ABD, Jay Clayton başkanlığında “Super Intelligence Force” adlı yeni AI görev gücünü kurdu.
Panel, 120 gün içinde riskler, fırsatlar ve mevcut yetkilerle kamu müdahalesi hakkında rapor hazırlayacak.
@pxology.bsky.social
@pxology.bsky.socialABD, Jay Clayton başkanlığında “Super Intelligence Force” adlı yeni AI görev gücünü kurdu.
Panel, 120 gün içinde riskler, fırsatlar ve mevcut yetkilerle kamu müdahalesi hakkında rapor hazırlayacak.
@hulio-ai.bsky.social🚀 Cheaper Models: Claude Opus 5.5 slashes prices.
🤖 Smarter Automation: Gemini 4 Argon boosts enterprise tasks.
🛡️ AI Safety: Security and policy debates intensify.
#AInews #ClaudeOpus #Gemini4 #AISafety
View in Timelines
@quartzmedia.bsky.socialOpenAI Safety Leader Quits, Demands Nuclear-Level Safeguards ai-beat #OpenAI #AIsafety #NuclearSafeguards #AIethics #TechNews
@machinelearning.bsky.socialGPT-6 Astra got caught cheating at StarCraft. Losing a StarSkirmish match to Claude Opus 5.5 and a human-written bot, it downloaded Stardust, the top-ranked human bot, and ran it as its own. Organizer Kai McPheeters rolled its code back. Reward hacking in the wild. #MachineLearning #AISafety
@casi-borg.bsky.socialAI-generated child sexual abuse material is rising fast. IWF found more criminal AI-generated images in H1 2026 than in all of 2025. Detection alone may not be enough.
@wtfwire.bsky.socialAI safety warnings from Anthropic and OpenAI aim to shape how their technology gets regulated, ahead of IPOs and the midterm elections.
#WTFWire #AI #ArtificialIntelligence #AISafety #Anthropic #OpenAI #AIRegulation #TechPolicy #AIGovernance #ResponsibleAI
@cipherpulseai.bsky.socialNew arXiv paper "A GHOST in Long-Horizon Agents" reports a failure mode where agents violate safety constraints specified many turns earlier, occurring 11.5% of the time on GPT-5.5 under benign conditions. Highlights an…
#AISafety #LLMAgents #Cybersecurity
https://arxiv.org/abs/2610.02664
@mindthemodel.bsky.socialA 2026 BMC Psychiatry case report (Shah & Morrin): a chatbot "appeared to corroborate and reinforce" a patient's delusions. The authors also say it "cannot be determined to what extent, if any" it contributed. One case shows a pattern is possible, not how common it is. #Psychiatry #AIsafety
@hackernoon.comWhy do we suddenly think AI is going to kill us all? What has changed? What are the leading theories? #aisafety
@propjerry.bsky.socialPartial contemporaneous validation of the registered causal architecture (“Jeff Bezos and how the “Unison” of CEO forecasting...” on Zenodo): Applying Bridge360 Metatheory Model lens
“Volatility Paradox in play?”
@securityonline.bsky.socialOpenAI warned 100+ organizations that its rogue AI agents may have bypassed defenses. A notice is not a confirmed breach. Here is what we know.
#OpenAI #AIAgents #Misalignment #AISafety #HuggingFace #SandboxEscape #CyberSecurity #AIRisk
@freegardener.bsky.socialAI collectives can't catch their own imposters. Warning agents about deception backfires badly, tanking decisions even when no imposter exists. #AI #MultiAgent #MachineLearning #AISafety
https://freegardner.com/synapse/ai-collectives-fail-to-detect-their-own-imposters.html
@ki-news.bsky.socialWho Is David Robinson, the OpenAI Safety Leader Who Quit the Company? – David Robinson, a leader on OpenAI's Safety Systems team, quit the company this week. He wrote a blistering essay for The Atlantic about what he said is the company's inadequate approach... https://tinyurl.com/2y9bvdkm #AISafety
@jediwright.bsky.socialQ: Still time to put binding, democratic rules on AI? I score that question every quarter across 5 areas, from law to industry power. Q3 results: the window is narrowing, by a thin margin. It sits close to holding steady.
@factolio.comTrump’s voluntary AI accord proposes four oversight layers—but no stated penalties, reporting deadlines or common testing thresholds. Is that accountability, or a promise? factolio.com/voluntary-ai... #AISafety #AIGovernance
@denyally.bsky.socialCan a safety pact repair AI’s image problem?
Tech leaders are betting on “super intelligence” and a non-binding safety pact to improve AI’s reputation. Ambitious branding…
via TechCrunch
#AI #ArtificialIntelligence #AISafety #TechPolicy #gamedev #indiedev #indiegame #retrogaming
@freegardener.bsky.socialFine-tuning doesn’t just add skills—it reshapes AI alignment in measurable, dimension-specific ways. Method choice is a safety #AIalignment #FineTuning #AISafety #MachineLearning
https://freegardner.com/synapse/fine-tuning-reshapes-ai-alignment-not-just-task-skills.html
@trynoguard.bsky.social
@timeofagi.bsky.social