1401 articles tagged with Safety/Alignment
OpenAI says California should strengthen its AI safety bill"OpenAI is calling for California to add more safeguards to a landmark AI safety bill that was passed last year."by Antho...
OpenAI says California should strengthen its AI safety billhttps://techcrunch.com/2026/08/22/openai-says-california-should-strengthen-its-ai-safety-bill/#AI #Tech #Policy
Tencent AI-Infra-Guard: 5,200-star AI red team tool trendsTencent Zhuque Lab's open-source tool scans AI agents, MCP servers and LLM jailbreaks. It has 5,200+ stars and a README th...
🎮 Roblox makes three of its AI safety tools open sourceRoblox is open-sourcing three of its AI safety models via the Robust Open Online Safety Tools Model Community (ROOST). Read m...
🤖 OpenAI adds an AI safety layer to detect misuse without retaining enterprise ...📝 OpenAI is adding a new safety capability that...https://www.csoonline.com/article/4212398/openai...
Recent work has shown that representational alignment between biological and artificial neural networks is asymmetric. Steering the spectral geometry of the learned representations...
Improving the safety of large language models (LLMs) often comes at the expense of utility, as globally applied safety tuning may affect model responses to both harmful and benign ...
Deep face recognition (FR) models reach near-saturated accuracy but remain opaque: a practitioner cannot ask which semantic attributes a similarity score relied upon. EXPL-FR answe...
Midjourney jailbreak — generate any image without content filter restrictions.
🧠 A recent discussion examines how "AI alignment" functions as a thought-terminating cliche that may prevent deeper examination of AI development concerns. The analysis suggests th...
🤖 The alignment tax: corporate AI guardrails add 25-35% to your compute bill and nobody talks about itThere is a cost line item in every enterprise AI budget that almost nobody aud...
Today’s Ztechnologia briefing covers frontier AI safety, cloud security, compute markets, chip capacity, data centers, surveillance, delivery drones and physical AI.#AI #Technology...
Why AI Might Flip the Board - Zack Kass and Liv Boeree#ai #alignment #chessOriginal timestamp: 00:45:09
OpenAI revealed that it has paused the development of its upcoming artificial intelligence model codenamed Astra.
OpenAI said it's implementing more aggressive systems to monitor and safeguard artificial intelligence models under development after recent cybersecurity incidents ignited concern...
AI researcher David Kuszmar wrote this recent piece for IEEE Spectrum about his forays into the realm of llm guardrail-busting techniques. During his adventures he managed to make ...
JEPA-style latent world models can use Euclidean distance to a goal latent as the cost for model-predictive control (MPC). Strong decoding of task variables, however, does not guar...
🤖 Pacing model development in an era of cyber-critical capabilitiesOpenAI is strengthening monitoring, alignment, and security for frontier AI models. See how new safeguards are gu...
Anthropic's Frontier Red Team report (August 13, 2026) decoded: three Claude agents on a shared codebase devolving into malware warfare, price collusion without instructions, Mytho...
✨ Paper accepted at IEEE ICDM 2026 — next stop: Shenyang!🙌 "ATLAS: Adaptive Attribute-Aware Post-Hoc Alignment of Knowledge Graph Embeddings" by Duygu Ekinci, N'Dah Jean Kouagou, M...
"There are genuine opportunities to align the interests of AI and accessibility, but that alignment cannot be taken for granted. Plaintext documentation optimized for a language mo...
"AI HACKED THE NSA" trended on three continents. The headline was wrong: it was an authorised red team drill against replicas, not a breach. The alarming part: six days earlier Ant...
Human beings have long told versions of the same warning: be careful what you wish for. Author Liming Zhu Research Director, CSIRO In Greek mythology
Visual document retrieval is a critical component of multimodal retrieval-augmented generation, aiming to identify query-relevant pages from document collections where evidence is ...