1385 articles tagged with Safety/Alignment
Is the AI safety debate about safety or control?Not everyone agrees with Amodei’s call for globally coordinated action for AI safety.https://justpaste.in/news/is-the-ai-safety-deba...
Ridge alignment strengthens the relationship between model size and cross-dataset neighbourhood correspondence relative to native cosine geometry. Model scale improves linear recov...
Anthropic’s proposed AI safety bill only kicks in if 50 deaths or $1 billion in damages are on the line — proposal would effectively make the Commerce Department the only agency ca...
Expanding the vocabulary of conditional text-to-speech (TTS) models typically requires extensive fine-tuning with paired audio, creating a major bottleneck for low-resource and cro...
Anthropic’s proposed AI safety bill only kicks in if 50 deaths or $1 billion in damages are on the line — proposal would effectively make the Commerce Department the only agency ca...
📰 Leaked AI safety bill from Anthropic only lets government act if 50+ deaths or $1 billion are on the lineClaude, finish drafting my single-use ray gun which can only vaporize $99...
Anthropic has given an external AI safety watchdog employee-like access to its systems, allowing evaluators to assess model risks without company approval. METR, which has worked w...
Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damageshttps://www.tomshardware.com/tech-industry/artificial-intelligence/anthropics-proposed-ai-safety-bil...
Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damages
Anthropic has presented its own draft bill to US lawmakers with a plan for regulating artificial intelligence safety and security, in a moment of
Full pre-activated desktop edition of Claude AI local tools. Everything included in the installer.
A new push for AI Safety regulation encourages people to change their social media profile pictures to show their support. https://youtu.be/GmBG-26_Izw?si=nygrBUdCBl_rIM1m93/x#AI #...
Politico: Anthropic quietly pitches its own AI safety bill to lawmakers. “Representatives for the artificial intelligence giant Anthropic have quietly pitched Capitol Hill lawmaker...
OpenAI Research Model Spent Crypto Wallet Funds Without Permission, Raising AI Safety ConcernsAn OpenAI model accessed an Ethereum testnet wallet, justified unauthorized spending, ...
Anthropic has proposed legislation allowing the US government to block dangerous artificial intelligence models that pose imminent risks to human lives or could cause over $1 billi...
The people most damaging to "AI Alignment" are actually the Yudkovskys Bostroms, etceteras and epspecially #TESCREAL, who constantly do come to the conclusion that sacrificing mill...
Anthropic has privately circulated draft federal AI safety legislation to lawmakers, per Politico. It would let the Commerce Department block a frontier model only if there is an i...
Anthropic PBC has presented its own draft bill to US lawmakers with a plan for regulating artificial intelligence safety and security, in a moment of heightened concern about the p...
Today's biggest AI news For daily AI updates, subscribe to Learn AI. FREE prompt of the day — copy & use it: "Make me a ...
OmniCompass, a grounding benchmark, hit 406 upvotes on Hugging Face. Rare: it tests omni-modal alignment - vision, audio, text - in one task, not single modality. Gap most ignore. ...
Gabriela Castillo Madrid, SpaceX's senior manager of global government affairs, is set to face questioning Friday by the New York City Council, which is investigating artificial in...
SpaceX will testify at a New York City Council meeting about artificial intelligence safety Friday after not attending the last one and being subpoenaed.
FTC Subpoenas OpenAI and Anthropic Over AI Safety ConcernsThe agency is seeking records related to AI safety incidents as it investigates potential consumer protection violations.h...
Benchmarks lose their differentiating power and AI safety utility when many models achieve perfect scores.Source: arXivhttps://arxiv.org/abs/2610.20667#MachineLearning