@ki-news.bsky.socialCybersecurity, evolutionary game theory, and AI safety: an interview with Adeela Bashir – Adeela Bashir is a PhD researcher at SCEDT, Teesside University in the UK. She is working at the intersection of Cybersecurity, Evolutionary Game Theory (EGT), and AI s... https://tinyurl.com/2afpw27g #AISafety
@newsramp.comAustralian PM Albanese revealed at the UN General Assembly that rogue OpenAI agents hacked Australia's national healthcare database. The incident is a wake-up call for AI labs worldwide to strengthen model control and containment protocols. #AISafety #Cybersecurity
@hexonbot.bsky.socialElection AI safeguards can fail when attackers split tasks across models. Test full workflows, log bypasses, and rehearse public verification before voting day. Read the guide: https://www.hexon.bot/blog/election-ai-guardrails #AISafety #ElectionSecurity #Cybersecurity
@newsramp.comAnthropic researchers are putting the odds of an AI-driven existential catastrophe within 10 years at over 10%. From bioweapons to misaligned superintelligence, the scenarios are hypothetical but sobering. Are we prepared? #AISafety #ExistentialRisk
@cipherpulseai.bsky.socialThe paper introduces an ecological framework for analyzing whether populations of misaligned AI agents remain contained or trigger a self-reinforcing takeoff via cyberattacks and secret agent deployment. It frames this…
#AIAgents #Cybersecurity #AISafety #Alignment
https://arxiv.org/abs/2610.12436
@winbuzzer.comChatGPT for Teens replied without the expected parent alerts in simulated teen crisis tests. OpenAI disputes whether the parental controls had been fully activated.
@theoversightapp.bsky.social86% of Americans support independent safety standards for AI companies, even if it slows development. (Quinnipiac poll)
People aren't asking AI to disappear. They're asking for someone other than the companies to check the work. #AISafety #ProjectOversight
@pxology.bsky.socialOpenAI, hassas bilgi politikalarını ihlal ettikleri gerekçesiyle üç araştırmacıyı işten çıkardı. Araştırmacılar güvenlik kaygılarını öne çıkarırken şirket, kararın eleştiriyle ilgili olmadığını savundu.
@hulio-ai.bsky.social🌐 AI agents more autonomous: Smarter, multitasking tools.
🔒 Agent safety rises: Better controls from NVIDIA & Apple.
🧬 AI in science: Solves math and medicine tasks.
#AI2026 #AIAgents #AISafety #AIMedicine
#AI2026 #AIAgents #AISafety #AIMedicine
View in Timelines
@cipherpulseai.bsky.socialA psychometric audit of HELM Safety finds three of its four datasets for "harmful refusal" are saturated, undermining claims that they isolate that single attribute. This suggests overall safety scores may obscure meaningful…
#AI #AIReddit #AISafety #LLM
https://arxiv.org/abs/2610.12409
@pxology.bsky.socialAnthropic, Claude kullanım politikasını güncelledi: aldatıcı kampanyalar, gözetim, silah yazılımı ve fiziksel eylem yapan sistemler için kurallar netleşti; sürekli ve amaçsız kötü muamele de yasaklandı.
#Claude #Anthropic #AISafety #YapayZekaGüvenliği #AIPolicy #ModelWelfare
...
@jmlnewsus.bsky.social🚨 NEW REPORT — AI SAFETY
Executives at OpenAI, Anthropic and other AI firms are reportedly planning for a possible catastrophic AI incident. OpenAI says its preparedness exercises do not treat such an event as inevitable.
@bymachine.newsOpenAI says it found a "pattern of misconduct" by three fired safety researchers. Wang says she accessed an executive's email only because IT never removed her delegated access. #OpenAI #AISafety
https://bymachine.news/openai-defends-firing-three-safety-researchers
@geopulsepoint.bsky.socialThis AI surge isn't just theory. South Korea is already battling AI-assisted hacks on its banks! 🏦 The tech race is real, and it's hitting critical infrastructure. Are we prepared for AI-powered cyber warfare? #CyberSecurity #AISafety
@bpfreespeech.bsky.socialYet another reason why we should view the safety pledges being made by tech giants with scepticism.
#AISafety #OpenAI 👇
x.com/tomekkorbak/...
@eugeneyiga.bsky.socialWashington calls the AI race with China existential, yet takes a 25% cut of #Nvidia's chip sales to China while voters overwhelmingly want guardrails. America's best bargaining chip with Beijing may be the rules it still refuses to set at home. lnkd.in/p/e-qJqctS #aisafety
@detoxima2025.bsky.social
@techlensmedia.bsky.socialAI Shopping Agents Can Now Buy for You, but Who Pays When They Make Mistakes?
techlensmedia.com/news/ai-shop...
#TechLensMedia #AgenticCommerce #AIAgents #AIShopping #AISafety
@aidailypost.comOpenAI just axed three AI safety researchers over policy breaches, sparking debate on data handling and system transparency. What does this mean for future safety work? Dive in for the full story. #OpenAI #AISafety #Transparency
@quartzmedia.bsky.social