The European Commission asked for input as they discussed proposing targeted amendments to the EU AI Act through its Digital Omnibus initiative, seeking to streamline certain procedures and reduce regulatory burdens. We provided technical input to inform this process.
Our submission addresses critical implementation challenges including high-risk AI classification, General Purpose AI provider obligations, and standards development timelines. Drawing on our participation in the EU’s Code of Practice and CEN-CENELEC standardization efforts, we recommend clearer guidance on scientific research exemptions, refined classification criteria for Annex I high-risk systems to prevent over-regulation of low-risk AI components and clearer allocation of responsibilities in complex GPAI value chains to avoid regulatory gaps and unintended incentives to unsafe or non-compliant behaviour.