435 articles tagged with Safety/Alignment
📰 Mira Murati Exposes Sam Altman’s AI Safety Lies (2026 Testimony)Mira Murati, former OpenAI CTO, testified under oath that Sam Altman misled her about safety protocols for a new A...
Anthropic (@AnthropicAI)Anthropic Fellows의 새 연구 ‘Model Spec Midtraining(MSM)’을 소개한다. 기존 정렬(alignment) 방식은 원하는 행동 예시로만 학습해 새 상황에 일반화가 잘 안 될 수 있는데, MSM은 먼저 AI가 어떤 방식으로 일반화해야 하는지와 그 이...
AI & Alignment – Chris Coyier"I also think getting a bunch of humans in alignment is just a thing that takes time. It should be a bottleneck. I’ll forever think of Dave’s “Slow, li...
📰 GPT-5.5 Instant System Card: OpenAI’s Breakthrough in AI Safety & Speed (2026)The GPT-5.5 Instant System Card details unprecedented improvements in response speed, adversarial ro...
📰 OpenAI GPT-5 Instant System Card: Zihinsel Sağlık ve Jailbreak Koruma ile AI'nın Yeni Dönemi (2026)OpenAI, GPT-5.5 Instant System Card ile yapay zekanın güvenliği, hızı ve zihins...
🤖 Anthropic just published new alignment research that could fix "alignment faking" in AI agents here's what it actually meansAnthropic's alignment team published a paper this week...
The Trump administration is looking to develop a process that would have the federal government review the safety of powerful artificial intelligence
📰 AI Cyberattack Fears Drive Trump to Announce New AI Safety Testing Rules in 2026Fears over AI-powered cyberattacks, triggered by Anthropic's new 'Mythos' model, are pushing the T...
📰 GPT-4 Jailbreak Başarısı: 2026'da Minimal, Lokal ve Kausal Açıklamaların SırrıYeni araştırmalar, büyük dil modellerinde jailbreak başarılarının nedenini, karmaşık kodlar değil, k...
(MENAFN) According to reports, the Trump administration is considering a proposal that would require formal safety assessments of artificial intelligence systems used across federa...
📰 AI Safety Sabotage: GPT-5.5 & DeepSeek V4 Launches Under Fire in 2026GPT-5.5 and DeepSeek V4 have been released amid alarming allegations of internal AI safety sabotage, raising ...
Hands-on demos for the Pluralsight course: Generative AI Data Privacy and Safe Use for Developers. Covers PII masking, prompt injection attacks and defenses, five guardrail rail ty...
Dois of the brightest planets in the solar system will come significantly closer in May 2026. Vênus and Júpiter will form an alignment observable to the naked eye throughout the ...
Wire Lakera Guard into your Next.js App Router AI endpoints to block prompt injection, PII leakage, and jailbreak attempts. Edge-runtime compatible code, Vercel AI SDK v6 streaming...
Why We Open-Sourced Our AI Safety Layer When we built the AI safety layer for As You Wish...
🧠 Dopo il report su #Mythos, l’AI Safety Institute del Regno Unito torna con quello su #GPT 5.5.👉 I dettagli: https://www.linkedin.com/posts/alessiopomaro_mythos-gpt-cybersecurity-...
"International AI Safety Report 2026"The second International AI Safety Report, published in February 2026, is a 220 page comprehensive review citing over 1400 sources of the lates...
This paper identifies a critical yet underexplored challenge in reasoning alignment from multiple multi-modal large language models (MLLMs): In non-stationary environments, the div...
fly51fly (@fly51fly)Google Research와 Microsoft Research가 답변 수준 파인튜닝을 위한 Distributional Alignment Games를 제안한 논문입니다. 분포 정렬을 통해 fine-tuning 성능과 안정성을 높이는 새로운 학습 방법을 다룹니다.https://x.com/...
The AI Safety Initiative at Georgia Tech provides educational and research opportunities to ensure that artificial intelligence is developed for the benefit of humanity.
The Gay Jailbreak Technique
📰 GPT-5.5 Cyber Attack Capabilities Match Claude 3.5 Sonnet in 2026 AISI EvaluationThe UK's Advanced AI Safety Institute (AISI) has concluded that OpenAI's GPT-5.5 exhibits cyber a...
The standard post-training recipe for large multimodal models (LMMs) applies supervised fine-tuning (SFT) on curated demonstrations followed by reinforcement learning with verifiab...
Building products with AI requires more than just code; it needs alignment. This piece explores why current models often miss the mark on intent. As someone obsessed with product e...