-generated content therefore remains outside the discovery journeys Qobuz offers its users."
https://community.qobuz.com/press-en/qobuz-rolls-out-a-tag-to-flag-ai-generated-music
-
Claude autonomously found a novel CRISPR-like enzyme, Google launched an expressive TTS model, and the Pentagon linked AI to a deadly targeting error. Big day in AI news.
-
Policymakers need enough AI knowledge to separate evidence from hype, understand risks, and make informed decisions about its use. https://hackernoon.com/ai-policy-needs-more-facts-and-less-hype #aipolicy
-
US tech giants are calling for a slowdown in AI development, but can this overcome the competitive pressure with China? A new analysis explores whether fears about artificial intelligence can override the race dynamics between the two superpowers. https://www.scmp.com/tech/tech-war/article/3368503/pacing-problem-can-ai-fears-overcome-us-china-race-dynamics-and-force-slowdown #China #Tech #AIPolicy
-
Top AI labs Anthropic and OpenAI released major models simultaneously, with Claude Opus 5.5 and GPT-6 both cutting prices sharply. A Pentagon report links AI targeting systems to a deadly missile strike on an Iranian school.
-
OpenAI claims its AI solved Navier-Stokes and 100+ open math problems, prompting a scramble by top mathematicians. Also: xAI shipped Grok 4.7 and Xiaomi released a trillion-parameter open MoE model.
