#eu-ai-act — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #eu-ai-act, aggregated by home.social.
-
🧬🖥️ How do we keep #AI in #genomics safe?
A new paper coauthored by #RKI researchers explores risks such as data bias, quality issues, and #DualUse potential in #GenAI. It highlights how these risks emerge along the innovation pipeline and outlines strategies for safer, responsible use.
-
🎯Die Human‑Review‑Lücke im AI Act – Transparenzpflichten erklärt 🔗https://pwa.ist/human-review-lucke-eu-ai-act #ai #aiacttransparenz #aiact #euaiact #transparenz
-
FYI: EU AI Act gets its first real haircut - high-risk deadlines pushed to 2027: EU Council and Parliament today agreed to simplify AI Act rules under Digital Omnibus VII, fixing new high-risk deadlines and adding a ban on AI-generated CSAM. https://ppc.land/eu-ai-act-gets-its-first-real-haircut-high-risk-deadlines-pushed-to-2027/ #EUAIAct #ArtificialIntelligence #AIEthics #DataProtection #DigitalPolicy
-
ICYMI: EU AI Act gets its first real haircut - high-risk deadlines pushed to 2027: EU Council and Parliament today agreed to simplify AI Act rules under Digital Omnibus VII, fixing new high-risk deadlines and adding a ban on AI-generated CSAM. https://ppc.land/eu-ai-act-gets-its-first-real-haircut-high-risk-deadlines-pushed-to-2027/ #EUAIAct #ArtificialIntelligence #AIRegulation #DigitalOmnibus #HighRiskAI
-
Die EU verbietet ab Dezember 2026 KI-Modelle zur Erstellung sexualisierter Deepfakes ohne Einwilligung. Die Regulierung fokussiert sich direkt auf Entwickler und Werkzeuge statt nur auf Anwender. Gleichzeitig verschiebt sich die Wasserzeichenpflicht für Anbieter wie ChatGPT auf Ende 2026. Open-Source-Entwickler müssen künftig Filter implementieren.
-
EU AI Act gets its first real haircut - high-risk deadlines pushed to 2027: EU Council and Parliament today agreed to simplify AI Act rules under Digital Omnibus VII, fixing new high-risk deadlines and adding a ban on AI-generated CSAM. https://ppc.land/eu-ai-act-gets-its-first-real-haircut-high-risk-deadlines-pushed-to-2027/ #EUAIact #ArtificialIntelligence #AIlaw #DigitalOmnibus #AIethics
-
The EU AI Act makes human oversight a core requirement for high-risk AI systems—designed to prevent risks to fundamental rights and ensure human control. (Artificial Intelligence Act)
But what counts as “oversight” in practice?
If humans can’t meaningfully understand or challenge outputs, oversight risks becoming procedural.
New article:
https://medium.com/@biytelum/the-eu-ai-act-and-the-illusion-of-human-oversight-5ed38f40ff30 -
The EU AI Act classifies certain AI in asylum as “high-risk”—including systems that assist public authorities in examining applications.
Not all uses fall there.
But as AI influences decisions, expectations around oversight and traceability increase.
-
One of the hardest parts of implementing the EU AI Act:
Human oversight.
It’s easy to say a human is “in the loop.”
It’s harder to design systems where that oversight is actually meaningful:
• understanding outputs
• identifying errors
• intervening effectivelyThat’s a design problem—not just a policy one.
-
FYI: IAB Italia's AI white paper maps the future of marketing in Italy: IAB Italia published its AI white paper on April 13, covering generative AI, the EU AI Act, targeting, hyper-personalisation and the future of Italian marketing https://ppc.land/iab-italias-ai-white-paper-maps-the-future-of-marketing-in-italy/ #Marketing #ArtificialIntelligence #AIMarketing #GenerativeAI #EUAIACT
-
Man hat gerade das Gefühl, Europa verhandelt nicht nur Regeln für KI, sondern so etwas wie Haltung.
Es ist fast so, als würde man versuchen, etwas sehr Schnelles in eine Form zu bringen, ohne es dabei ganz zu verlieren.
Und vielleicht ist es genau das: nicht schneller zu werden als die Technik, sondern bewusst genug, um noch mitreden zu können.Ein bisschen Kontrolle.
Ein bisschen Zweifel.
Ein bisschen Europa 🙂#AI #EUAIAct #Digitalpolitik #TechUndGesellschaft #Verantwortung #Mastodon
-
«Schatten-HR — Bizarre Recruiting-Tool Gefahr bis zu 35 Millionen Euro:
Warum Low-Code-Tools in der Schatten-HR ohne IT-Segen gegen den EU AI Act verstoßen und Millionenstrafen riskieren.»Zu viele Verantwortliche sind leichtgläubig und wissen nicht mal selber was ihre Aufgabe ist. KI erhöht dessen naive Verhalten noch dazu.
🫤 https://www.it-daily.net/it-sicherheit/cloud-security/schatten-hr
#lawcode #itsicherheit #euaiact #aiact #ki #ai #eu #schattenseiten #hr #bizzarr
-
📢 Fantastic news from the Digital Science Center! 📢
The open‑access paper “Maximal Transparency for Online Recommender Systems” is out in Philosophy & Technology. A truly interdisciplinary effort across philosophy, bioinformatics, mathematics, computer science, and law.
Read the full article here: https://link.springer.com/article/10.1007/s13347-026-01084-7
#RecommenderSystems #Transparency #AIEthics #OpenAccess #InterdisciplinaryResearch #Philosophy #Technology #EUAIAct
1/5 -
German Chancellor Merz pushes for separate EU rules for industrial AI before the AI Act's August deadline. His argument: factory automation systems shouldn't face the same regulations as consumer chatbots. The timing matters as Germany targets 4x more AI processing capacity by 2030. Key question remains how to balance lighter rules with worker safety protections. #EUAIAct #IndustrialAI #AIRegulation https://www.implicator.ai/merz-seeks-industrial-ai-carve-out-before-eu-rules-bite-in-august/
-
Last week at the University of Applied Sciences St. Pölten, the CERTAIN Project hosted a technical workshop and presented at the SAINT Conference 2026.
Our focus: building trust directly into the ML pipeline.
As AI evolves, aligning with the EU AI Act and General Data Protection Regulation is a major challenge. CERTAIN tackles this by turning legal requirements into machine-readable processes.
-
“Human in the loop” is often treated as a safeguard in AI systems.
But under GDPR and the EU AI Act, the key question is whether that oversight is meaningful.
If reviewers rely on opaque outputs or constrained processes, the safeguard may be more formal than real.
-
New blog: AI Governance and Regulatory Compliance — Ensuring Ethical and Non-Discriminatory AI Use. Dive into practical guidance on aligning AI systems with frameworks like the EU AI Act and NIST, and learn why automated compliance is becoming a business imperative for transparency and fairness. Read more: https://wix.to/3ZLfsNO
-
Benvenuti nell'era dell'AI slop: perché la pigrizia algoritmica ci sta allontanando dai social
Scopri come l'AI slop sta influenzando i social e come possiamo difenderci! Leggi l'articolo per capire come restare autentici in un mondo digitale sempre più standardizzato. #AIslop #autenticità #socialmedia
@tecnologia -
Third post this week on compliance, but this one is on the clock.
The EU AI Act deadline is August 2, 2026. That's four months away. And unlike GDPR, which most SMBs at least pretend to take seriously, AI governance is currently a blank page at almost every mid-market company I talk to.
Here's the uncomfortable part: you're already a "deployer" under the Act. If your team uses ChatGPT, Copilot, Gemini, or any AI feature buried inside your CRM or HR platform, you have obligations. Not because you built the AI — because you chose to use it.
The obligations themselves are manageable. Inventory your AI services. Classify each one by risk. Document human oversight for anything that affects people. Integrate it with your existing ROPA. None of that is hard. What's hard is that almost nobody has started.
One thing to know that often gets missed: the AI literacy requirement (Article 4) already took effect in February 2025. If your staff use AI tools, they're already supposed to have training. That deadline is in the past, not the future.
I wrote a guide on what SMBs actually need to do. Four steps, mapped against the August 2026 deadline, written for people who don't have a Chief AI Officer.
https://readmodel.com/blog/article.php?slug=ai-governance-guide
#EUAIAct #AIGovernance #GDPR #readmodel #locaverdi -
Der EU AI Act ist in Teilen in Kraft. Was das konkret bedeutet:
KI-Systeme, die mit Menschen interagieren, müssen sich als KI zu erkennen geben. Immer.
Das gilt auch für KI-Telefonassistenten, Chatbots und automatisierte Kundenkommunikation.
Ich finde das richtig. Transparenz war schon immer gut für das Vertrauen. Siehe Influencer Marketing.
-
AI Act: Who is actually a deployer, and what must they do today?
Alarmist posts keep circulating — fabricated deadlines, non-existent penalties, overstated obligations.
My new analysis: what "deployer" means, real timelines, actual penalties, the Commission's position on AI agents, and 6 practical steps.
Regulatory precision is not an academic luxury. It is a professional responsibility.👉 https://www.nicfab.eu/en/posts/deployer-ai-agents/
#AIAct #EUAIAct #AIGovernance #ArtificialIntelligence #Compliance #TechLaw #AI
-
Ti sei accorto che i feed dei social sono pieni di immagini e testi tutti uguali? Benvenuti nell'era dell'AI Slop. In questo video esploriamo perché la "sbobba" digitale sta invadendo il web, come i grandi brand (Lego, Dove, Nik
@tecnologia
https://www.youtube.com/watch?v=6acKun3xSQI#AISlop #IntelligenzaArtificiale #Staipa #Divulgazione #EUAIACT #Creatività #UmanesimoDigitale #PensieroCritico #Divulgazione #DivulgazioneInformatica #Informatica #DivulgazioneTecnologica
-
AI agents are becoming operational infrastructure—but they’re also creating cross-border regulatory exposure.
For Latin American companies, GDPR and the EU AI Act can apply based on who you interact with, not where you’re based.
That threshold is easier to cross than it looks.
More here:
-
EU-Institutionen untersagen ihren Presseteams die Nutzung vollständig KI-generierter Bilder und Videos in der offiziellen Kommunikation. KI-Modelle dürfen nur noch zur qualitativen Aufwertung echter Aufnahmen eingesetzt werden, etwa für Upscaling oder Rauschunterdrückung. Experten kritisieren das Verbot und fordern einen transparenten Umgang durch sichtbare Wasserzeichen gemäß dem europäischen AI Act.
#EUAIAct #GenerativeAI #Deepfakes #Datenschutz #News
https://www.all-ai.de/news/news26/eu-ki-bilder-verbot -
Enforce at the Irish Council for Civil Liberties is hiring an expert in #AI.
Full-time role. Remote in the EU.
Apply early. Deadline 5pm Dublin time on 15 April 2026.
#AIPolicy #EUAIAct #DigitalRights #TechPolicy
https://www.iccl.ie/digital-data/enforce-is-hiring-ai-expert/