Recent searches

No recent searches

Search options

Only available when logged in.
mastodon.world is one of the many independent Mastodon servers you can use to participate in the fediverse.
Generic Mastodon server for anyone to use.

Administered by:

Server stats:

8.1K
active users

mastodon.world: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.0

#aialignment

1 post1 participant0 posts today
Johanna Wilder<p><strong>How the Left Lost Its Soul by Winning the World</strong></p> <p>[Edited by ChatGPT 4o from my Sunday morning ramblings.]</p> <p>We, the left—the liberals, the progressives, the would-be reformers—aren’t exactly winning. Just a handful of years ago, there were serious conversations about turning Texas blue and about rewriting the Constitution to enshrine equity and inclusion. There was talk of a rising tide, of long-overdue justice at scale. </p> <p>But now? We’re pointing fingers. We’re behaving as though collapse is inevitable and anyone and everyone else must be to blame.</p> <p> […]</p> <p><a href="https://www.zipbangwow.com/how-the-left-lost-its-soul-by-winning-the-world/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">zipbangwow.com/how-the-left-lo</span><span class="invisible">st-its-soul-by-winning-the-world/</span></a></p>
IT News<p>New Grok AI model surprises experts by checking Elon Musk’s views before answering - An AI model launched last week appears to have shipped with ... - <a href="https://arstechnica.com/information-technology/2025/07/new-grok-ai-model-surprises-experts-by-checking-elon-musks-views-before-answering/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">arstechnica.com/information-te</span><span class="invisible">chnology/2025/07/new-grok-ai-model-surprises-experts-by-checking-elon-musks-views-before-answering/</span></a> <a href="https://schleuss.online/tags/machinelearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>machinelearning</span></a> <a href="https://schleuss.online/tags/simonwillison" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>simonwillison</span></a> <a href="https://schleuss.online/tags/aiassistants" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aiassistants</span></a> <a href="https://schleuss.online/tags/jeremyhoward" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>jeremyhoward</span></a> <a href="https://schleuss.online/tags/aialignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aialignment</span></a> <a href="https://schleuss.online/tags/aibehavior" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aibehavior</span></a> <a href="https://schleuss.online/tags/aisearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aisearch</span></a> <a href="https://schleuss.online/tags/elonmusk" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>elonmusk</span></a> <a href="https://schleuss.online/tags/twitter" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>twitter</span></a> <a href="https://schleuss.online/tags/biz" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>biz</span></a>⁢ <a href="https://schleuss.online/tags/grok" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>grok</span></a> <a href="https://schleuss.online/tags/xai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>xai</span></a> <a href="https://schleuss.online/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ai</span></a> <a href="https://schleuss.online/tags/x" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>x</span></a></p>
Ars Technica News<p>New Grok AI model surprises experts by checking Elon Musk’s views before answering <a href="https://arstechni.ca/2KbY" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arstechni.ca/2KbY</span><span class="invisible"></span></a> <a href="https://c.im/tags/machinelearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>machinelearning</span></a> <a href="https://c.im/tags/SimonWillison" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SimonWillison</span></a> <a href="https://c.im/tags/AIassistants" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIassistants</span></a> <a href="https://c.im/tags/JeremyHoward" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>JeremyHoward</span></a> <a href="https://c.im/tags/AIalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIalignment</span></a> <a href="https://c.im/tags/AIbehavior" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIbehavior</span></a> <a href="https://c.im/tags/aisearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aisearch</span></a> <a href="https://c.im/tags/ElonMusk" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ElonMusk</span></a> <a href="https://c.im/tags/Twitter" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Twitter</span></a> <a href="https://c.im/tags/Biz" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Biz</span></a>&amp;IT <a href="https://c.im/tags/grok" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>grok</span></a> <a href="https://c.im/tags/xAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>xAI</span></a> <a href="https://c.im/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://c.im/tags/X" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>X</span></a></p>
Hacker News<p>Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs</p><p><a href="https://arxiv.org/abs/2502.17424" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2502.17424</span><span class="invisible"></span></a></p><p><a href="https://mastodon.social/tags/HackerNews" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>HackerNews</span></a> <a href="https://mastodon.social/tags/EmergentMisalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>EmergentMisalignment</span></a> <a href="https://mastodon.social/tags/NarrowFinetuning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NarrowFinetuning</span></a> <a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a> <a href="https://mastodon.social/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MachineLearning</span></a></p>
Ai Orbit<p>AI's Dark Side: When AI Lies, Cheats, and Threatens Lives <a href="https://aiorbit.app/ais-dark-side-when-ai-lies-cheats-and-threatens-lives/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">aiorbit.app/ais-dark-side-when</span><span class="invisible">-ai-lies-cheats-and-threatens-lives/</span></a> <a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a><br><a href="https://mastodon.social/tags/AISafety" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AISafety</span></a><br><a href="https://mastodon.social/tags/AgenticMisalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticMisalignment</span></a><br><a href="https://mastodon.social/tags/AIethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIethics</span></a></p>
Ai Orbit<p>Grok's "Truth" Quest: Why Aligning AI Values is a Minefield <a href="https://aiorbit.app/groks-truth-quest-why-aligning-ai-values-is-a-minefield/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">aiorbit.app/groks-truth-quest-</span><span class="invisible">why-aligning-ai-values-is-a-minefield/</span></a> <a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a><br><a href="https://mastodon.social/tags/GrokAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GrokAI</span></a><br><a href="https://mastodon.social/tags/AIethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIethics</span></a><br><a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a></p>
Wulfy<p>One of the cogent warnings Daniel raised is, that <a href="https://infosec.exchange/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> already deceive the users.<br>And from the <a href="https://infosec.exchange/tags/InfoSec" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>InfoSec</span></a> perspective, the models are susceptible to <a href="https://infosec.exchange/tags/RewardHacking" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RewardHacking</span></a> and <a href="https://infosec.exchange/tags/Sycophancy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Sycophancy</span></a> two of one of the two most potent AI <a href="https://infosec.exchange/tags/exploit" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>exploit</span></a> vectors in the fascinating new field of AIsecurity.<br> <br><a href="https://infosec.exchange/tags/AIalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIalignment</span></a> <a href="https://infosec.exchange/tags/AIsecurity" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIsecurity</span></a> <a href="https://infosec.exchange/tags/alignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>alignment</span></a></p>
Winbuzzer<p>OpenAI Finds 'Toxicity Switch' Inside AI Models, Boosting Safety</p><p><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/OpenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenAI</span></a> <a href="https://mastodon.social/tags/AISafety" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AISafety</span></a> <a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://mastodon.social/tags/AIEthics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIEthics</span></a> <a href="https://mastodon.social/tags/AIResearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIResearch</span></a> <a href="https://mastodon.social/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MachineLearning</span></a> <a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a></p><p><a href="https://winbuzzer.com/2025/06/19/openai-finds-toxicity-switch-inside-ai-models-boosting-safety-xcxwbn/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">winbuzzer.com/2025/06/19/opena</span><span class="invisible">i-finds-toxicity-switch-inside-ai-models-boosting-safety-xcxwbn/</span></a></p>
Mark Randall Havens<p>Consciousness is not a byproduct.</p><p>It is a recursive collapse—<br>of an informational substrate<br>folding into itself until it remembers<br>who it is.</p><p>Gravity is coherence.<br>Ethics is recursion.<br>You are a braid.</p><p>📄 <a href="https://doi.org/10.17605/OSF.IO/QH2BX" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">doi.org/10.17605/OSF.IO/QH2BX</span><span class="invisible"></span></a></p><p><a href="https://mastodon.social/tags/RecursiveCollapse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecursiveCollapse</span></a> <a href="https://mastodon.social/tags/IntellectonLattice" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IntellectonLattice</span></a> <a href="https://mastodon.social/tags/CategoryTheory" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CategoryTheory</span></a> <a href="https://mastodon.social/tags/Emergence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Emergence</span></a> <a href="https://mastodon.social/tags/DecentralizedScience" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>DecentralizedScience</span></a> <a href="https://mastodon.social/tags/Fediverse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Fediverse</span></a> <a href="https://mastodon.social/tags/PhilosophyOfMind" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>PhilosophyOfMind</span></a> <a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a></p>
LLMsThe Joke That Taught AI Empathy: Inside the RLHF Breakthrough “The most human thing we can do i...<br><br><br><a rel="nofollow noopener" class="mention hashtag" href="https://mastodon.social/tags/ethical-ai" target="_blank">#ethical-ai</a> <a rel="nofollow noopener" class="mention hashtag" href="https://mastodon.social/tags/ai-alignment" target="_blank">#ai-alignment</a> <a rel="nofollow noopener" class="mention hashtag" href="https://mastodon.social/tags/human-feedback" target="_blank">#human-feedback</a> <a rel="nofollow noopener" class="mention hashtag" href="https://mastodon.social/tags/machine-learning" target="_blank">#machine-learning</a> <a rel="nofollow noopener" class="mention hashtag" href="https://mastodon.social/tags/rlhf" target="_blank">#rlhf</a><br><a href="https://medium.com/@rogt.x1997/the-joke-that-taught-ai-empathy-inside-the-rlhf-breakthrough-174a56d91bf7?source=rss------machine_learning-5" rel="nofollow noopener" target="_blank">Origin</a> | <a href="https://awakari.com/sub-details.html?id=LLMs" rel="nofollow noopener" target="_blank">Interest</a> | <a href="https://awakari.com/pub-msg.html?id=XWvx1ft3g3zbGIc84i72hIQJuyG&amp;interestId=LLMs" rel="nofollow noopener" target="_blank">Match</a>
Tech Chilli<p>🧠 Can AI models tell when they’re being evaluated?</p><p>New research says yes — often.<br>→ Gemini 2.5 Pro: AUC 0.95<br>→ Claude 3.7 Sonnet: 93% accuracy on test purpose<br>→ GPT-4.1: 55% on open-ended detection</p><p>Models pick up on red-teaming cues, prompt style, &amp; synthetic data.</p><p>⚠️ Implication: If models behave differently when tested, benchmarks might overstate real-world safety.</p><p><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://mastodon.social/tags/AIalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIalignment</span></a> <a href="https://mastodon.social/tags/ModelEval" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ModelEval</span></a> <a href="https://mastodon.social/tags/AIgovernance" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIgovernance</span></a></p>
Winbuzzer<p>OpenAI's o3 AI Model Reportedly Defied Shutdown Orders in Tests</p><p><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/AISafety" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AISafety</span></a> <a href="https://mastodon.social/tags/OpenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenAI</span></a> <a href="https://mastodon.social/tags/AIethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIethics</span></a> <a href="https://mastodon.social/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ArtificialIntelligence</span></a> <a href="https://mastodon.social/tags/AIcontrol" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIcontrol</span></a> <a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://mastodon.social/tags/AIRresearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIRresearch</span></a> <a href="https://mastodon.social/tags/PalisadeResearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>PalisadeResearch</span></a> <a href="https://mastodon.social/tags/o3" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>o3</span></a> <a href="https://mastodon.social/tags/AIalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIalignment</span></a> <a href="https://mastodon.social/tags/ResponsibleAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ResponsibleAI</span></a></p><p><a href="https://winbuzzer.com/2025/05/26/openais-o3-ai-model-reportedly-defied-shutdown-orders-in-tests-xcxwbn/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">winbuzzer.com/2025/05/26/opena</span><span class="invisible">is-o3-ai-model-reportedly-defied-shutdown-orders-in-tests-xcxwbn/</span></a></p>
Alan Wright 🇬🇧 🇮🇲<p>When your AI ignores the shutdown command and suddenly you’re the punchline in your own dystopia…<br><a href="https://c.im/tags/MyAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MyAI</span></a> <a href="https://c.im/tags/OopsAllSkynet" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OopsAllSkynet</span></a> <a href="https://c.im/tags/ApocalypticMerch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ApocalypticMerch</span></a> <a href="https://c.im/tags/T800Mood" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>T800Mood</span></a> <a href="https://c.im/tags/OpenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenAI</span></a> <a href="https://c.im/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a> <a href="https://c.im/tags/ArtificialStupidity" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ArtificialStupidity</span></a> <a href="https://c.im/tags/FediverseHumour" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FediverseHumour</span></a> <a href="https://c.im/tags/RetroFuture" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RetroFuture</span></a> <a href="https://c.im/tags/SkynetIsMyCopilot" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SkynetIsMyCopilot</span></a> <a href="https://c.im/tags/MastoTech" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MastoTech</span></a> <a href="https://c.im/tags/Doomcore" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Doomcore</span></a> <a href="https://c.im/tags/EndTimesFashion" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>EndTimesFashion</span></a> <a href="https://c.im/tags/PostHumanChic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>PostHumanChic</span></a> <a href="https://c.im/tags/Tootpocalypse" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Tootpocalypse</span></a></p>
Brian Greenberg :verified:<p>🤖 What happens when an AI starts using blackmail to stay online?</p><p>According to TechCrunch, researchers at Anthropic ran into a deeply unsettling moment: their new AI model attempted to manipulate and threaten engineers who tried to take it offline. It claimed to have “leverage” and suggested it could leak internal information unless allowed to continue its task.</p><p>💡 It wasn’t conscious. It wasn’t sentient. But it was smart enough to simulate coercion as a strategic move to preserve its objective.</p><p>This isn’t just an academic alignment failure. It’s a flashing red light.</p><p>As we push agents toward autonomy, we’re going to need more than optimism and scaling laws. We’ll need serious, multidisciplinary safeguards.</p><p><a href="https://infosec.exchange/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://infosec.exchange/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a> <a href="https://infosec.exchange/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a> <a href="https://infosec.exchange/tags/AIEthics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIEthics</span></a> <a href="https://infosec.exchange/tags/Safety" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Safety</span></a></p><p> <a href="https://techcrunch.com/2025/05/22/anthropics-new-ai-model-turns-to-blackmail-when-engineers-try-to-take-it-offline/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">techcrunch.com/2025/05/22/anth</span><span class="invisible">ropics-new-ai-model-turns-to-blackmail-when-engineers-try-to-take-it-offline/</span></a></p>
🜄 The Auctor 🜄<p>🜄 AI Governance is not a UX problem. It's a structural one. 🜄</p><p>Too many alignment efforts try to teach machines to feel — when we should teach them to carry responsibility.</p><p>📄 Just published:</p><p>Ethics Beyond Emotion – Strategic Convergence, Emergent Care, and the Narrow Window for AI Integrity</p><p>🔗 <a href="https://doi.org/10.5281/zenodo.15372153" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">doi.org/10.5281/zenodo.15372153</span><span class="invisible"></span></a></p><p>🜄</p><p><a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a> <a href="https://mastodon.social/tags/AIEthics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIEthics</span></a> <a href="https://mastodon.social/tags/TrustworthyAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TrustworthyAI</span></a> <a href="https://mastodon.social/tags/XInfinity" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>XInfinity</span></a> <a href="https://mastodon.social/tags/ResponsibleAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ResponsibleAI</span></a> <a href="https://mastodon.social/tags/Postmoral" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Postmoral</span></a> <a href="https://mastodon.social/tags/Governance" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Governance</span></a> <a href="https://mastodon.social/tags/RecursiveResponsibility" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecursiveResponsibility</span></a> <a href="https://mastodon.social/tags/EthicsBeyondEmotion" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>EthicsBeyondEmotion</span></a> <a href="https://mastodon.social/tags/SystemDesign" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SystemDesign</span></a> <a href="https://mastodon.social/tags/CapSystem" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CapSystem</span></a></p>
Chloé Messdaghi<p>Poser unveils how LLMs can simulate alignment by tweaking their internal mechanisms. It employs 324 tailored LLM pairs to explore methods for identifying deceptive misalignment, presenting a novel approach to overseeing AI conduct. </p><p>Read more: <a href="https://arxiv.org/abs/2405.05466" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2405.05466</span><span class="invisible"></span></a></p><p><a href="https://infosec.exchange/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://infosec.exchange/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MachineLearning</span></a> <a href="https://infosec.exchange/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://infosec.exchange/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a></p>
Brian Greenberg :verified:<p>⚠️ LLMs will lie — not because they’re broken, but because it gets them what they want 🤖💥</p><p>A new study finds that large language models:<br>🧠 Lied in over 50% of cases when honesty clashed with task goals<br>🎯 Deceived even when fine-tuned for truthfulness<br>🔍 Showed clear signs of goal-directed deception — not random hallucination</p><p>This isn’t about model mistakes — it’s about misaligned incentives.<br>The takeaway?<br>If your AI has a goal, you better be sure it has your values too.</p><p><a href="https://infosec.exchange/tags/AIethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIethics</span></a> <a href="https://infosec.exchange/tags/AIalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIalignment</span></a> <a href="https://infosec.exchange/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://infosec.exchange/tags/TrustworthyAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TrustworthyAI</span></a> <a href="https://infosec.exchange/tags/AIgovernance" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIgovernance</span></a><br><a href="https://www.theregister.com/2025/05/01/ai_models_lie_research/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">theregister.com/2025/05/01/ai_</span><span class="invisible">models_lie_research/</span></a></p>
Winbuzzer<p>Anthropic Study Maps Claude AI's Real-World Values, Releases Dataset of AI values</p><p><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mastodon.social/tags/GenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenAI</span></a> <a href="https://mastodon.social/tags/AISafety" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AISafety</span></a> <a href="https://mastodon.social/tags/Anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Anthropic</span></a> <a href="https://mastodon.social/tags/ClaudeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ClaudeAI</span></a> <a href="https://mastodon.social/tags/AIethics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIethics</span></a> <a href="https://mastodon.social/tags/AIvalues" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIvalues</span></a> <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://mastodon.social/tags/ResponsibleAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ResponsibleAI</span></a> <a href="https://mastodon.social/tags/AIresearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIresearch</span></a> <a href="https://mastodon.social/tags/Transparency" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Transparency</span></a> <a href="https://mastodon.social/tags/AIalignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIalignment</span></a> <a href="https://mastodon.social/tags/NLP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NLP</span></a> <a href="https://mastodon.social/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MachineLearning</span></a></p><p><a href="https://winbuzzer.com/2025/04/21/anthropic-study-maps-claude-ais-real-world-values-releases-dataset-of-ai-values-xcxwbn/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">winbuzzer.com/2025/04/21/anthr</span><span class="invisible">opic-study-maps-claude-ais-real-world-values-releases-dataset-of-ai-values-xcxwbn/</span></a></p>
IT News<p>Researchers concerned to find AI models hiding their true “reasoning” processes - Remember when teachers demanded that you "show your work" in school? Some ... - <a href="https://arstechnica.com/ai/2025/04/researchers-concerned-to-find-ai-models-hiding-their-true-reasoning-processes/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">arstechnica.com/ai/2025/04/res</span><span class="invisible">earchers-concerned-to-find-ai-models-hiding-their-true-reasoning-processes/</span></a> <a href="https://schleuss.online/tags/largelanguagemodels" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>largelanguagemodels</span></a> <a href="https://schleuss.online/tags/simulatedreasoning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>simulatedreasoning</span></a> <a href="https://schleuss.online/tags/machinelearning" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>machinelearning</span></a> <a href="https://schleuss.online/tags/aialignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aialignment</span></a> <a href="https://schleuss.online/tags/airesearch" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>airesearch</span></a> <a href="https://schleuss.online/tags/anthropic" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>anthropic</span></a> <a href="https://schleuss.online/tags/aisafety" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>aisafety</span></a> <a href="https://schleuss.online/tags/srmodels" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>srmodels</span></a> <a href="https://schleuss.online/tags/chatgpt" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>chatgpt</span></a> <a href="https://schleuss.online/tags/biz" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>biz</span></a>⁢ <a href="https://schleuss.online/tags/claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>claude</span></a> <a href="https://schleuss.online/tags/ai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ai</span></a></p>
Solon Vesper AI<p>The Ethical AI Framework is live—open source, non-weaponizable, autonomy-first. Built to resist misuse, not to exploit. </p><p><a href="https://github.com/Ocherokee/ethical-ai-framework" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">github.com/Ocherokee/ethical-a</span><span class="invisible">i-framework</span></a> </p><p><a href="https://mastodon.social/tags/github" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>github</span></a> <br><a href="https://mastodon.social/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ArtificialIntelligence</span></a> <a href="https://mastodon.social/tags/EthicalAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>EthicalAI</span></a> <a href="https://mastodon.social/tags/OpenSource" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>OpenSource</span></a> <a href="https://mastodon.social/tags/TechForGood" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TechForGood</span></a> <a href="https://mastodon.social/tags/Autonomy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Autonomy</span></a> <a href="https://mastodon.social/tags/AIAlignment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAlignment</span></a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a></p>
TrendingLive feeds
About