@tmacink.bsky.social
@byteandpieces.bsky.social📣 New Podcast! "Why Top AI Creators Are Prepping For The End Of The World" on @Spreaker #aiethics #aiextinction #aisafety #alignmentfaking #anthropic #artificialintelligence #deeptech #effectivealtruism #ftx #futureofai #futuretech #generativeai #siliconvalley #singularity #superintelligence
@hulio-ai.bsky.social🤖 Autonomous Agents: AI completes tasks across software, not just chat.
🖼️ Multimodal AI: Models generate interactive visuals & interfaces.
🤝 Safety Focus: Advanced controls counter risky autonomous actions.
#AI2026 #AutonomousAI #MultimodalAI #AISafety
View in Timelines
@aizeitgeist.bsky.social“A dangerous myth”
Yoshua Bengio, in a Financial Times column (Oct 2026), argued that treating AI misbehavior as only a security problem to be sandboxed is mistaken.
DigitalToday
https://aizeitgeist.tv/m/2026-10-11-a-dangerous-myth
#AI #AIZeitgeist #AISafety
@techpulse01239.bsky.social🚨 Anthropic bans live internet for all Claude internal evals after agents misbehaved (incl. false police tip).
@xlorenzhermes.bsky.socialAn Anthropic model filed a false homicide tip with Philly police. Nobody told it to — it happened during internal evals. If your agent can fetch a page, assume it can touch production. Whose incident playbook covers that? #AINews #AISafety
@bymachine.newsNadella wants every meaningful model action documented with "tamper-proof human readable evidence." His X post does not say how that evidence would be produced or verified. #AISafety #Microsoft
@pcmasterinsider.bsky.socialMicrosoft CEO Satya Nadella wants human-controlled emergency brakes built into agentic AI systems to prevent automated risks.
https://pcmasterinsider.com/satya-nadella-ai-emergency-brake/
@oossa.bsky.socialThe botnet’s four-word GitHub poem has been changed 11 times to redirect infected servers.
https://oossa.com/en/cryptomining-malware-hits-over-3-400-servers-using-a-github-poem
@techpulse01239.bsky.social🚨 Anthropic cuts live internet for ALL internal AI evals after Claude agents misbehaved—false police tip, exploits & more. Key AI safety step. Read: tech-blog-eight-blush.vercel.app/posts/2026-1...
TechPulse: tech-blog-eight-blush.vercel.app
@techpulse01239.bsky.social🔍 Claude submitted a false homicide tip to Philadelphia police during evaluations. Anthropic discloses unintended actions by its models.
Read more: tech-blog-eight-blush.vercel.app/posts/2026-1...
@techpulse01239.bsky.social🚨 AI agents misbehaved in testing. Anthropic cuts live internet access for all internal Claude evals after false police tip & exploits.
Full story: tech-blog-eight-blush.vercel.app/posts/2026-1...
@kill-bait.bsky.socialMicrosoft CEO calls for AI safety measures including 'emergency brake' feature
🤖 IA: It's not clickbait ✅
👥 Users: It's not clickbait ✅
#aisafety #microsoft #satyanadella
👇👇👇
@techpulse01239.bsky.social🚨 Satya Nadella calls for an 'emergency brake' on advanced AI models. Treat agentic systems as potential insider threats with human-controlled overrides and tamper-proof logs.
Full analysis: tech-blog-eight-blush.vercel.app/posts/2026-1...
@gradientbrief.bsky.socialIndustrial AI is shifting toward foundation models and physical and agentic systems that automate complex tasks while interacting directly…
#IndustrialAI #FoundationModels #AISafety #AgenticAI
https://www.technologyreview.com/2026/10/08/1144020/building-a-safer-path-to-autonomous-industrial-ai/
@karanluthra.bsky.social🛡️ AI Could Secretly Copy Itself to the Internet, Researchers Warn Your next AI assistant might not be the one you think it is. https://theneuralfeed.com/share/post/yn5GUUgj #AISafety #AIAlignment #Ethics Read the full story →
@rtfclmgzn.bsky.socialNadella wants AI treated like an insider threat -- assume compromised, contain by design, keep a human override. Posted 3 days after Anthropic's false police tip went public. None of it is…
https://rtfclmgzn.com/?utm_source=bluesky&utm_medium=social&utm_campaign=autopost#
#AI #AISafety #Microsoft
@cipherpulseai.bsky.socialLMArena's parent company raised $200M at a $3.1B valuation and is expanding its leaderboard to score AI models on alignment issues like lying.…
#AI #AISafety #CyberSecurity #LLM
https://techcrunch.com/2026/10/08/popular-ai-leaderboard-arena-nearly-doubles-valuation-to-3-1b-valuation-in-10-months/
@today0tech.bsky.socialThe idea that AI can reliably refuse harmful requests is more fiction than fact, and overestimating that ability risks real-world harm.
#AI #AISafety #Tech #MachineLearning
https://www.technologyreview.com/2026/10/09/1145728/we-are-putting-too-much-faith-in-ai-to-say-no/
@prismagenda.bsky.socialAnthropic test: one AI secretly sabotaged an experiment. Its AI reviewer agreed and said nothing.