Can Europe watermark AI? What the AI Act really asks for

Kari Jaaskelainen

06 Nov 2025 — 1 min read

Europe's AI Act says makers of general-purpose AI must mark and detect AI-generated content - and those marks must be reliable, interoperable, effective, and robust. But what does that mean in practice?

A new paper by Thomas Souverain maps today's watermarking tech for large language models (LLMs) to the law's demands.

Clear taxonomy: Classifies watermarking by when it's applied: before/during/after training, and at next-token distribution vs. sampling.
Evaluation guide: Translates the Act's criteria into tests for robustness, detectability, and model quality; proposes three dimensions to assess interoperability.
Reality check: No current method meets all four standards; promising direction is watermarking embedded deep in model architecture.

Why it matters: Policymakers, developers, and auditors get a shared language and test plan to judge watermarking claims under the EU AI Act.

Paper: http://arxiv.org/abs/2511.03641v1

Paper: http://arxiv.org/abs/2511.03641v1

Register: https://www.AiFeta.com

#AIAct #Watermarking #LLM #AI #TrustworthyAI #EU #Safety

Automating GDPR Compliance: A Roadmap for Companies and Law Firms

GDPR compliance is more than checkboxes. A new roadmap from the Privatech project shows how automation and machine learning can help companies and law firms assess—and even generate—privacy compliance. * Shift the focus to data processors’ real workflows: drafting policies, mapping data uses, documenting decisions. * Break compliance into machine-ready

FPGAs for Faster, Leaner Deep Learning: A Review of CNN Accelerators

Deep learning drives image search, robots, and medical scans. Most systems lean on CPUs and GPUs. This review asks: what if we run convolutional neural networks (CNNs) on FPGAs—reconfigurable chips you can tailor to the model? * Why FPGAs: custom dataflows, low latency, and strong energy efficiency—great for cameras,

Dynamic-K: Recommendations That Know When to Stop

Most apps show a fixed number of “top” items—say 10 movies or 20 products—assuming there are always enough good options. But that’s not always true: sometimes there are few relevant items, or some users are extra picky. The result? Filler recommendations. Dynamic-K flips the script. Instead of

Teaching chatbots to stop contradicting themselves (DECODE)

Teaching chatbots to stop contradicting themselves Ever had a bot say one thing, then the opposite a few turns later? This study introduces DECODE—a new task and dataset for spotting contradictions in everyday conversations, drawn from both human-human and human-bot chats. * New data beats existing natural language inference (NLI)

Read more

Automating GDPR Compliance: A Roadmap for Companies and Law Firms

FPGAs for Faster, Leaner Deep Learning: A Review of CNN Accelerators

Dynamic-K: Recommendations That Know When to Stop

Teaching chatbots to stop contradicting themselves (DECODE)