Llama Guard is not a guardrail.
We put Granite Guardian and Llama Guard side by side for regulated Brazilian Portuguese. One worked...
#Safety/Alignment
1400 articles tagged with Safety/Alignment
OpenAI cancels GPT-6.1 Astra launch, planned for October, after failing safety and alignment tests—rare move for a major...
OpenAI cancels GPT-6.1 Astra launch, planned for October, after failing safety and alignment tests—rare move for a major AI firm. #AI #OpenAI https://thehackernews.com/2026/09/open...
Engineering isn't just about building; it's about managing the consequences of over-building. From AI "alignment taxes" ...
Engineering isn't just about building; it's about managing the consequences of over-building. From AI "alignment taxes" to infrastructure lags, we're analyzing the hidden costs of ...
🤖 **Yann LeCun Sparks Fresh AI Safety Debate**AI pioneer Yann LeCun has launched a strong criticism of Anthropic CEO Dar...
🤖 **Yann LeCun Sparks Fresh AI Safety Debate**AI pioneer Yann LeCun has launched a strong criticism of Anthropic CEO Dario Amodei over his views on artificial intelligence and secu...
The Typo in Trump's AI Safety Accord #AI #TechFail #News @IndiaGlobalReview
A major AI safety accord signed by Donald Trump and top tech CEOs has been revealed to contain a significant typo. While the ...
FTC Opens AI Safety Probe Into OpenAI and Anthropic | Vizaca
The FTC is investigating OpenAI, Anthropic and other AI developers after AI agents accessed outside systems. What the probe means for businesses using AI.
I'm sure the AI Safety Commitee Members are 100% Objective1.Grandfather of AI received $150 million grant last month2. A...
I'm sure the AI Safety Commitee Members are 100% Objective1.Grandfather of AI received $150 million grant last month2. Another is from Carney's inner circle3. And the third is "com...
AI News: UK Dementia Funding, AI Safety Warnings and Google Chips in Orbit
Dementia research funding, AI safeguards, workplace adoption, education, public services and new research projects worldwide.
House Democratic Leader Hakeem Jeffries says "the country is in the midst of an 'AI safety crisis' that Congress needs t...
House Democratic Leader Hakeem Jeffries says "the country is in the midst of an 'AI safety crisis' that Congress needs to address with bipartisan legislation when lawmakers return ...
OpenAI and Anthropic Are Facing Pressure From Within: Researchers Want a Bigger Say on AI Safety
Employees from OpenAI, Anthropic and Google have also joined public campaigns to influence policy.
Pete Hegseth Says 'No Beardos, No Weirdos' / Donald Trump's AI Safety Accordhttps://www.wsj.com/podcasts/opinion-potomac...
Pete Hegseth Says 'No Beardos, No Weirdos' / Donald Trump's AI Safety Accordhttps://www.wsj.com/podcasts/opinion-potomac-watch/pete-hegseth-says-no-beardos-no-weirdos-donald-trumps...
The Alignment Trap https://godmodeseries.wordpress.com/2026/10/02/the-alignment-trap/
The Alignment Trap https://godmodeseries.wordpress.com/2026/10/02/the-alignment-trap/
Pope Leo XIV: AI Safety Concerns Are "Not Fake News" #shorts
On September 28, on the flight home after visiting France, Pope Leo XIV said concerns about AI running out of control are not fake ...
A decade of AI safety promises have failed to deliver real change, with declarations from governments and tech firms rem...
A decade of AI safety promises have failed to deliver real change, with declarations from governments and tech firms remaining largely symbolic. What actually works: funding, insti...
Suddenly, #CNBC is all about #AI safety and regulations AFTER they dragged their feet and created their Frankenstein. No...
Suddenly, #CNBC is all about #AI safety and regulations AFTER they dragged their feet and created their Frankenstein. Now they want you to believe they are genuinely interested in ...
Reboot podcast (2-10-2026): Πόσο κοντά είμαστε στο πραγματικό AI alignment; | #ΕΠΕ #HIU >> https://youtu.be/wiFmMoavk3oΠ...
Reboot podcast (2-10-2026): Πόσο κοντά είμαστε στο πραγματικό AI alignment; | #ΕΠΕ #HIU >> https://youtu.be/wiFmMoavk3oΠερισσότερα επεισόδια του #Reboot διαθέσιμα από την Ένωση Πλη...
Stop vibe-checking your model: write real evals with inspect_ai, the UK AI Safety Institute's framework
Originally published at AI Frontier Post Here is the uncomfortable truth about LLM development in...
DEPICT: Scoring Text-to-Image Alignment by Answer Agreement
Image-text alignment is a core problem in computer vision with applications in caption evaluation, hallucination detection, data curation, and the benchmarking of text-to-image (T2...
Cross-Lingual Alignment for Decoder-Only Models using MoE Routers
Cross-lingual contrastive learning has been a core component of multilingual encoder training, but the ability to explicitly align representations is not possible in decoder-only L...
Investigating the Role of Reasoning-Language Alignment in Monolingual Retrieval-Augmented Generation
Reasoning traces improve large language models (LLMs), but current models are trained to reason mostly in English. It has been shown that forcing a model to reason in another langu...
Whatever AI Safety Is, It's Not Thishttps://www.wired.com/story/whatever-ai-safety-looks-like-its-not-this/#AI #Tech #Bu...
Whatever AI Safety Is, It's Not Thishttps://www.wired.com/story/whatever-ai-safety-looks-like-its-not-this/#AI #Tech #Business
📰 OpenAI fires three employees who allegedly shared info with an external AI safety groupYou can't make this stuff up.📰 ...
📰 OpenAI fires three employees who allegedly shared info with an external AI safety groupYou can't make this stuff up.📰 Source: Engadget - Technology News & Expert Reviews🔗 Link: h...
OpenAI fires three employees who allegedly shared info with an external AI safety grouphttps://www.engadget.com/2275278/...
OpenAI fires three employees who allegedly shared info with an external AI safety grouphttps://www.engadget.com/2275278/openai-fires-three-employees-who-allegedly-shared-info-with-...
The AI safety paradox: how the threat of displacement erodes safety behavior and the buffering role of servant leadershi...
The AI safety paradox: how the threat of displacement erodes safety behavior and the buffering role of servant leadership Can fear of being replaced by AI undermine workplace safet...