Stop before you update your compliance calendar.
According to available legal analysis from the Austrian Economic Chamber (WKO), the European Parliament has voted to delay the application of certain high-risk AI system rules under the EU AI Act. The vote is real. The delay is not yet enacted. Council of the European Union approval is still required for any timeline change to take effect.
That distinction is not administrative fine print. It’s the operative compliance fact.
What the Parliament Vote Reportedly Proposes
For high-risk AI systems specifically listed in the EU AI Act regulation, the proposed new application date is reportedly December 2, 2027, pushed back from the current August 2026 milestone. For AI systems covered by EU sectoral legislation on safety and market surveillance, the proposed date moves further, to reportedly August 2, 2028.
A third date applies to a narrower obligation: providers may reportedly have until November 2, 2026, to comply with rules on watermarking AI-created audio, image, video, and text content. That deadline, notably, is earlier than the high-risk system delays would provide.
The Digital Omnibus on AI is the legislative vehicle. The Council reportedly adopted its general approach on March 13, 2026. EP IMCO and LIBE Committees reportedly adopted their joint position on March 18, 2026. The EP committee position reportedly includes a ban on AI systems designed to generate non-consensual intimate images.
Why the Qualifier Matters
TJS has previously published on the August 2, 2026 enforcement timeline and the 120-day operational compliance countdown. Those pieces reflect the currently enacted EU AI Act deadlines. They remain accurate until Council approval of any delay is confirmed.
Organizations that have built compliance programs toward August 2026 milestones should not pause work based on this vote alone. Compliance preparation that gets done early creates less risk than preparation that gets delayed and then needs to accelerate if Council approval doesn’t materialize on schedule.
The sourcing for this item is also an important transparency note. The two primary sources cited by the Wire were found to be content mismatches, they resolve to journalism-and-society research pages, not EU legislative content. The claims here rest on a single legal analysis source (WKO) whose full content was not available for direct review. All specific dates in this brief are qualified as reported and proposed, not confirmed.
What to Watch
– Council of the European Union vote on the Digital Omnibus on AI, this is the trigger event that would confirm or reject the proposed delay – Any official European Parliament press release confirming the vote and proposed dates – Whether the watermarking November 2026 deadline survives Council deliberations unchanged – Organizations with August 2026 compliance programs: monitor Council proceedings, don’t stand down
TJS Synthesis
If confirmed, the proposed delay would be the most significant EU AI Act timeline shift since the regulation passed. December 2027 for listed high-risk systems gives organizations substantially more runway, but “if confirmed” is doing heavy lifting in that sentence. The compliance professional’s posture right now is: continue current preparations, log this development as a material pending variable, and watch for Council action. Don’t build a compliance strategy around a proposed date that still requires approval.
*Prior TJS coverage of the August 2, 2026 EU AI Act enforcement deadline remains operative until Council approval of any delay is officially confirmed.*