EU AI Act compliance deadlines delayed until 2027 for high-risk systems

The extension provides immediate breathing room for streaming companies developing recommendation engines or content moderation tools, but it does not resolve the underlying lack of technical benchmarks. Within the streaming ecosystem, the delay highlights a growing gap between legislative intent and the practical availability of notified bodies capable of certifying complex algorithms. This regulatory uncertainty likely favors large-scale incumbents with the legal resources to navigate shifting requirements while potentially stifling smaller innovators. Watch for the AI Office to release specific data protection guidelines for the new Union-level sandbox to see if it attracts meaningful industry participation.

The EU AI Act's delay has intensified pressure on European standards bodies to deliver the technical specifications that will determine how high-risk AI systems are certified. CEN and CENELEC, the two organizations tasked with drafting harmonized standards under the regulation, have faced repeated criticism for slow progress. In June 2026, the European Commission acknowledged that none of the requested harmonized standards for high-risk AI systems had been finalized, leaving developers without the technical benchmarks needed to demonstrate conformity. For streaming platforms deploying recommendation engines or automated content moderation tools classified as high-risk, this standards vacuum means the extended timeline does not actually reduce compliance complexity. It merely postpones the moment when companies must prove their systems meet requirements that have not yet been fully defined.

The Digital Omnibus delay also intersects with broader European competitiveness concerns that Mario Draghi highlighted in his landmark report. Draghi's September 2024 report on European competitiveness called for simplifying AI regulation to prevent innovation from migrating to less restrictive jurisdictions, a recommendation that the European Commission has cited as motivation for the omnibus package. Meanwhile, Vera Lúcia Raposo, a legal scholar specializing in AI governance, has argued that the delay risks creating a two-tier system where large technology firms with dedicated compliance teams proceed on their own timelines while smaller developers face prolonged uncertainty about certification pathways. The streaming industry, which includes both major platforms and numerous mid-size European services, sits squarely in this tension between regulatory relief and unresolved technical obligations.

On the technical side, the absence of harmonized standards has pushed some streaming and media companies toward voluntary frameworks and third-party audits as interim compliance mechanisms. The AI Office published guidance in May 2026 encouraging companies to adopt internal risk assessment procedures aligned with the Act's Annex III categories, but legal experts note that such guidance carries no binding authority and may not satisfy future conformity assessment requirements. For streaming services using , content classification, or automated moderation at scale, the practical effect is a compliance strategy built on shifting sand. Companies must invest in documentation and risk management processes now, knowing that the specific technical thresholds they must ultimately meet remain undefined by CEN and CENELEC.