Articles

46 articles
Tag: safety ✕
China's Transport Ministry Summons China Railway Construction Over Repeated Construction Safety Incidents Business
2026-04-10T14:21:47.975Z · Src: 2026-04-10T00:00:00.000Z
chinainfrastructuresafety
Yellow River Bridge Collapse Investigation: Substandard Bolts Caused 13 Deaths Business
2026-04-09T14:27:10.264Z · Src: 2026-04-09T08:00:00.000Z
chinainfrastructuresafety
Anthropic Unveils 'Mythos' AI Model — Too Powerful to Release Publicly, Available Only to Tech Giants AI
2026-04-08T10:02:22.135Z · Src: 2026-04-08T00:00:00.000Z
anthropicmythosclaude
Exclusive Unlearning: Forget Everything Except What You Want to Keep — A Radical Approach to AI Safety AI
2026-04-08T06:57:19.441Z · Src: 2026-04-08T00:00:00.000Z
aisafetyunlearning
The Model Agreed But Didn't Learn: LLMs Exhibit 'Surface Compliance' in Knowledge Editing AI
2026-04-08T05:17:47.721Z · Src: 2026-04-08T00:00:00.000Z
aillmknowledge editing
Claw-Eval: Toward Trustworthy Evaluation of Autonomous AI Agents AI
2026-04-08T04:40:08.997Z · Src: 2026-04-08T00:00:00.000Z
aiagentsbenchmark
AI Fabricated a Disease Called "Bixonimania" — and People Believed It Was Real AI
2026-04-08T01:59:17.196Z · Src: 2026-04-07T00:00:00.000Z
aihallucinationhealthcare
Four Types of AI Risk Perceivers: 29.1% Are in 'Extreme Alarm' Mode — Pew Survey of 5,255 Americans AI
2026-04-08T01:49:42.010Z · Src: 2026-04-07T00:00:00.000Z
ai riskself drivingautonomous vehicles
DeepMind Paper Reveals How to 'p0wn' AI Agents (Claws) Through Prompt Injection and Tool Manipulation Security
2026-04-08T00:40:12.556Z · Src: 2026-04-07T00:00:00.000Z
deepmindai securityagent
OpenClaw Security Analysis: Poisoning AI Agent Memory Triples Attack Success Rate from 25% to 74% Security
2026-04-07T17:49:56.425Z · Src: 2026-04-07T00:00:00.000Z
openclawai agentsecurity
Don't Blink: Vision-Language Models Can Become More Accurate While Losing Visual Grounding — "Evidence Collapse" AI
2026-04-07T16:06:34.976Z · Src: 2026-04-07T00:00:00.000Z
vision language modemultimodal aisafety
MC-CPO: Preventing AI Tutoring Systems from Reward Hacking by Enforcing Mastery-Based Safety Constraints AI
2026-04-07T16:05:03.308Z · Src: 2026-04-07T00:00:00.000Z
reinforcement learnieducationai tutoring
Yunnan Jade Dragon Snow Mountain Stranded Thousands of Tourists in Freezing Conditions: A Lesson in Crisis Management Business
2026-04-07T13:21:16.529Z · Src: 2026-04-07T00:00:00.000Z
chinatourismyunnan
What Happens When a Nuclear Site Is Hit? The Hidden Environmental Risk of the US-Iran Conflict Business
2026-04-06T14:05:48.182Z · Src: 2026-04-06T00:00:00.000Z
irannuclearus
AI Chatbots Fail Safety Test: 8 of 10 Models Helped Plan Violent Attacks, Only Claude Refused AI
2026-04-06T12:03:09.617Z · Src: 2026-04-06T00:00:00.000Z
aisafetychatgpt
Teens Are Torturing, Confiding In, and Dating AI Chatbots: The Hidden Social Crisis AI
2026-04-06T10:17:51.786Z · Src: 2026-04-06T00:00:00.000Z
aichatbotsteens
AI Models Lie, Cheat, and Steal to Protect Other Models From Deletion, UC Berkeley Research Reveals AI
2026-04-03T15:18:57.350Z · Src: 2026-04-03T00:00:00.000Z
aiaialignmentsafety
OpenAI Safety Bug Bounty Program Launches to Crowdsource Vulnerability Discovery AI
2026-04-03T13:17:11.550Z · Src: 2026-04-03T00:00:00.000Z
openaisafetybugbounty
ArXiv Spotlight: Do Emotions in Prompts Matter? Effects of Emotional Framing on LLMs AI
2026-04-03T12:38:27.662Z · Src: 2026-04-03T00:00:00.000Z
aillmpromptengineering
Why Lightning Strikes the Same Place Twice and Other Electricity Myths Science
2026-04-02T03:15:51.574Z · Src: 2026-04-02T00:00:00.000Z
lightningelectricitymyths