Two individual rights entered the picture, and eight modules went back for review
Articles 85 and 86 of the EU AI Act have applied since 2 August 2026, and anchoring them sent one competency and eight LearnWize modules back for review on 24 August 2026.
- In force since
- 2 August 2026
- Reviewed and published
- 24 August 2026
- modules reviewed
- 8
- competencies touched
- 1
Enforcement is no longer only a matter between a supervisor and an organisation. An employee, a customer or a rejected applicant can go to the market surveillance authority directly, and can ask what an AI system contributed to a decision that affected them. That changes what your people need to recognise, so it changed our content.
What changed
- Article 85 gives anyone with reason to believe the regulation was breached a route to the competent market surveillance authority. That route does not run through the organisation.
- Article 86 lets a person affected by a decision a deployer took on the output of an Annex III high-risk system ask for an explanation of the role that system played and of the main elements of the decision.
- Both have applied since 2 August 2026 and both were added to the published implementation graph in release 2.1.0.
What it means for your people
- Your first sign of a problem may be a call from a supervisor rather than an internal complaint, so the people closest to the decision need to recognise the situation early.
- A hiring manager, a claims handler or a case worker can be asked to explain what the system contributed. That is a conversation skill, not a legal filing.
- Recording who was trained on this, and when, is the part that is easy to arrange in advance and impossible to arrange afterwards.
What we changed in the platform
- Anchored both obligations to the enforcement competency in the anchor file, pinned to graph release 2.1.0 with the payload hash of each object version.
- Reviewed the eight modules that serve that competency across all three depth levels.
- Left every other competency untouched, because nothing else in this change affects them. A review that touches everything proves nothing.
Obligations this entry is anchored to
Article 85: right to lodge a complaint with the market surveillance authority
praxikon:eu:ai-act:obligation:article-85-right-to-complain
sha256 721db6d8aebb1ea9c2681a8ed7ba0daa59165fc2d8f3dc8d8974695ff70869fe
Article 86: right to an explanation of individual decision-making
praxikon:eu:ai-act:obligation:article-86-right-to-explanation
sha256 fdff4b46310b77efe7995b1f93427a199d0ce4ca2d3d59ce4cb9db0cc640944c
Each obligation carries the payload hash of the object version this entry was written against, so the claim can be verified rather than trusted. Implementation graph release: 2.1.0.
Questions people ask about this change
- Does Article 86 apply to every AI system we use?
- No. It concerns decisions a deployer takes on the output of a high-risk system listed in Annex III. The scope is narrower than most summaries suggest, which is exactly why role-level recognition matters more than a general awareness session.
- Why does a change in the law send training modules back for review?
- Because a course silently goes stale the day the law moves. Every LearnWize competency is anchored to the obligations it teaches, so a material change in one of those obligations names the competencies and modules that need a human to look at them again.
Official sources
See which of your roles this reaches
The 5-minute scan maps your AI use to the people around it and shows where evidence is still thin.