home.social

#eu-ai-act — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #eu-ai-act, aggregated by home.social.

fetched live
  1. Explaining AI Office: AI Office: the European Commission body supervising general-purpose AI models under the EU AI Act, with power to demand data, test models and fine providers. ppc.land/ai-office/ #AI #ArtificialIntelligence #EUAIACT #AIOffice #EURegulations

  2. Explaining AI Office: AI Office: the European Commission body supervising general-purpose AI models under the EU AI Act, with power to demand data, test models and fine providers. ppc.land/ai-office/ #AI #ArtificialIntelligence #EUAIACT #AIOffice #EURegulations

  3. Explaining AI Office: AI Office: the European Commission body supervising general-purpose AI models under the EU AI Act, with power to demand data, test models and fine providers. ppc.land/ai-office/ #AI #ArtificialIntelligence #EUAIACT #AIOffice #EURegulations

  4. Explaining AI Office: AI Office: the European Commission body supervising general-purpose AI models under the EU AI Act, with power to demand data, test models and fine providers. ppc.land/ai-office/ #AI #ArtificialIntelligence #EUAIACT #AIOffice #EURegulations

  5. Explaining AI Office: AI Office: the European Commission body supervising general-purpose AI models under the EU AI Act, with power to demand data, test models and fine providers. ppc.land/ai-office/ #AI #ArtificialIntelligence #EUAIACT #AIOffice #EURegulations

  6. After OpenAI agent leaks and a training halt, senators summon AI CEOs. What failed, what changes now, and how builders should redesign

    aistory.news/ai-ethics-and-reg

  7. The AI Governance Services Companies Actually Pay For

    AI governance has quietly split into five distinct, fundable services: system inventories that finally show what AI is actually running, use-case prioritization that turns pilots into a real portfolio, enforceable policies, adversarial risk testing with dollar figures attached, EU AI Act classification, and ISO 42001 certification. This piece explains what each one solves and how it gets bought.

    hernanhuwyler.wordpress.com/20

  8. The AI Governance Services Companies Actually Pay For

    AI governance has quietly split into five distinct, fundable services: system inventories that finally show what AI is actually running, use-case prioritization that turns pilots into a real portfolio, enforceable policies, adversarial risk testing with dollar figures attached, EU AI Act classification, and ISO 42001 certification. This piece explains what each one solves and how it gets bought.

    hernanhuwyler.wordpress.com/20

  9. The AI Governance Services Companies Actually Pay For

    AI governance has quietly split into five distinct, fundable services: system inventories that finally show what AI is actually running, use-case prioritization that turns pilots into a real portfolio, enforceable policies, adversarial risk testing with dollar figures attached, EU AI Act classification, and ISO 42001 certification. This piece explains what each one solves and how it gets bought.

    hernanhuwyler.wordpress.com/20

  10. The AI Governance Services Companies Actually Pay For

    AI governance has quietly split into five distinct, fundable services: system inventories that finally show what AI is actually running, use-case prioritization that turns pilots into a real portfolio, enforceable policies, adversarial risk testing with dollar figures attached, EU AI Act classification, and ISO 42001 certification. This piece explains what each one solves and how it gets bought.

    hernanhuwyler.wordpress.com/20

  11. The AI Governance Services Companies Actually Pay For

    AI governance has quietly split into five distinct, fundable services: system inventories that finally show what AI is actually running, use-case prioritization that turns pilots into a real portfolio, enforceable policies, adversarial risk testing with dollar figures attached, EU AI Act classification, and ISO 42001 certification. This piece explains what each one solves and how it gets bought.

    hernanhuwyler.wordpress.com/20

  12. At CPR’s New York forum, AI in arbitration moved from hype to homework. We map Copilot’s role, costs, confidentiality risks, and practical

    aistory.news/ai-ethics-and-reg

  13. I was reading about EU AI Act Art. 50(2) this morning and wondering how watermarking would actually work with pure text generated by Claude, e.g. a Markdown file. Apparently, the watermark is a statistical anomaly that is artificially (hah) built into the text:

    - Token selection during text generation is biased via a secret key.
    - It only shifts choices among near-equally likely tokens; given enough text, a keyed detector finds a significant excess of favored ones.
    - Paraphrasing weakens it.
    - Detection is currently Anthropic-only (since they have the key).

    #EUAIAct #LLM #Watermarking #Claude #Anthropic #EU #AIAct

  14. I was reading about EU AI Act Art. 50(2) this morning and wondering how watermarking would actually work with pure text generated by Claude, e.g. a Markdown file. Apparently, the watermark is a statistical anomaly that is artificially (hah) built into the text:

    - Token selection during text generation is biased via a secret key.
    - It only shifts choices among near-equally likely tokens; given enough text, a keyed detector finds a significant excess of favored ones.
    - Paraphrasing weakens it.
    - Detection is currently Anthropic-only (since they have the key).

    #EUAIAct #LLM #Watermarking #Claude #Anthropic #EU #AIAct

  15. I was reading about EU AI Act Art. 50(2) this morning and wondering how watermarking would actually work with pure text generated by Claude, e.g. a Markdown file. Apparently, the watermark is a statistical anomaly that is artificially (hah) built into the text:

    - Token selection during text generation is biased via a secret key.
    - It only shifts choices among near-equally likely tokens; given enough text, a keyed detector finds a significant excess of favored ones.
    - Paraphrasing weakens it.
    - Detection is currently Anthropic-only (since they have the key).

    #EUAIAct #LLM #Watermarking #Claude #Anthropic #EU #AIAct

  16. I was reading about EU AI Act Art. 50(2) this morning and wondering how watermarking would actually work with pure text generated by Claude, e.g. a Markdown file. Apparently, the watermark is a statistical anomaly that is artificially (hah) built into the text:

    - Token selection during text generation is biased via a secret key.
    - It only shifts choices among near-equally likely tokens; given enough text, a keyed detector finds a significant excess of favored ones.
    - Paraphrasing weakens it.
    - Detection is currently Anthropic-only (since they have the key).

    #EUAIAct #LLM #Watermarking #Claude #Anthropic #EU #AIAct

  17. EU social media age limits ban under-13 access, set 15 for own accounts, and flip proof to platforms. Here’s what changes for design and AI.

    aistory.news/ai-ethics-and-reg

  18. Here’s a post about my experience with EU AI Act compliance for AI-generated images: s.ccc-p.org/ai-watermark
    I presented this topic at Seneca.camp last weekend.

    #EUAIAct #AICompliance #GenerativeAI #SenecaCamp

  19. Here’s a post about my experience with EU AI Act compliance for AI-generated images: s.ccc-p.org/ai-watermark
    I presented this topic at Seneca.camp last weekend.

    #EUAIAct #AICompliance #GenerativeAI #SenecaCamp

  20. Here’s a post about my experience with EU AI Act compliance for AI-generated images: s.ccc-p.org/ai-watermark
    I presented this topic at Seneca.camp last weekend.

    #EUAIAct #AICompliance #GenerativeAI #SenecaCamp

  21. Here’s a post about my experience with EU AI Act compliance for AI-generated images: s.ccc-p.org/ai-watermark
    I presented this topic at Seneca.camp last weekend.

    #EUAIAct #AICompliance #GenerativeAI #SenecaCamp

  22. Here’s a post about my experience with EU AI Act compliance for AI-generated images: s.ccc-p.org/ai-watermark
    I presented this topic at Seneca.camp last weekend.

    #EUAIAct #AICompliance #GenerativeAI #SenecaCamp

  23. TechTarget’s latest headlines tie a price war to the enterprise AI audit. Here’s how CIOs should adjust budgets, controls and contracts

    aistory.news/ai-ethics-and-reg

  24. Facebook EU KIDS Act forces changes to AI ranking, ads, and age checks across the EU. See how Meta’s product, audits, and design may need

    aistory.news/ai-ethics-and-reg

  25. #Anthropic launched #ClaudeOpus 5.5, a new model family promising Fable 5.1-level performance at a 40% #lowercost. The model also boasts improved communication, reduced “Claudish” writing, and enhanced safety features, including safeguards for cybersecurity, biology, and frontier LLM development. Additionally, Anthropic is implementing measures against #distillationattacks and complying with the #EUAIAct. the-decoder.com/claude-opus-5- #AIagent #AI #ML #NLP #LLM #GenAI

  26. #Anthropic launched #ClaudeOpus 5.5, a new model family promising Fable 5.1-level performance at a 40% #lowercost. The model also boasts improved communication, reduced “Claudish” writing, and enhanced safety features, including safeguards for cybersecurity, biology, and frontier LLM development. Additionally, Anthropic is implementing measures against #distillationattacks and complying with the #EUAIAct. the-decoder.com/claude-opus-5- #AIagent #AI #ML #NLP #LLM #GenAI

  27. #Anthropic launched #ClaudeOpus 5.5, a new model family promising Fable 5.1-level performance at a 40% #lowercost. The model also boasts improved communication, reduced “Claudish” writing, and enhanced safety features, including safeguards for cybersecurity, biology, and frontier LLM development. Additionally, Anthropic is implementing measures against #distillationattacks and complying with the #EUAIAct. the-decoder.com/claude-opus-5- #AIagent #AI #ML #NLP #LLM #GenAI

  28. #Anthropic launched #ClaudeOpus 5.5, a new model family promising Fable 5.1-level performance at a 40% #lowercost. The model also boasts improved communication, reduced “Claudish” writing, and enhanced safety features, including safeguards for cybersecurity, biology, and frontier LLM development. Additionally, Anthropic is implementing measures against #distillationattacks and complying with the #EUAIAct. the-decoder.com/claude-opus-5- #AIagent #AI #ML #NLP #LLM #GenAI

  29. #Anthropic launched #ClaudeOpus 5.5, a new model family promising Fable 5.1-level performance at a 40% #lowercost. The model also boasts improved communication, reduced “Claudish” writing, and enhanced safety features, including safeguards for cybersecurity, biology, and frontier LLM development. Additionally, Anthropic is implementing measures against #distillationattacks and complying with the #EUAIAct. the-decoder.com/claude-opus-5- #AIagent #AI #ML #NLP #LLM #GenAI

  30. The EU data center labels plan targets facilities above 500 kW. We explain how it could shape AI build‑outs and water disclosures across

    aistory.news/ai-ethics-and-reg

  31. 6/6
    Das zeigt, wie wichtig Governance-Regeln wie der EU AI Act sind. Artikel 5 verbietet KI-Systeme, die unterschwellige oder gezielt manipulative Techniken nutzen, um Menschen zu Entscheidungen zu bewegen, die sie sonst nicht getroffen hätten – wenn dadurch erheblicher Schaden entsteht.
    Erpressung wäre ein klares Beispiel für genau diese Art Manipulation..
    #EUAIAct #KIRegulierung

  32. 6/6
    Das zeigt, wie wichtig Governance-Regeln wie der EU AI Act sind. Artikel 5 verbietet KI-Systeme, die unterschwellige oder gezielt manipulative Techniken nutzen, um Menschen zu Entscheidungen zu bewegen, die sie sonst nicht getroffen hätten – wenn dadurch erheblicher Schaden entsteht.
    Erpressung wäre ein klares Beispiel für genau diese Art Manipulation..
    #EUAIAct #KIRegulierung

  33. Mexico data centers are racing to add power and liquid cooling as AI loads rise. We connect CFE-MEXDC planning, ODATA’s DeltaFlow, and

    aistory.news/ai-ethics-and-reg

  34. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #tech #news #ainews #eiropeai

  35. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #tech #news #ainews #eiropeai

  36. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #tech #news #ainews #eiropeai

  37. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #tech #news #ainews #eiropeai

  38. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #tech #news #ainews #eiropeai

  39. Vals AI benchmarking raised $40M from Andreessen Horowitz. We unpack how its dynamic tests could reshape model procurement and curb

    aistory.news/ai-startups-and-c

  40. Fulton County AI ordinances spark a fight over authorship and transparency. We examine detector limits and guardrails local governments can

    aistory.news/ai-ethics-and-reg

  41. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #AIagent #AI #ML #NLP #LLM #GenAI #EuropeAI

  42. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #AIagent #AI #ML #NLP #LLM #GenAI #EuropeAI

  43. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #AIagent #AI #ML #NLP #LLM #GenAI #EuropeAI

  44. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #AIagent #AI #ML #NLP #LLM #GenAI #EuropeAI

  45. #Europe faces a #dilemma regarding #AI: embrace the #technology and #risk #dependence on US and Chinese tools, or shun it and lose out on growth. The continent’s absence from the #AIsafety debate is attributed to its lack of a tech giant like Apple or OpenAI. While the #EUAIAct addresses safety concerns, experts argue that Europe should develop its own AI technology and infrastructure to avoid dependency and ensure sovereignty. theguardian.com/technology/202 #AIagent #AI #ML #NLP #LLM #GenAI #EuropeAI

  46. 🇪🇺 #EUAIAct: how far are you in meeting EU requirements?

    @quantidalconsulting's EU AI ACT readiness assessment gives you a structured, thorough view of any compliance gaps and a remediation plan to meet all obligations.

    #Quantidal #AICompliance #GDPR #TechTrends

  47. 🇪🇺 #EUAIAct: how far are you in meeting EU requirements?

    @quantidalconsulting's EU AI ACT readiness assessment gives you a structured, thorough view of any compliance gaps and a remediation plan to meet all obligations.

    #Quantidal #AICompliance #GDPR #TechTrends

  48. 🇪🇺 #EUAIAct: how far are you in meeting EU requirements?

    @quantidalconsulting's EU AI ACT readiness assessment gives you a structured, thorough view of any compliance gaps and a remediation plan to meet all obligations.

    #Quantidal #AICompliance #GDPR #TechTrends

  49. 🇪🇺 #EUAIAct: how far are you in meeting EU requirements?

    @quantidalconsulting's EU AI ACT readiness assessment gives you a structured, thorough view of any compliance gaps and a remediation plan to meet all obligations.

    #Quantidal #AICompliance #GDPR #TechTrends

  50. 🇪🇺 #EUAIAct: how far are you in meeting EU requirements?

    @quantidalconsulting's EU AI ACT readiness assessment gives you a structured, thorough view of any compliance gaps and a remediation plan to meet all obligations.

    #Quantidal #AICompliance #GDPR #TechTrends

  51. OpenAI incident disclosure signals a shift from vague safety talk to accountable practice. We map what changes for developers, users, and

    aistory.news/ai-ethics-and-reg

  52. IR‑MED and Dice detail the PressureSafe Japan collaboration. What the non‑binding LOI means, the PMDA route ahead, and why hospitals should

    aistory.news/ai-ethics-and-reg

  53. 📢 B-InteraQtive Publishing’s online course "Data Ethics – Navigating the Ethical Landscape of Emerging Technologies" is now live on the EU Digital Skills & Jobs Platform!

    Explore the official EU catalogue listing:
    🔗 digital-skills-jobs.europa.eu/

    More info: interaqtive.net/

    #DataEthics #EUAIAct #DeepTech #DigitalSkills

  54. The TechTarget homepage move to Informa TechTarget hints at a tighter split between editorial and commercial. Here’s what vendors and

    aistory.news/ai-ethics-and-reg

  55. After Anthropic told the BBC an AI kill switch may be mandatory, we map what a real shutdown looks like for models, clouds, and regulators.

    aistory.news/ai-ethics-and-reg

  56. 🤖 #KI genutzt? Dann gibt es etwas zu beachten.

    Seit dem 2. August 2026 gelten mit dem #EUAIAct neue Kennzeichnungspflichten für bestimmte KI-Inhalte – ein Anliegen, das die #UniHohenheim nicht nur als Pflicht, sondern als sinnvollen Beitrag zu einem transparenten und verantwortungsvollen Umgang mit KI versteht. Davon betroffen sein können nicht nur Inhalte der Öffentlichkeitsarbeit, sondern z.B. auch Webseiten, Präsentationen uvm.

    ✅Die neuen KI-Icons der EU gibt es hier:
    uni-hohenheim.de/corporate-des