The European Observatory on AI Standards is an independent initiative that reads the EU AI Act's harmonised standards against the Regulation itself, in the open, while the drafts are still open to comment. It is run by people who sit inside the rooms where these standards are written.
The Observatory is a collaboration between organisations working at the point where AI regulation, technical standardisation and applied ethics meet. Its contributors are active in CEN-CENELEC JTC 21 and in EU research, so the analysis is close to the source. That proximity is also a conflict of interest, and we state it plainly: the same people who read these drafts in public also work on them, and in some cases sell related services.
Piccadilly Labs advises providers, deployers and conformity teams on the AI Act's harmonised standards and on getting conformity work in place before the high-risk obligations apply.
ICARO Lab, the research laboratory of the Icaro Foundation, studies how AI systems behave under pressure, across languages, and once models and agents interact. Its work in Rome, with Sapienza University and Dexai, includes the widely reported finding that reformulating a prohibited request as verse defeats the safety filters of frontier models.
Dexai — Artificial Ethics, founded in Rome in 2021, gives operational form to AI ethics principles: transparency, fairness, human oversight, and measurable compliance. It has led EU Horizon research projects and takes part in CEN-CENELEC and ISO/IEC AI standardisation.
BeEthical is the AI and data ethics consultancy of Enrico Panai, convenor of CEN-CENELEC JTC 21 WG4, editor within ISO/IEC JTC 1 SC 42 WG3, and president of the Association of AI Ethicists. It advises large organisations on responsible AI.
