Updated 17 October 2026. This article has been revised to separate the policy timeline from the product argument. See the current AI Act timeline for the category-specific dates and primary reference.
Consultation is a legitimate part of rulemaking. The question worth asking is whether a proposed change solves an implementation problem without simply moving the cost onto the people affected.
Assess a proposal by its consequences
Consider a proposed reduction in reporting. Start by identifying who uses the information and what decision it supports. If two reports repeat the same evidence, consolidating them may help. If removing one leaves an affected person without an explanation, the consequence is different.
A useful submission to a consultation should describe the implementation problem, the alternative and how protection would be maintained. “This is burdensome” is a starting observation, not a complete argument. Show which work is duplicated and which evidence remains available.
The same standard should apply to objections. If a proposal is likely to weaken oversight, identify the missing control or remedy. Avoid making the case depend on guessing the intentions of people in the room.
For product teams, this is a useful habit beyond regulation. Requests to remove a review step should include what replaces it, who accepts the residual risk and how the result will be checked.
| Decision | Evidence to request |
|---|---|
| Current burden | Specific duplicated or unnecessary work |
| Proposed change | Alternative process |
| Protection retained | Evidence, owner and remedy |
What to do with this
A concrete alternative is harder to dismiss than a slogan. It also forces the proposer to account for what the change leaves behind. That is where the useful debate begins.
Record the decision, the responsible owner and the evidence still missing. The decision-record example shows a lightweight way to do that. Where a claim concerns system behaviour, connect it to evaluation evidence. These are proposed product practices, not a finding that a system complies with the law.
Primary timeline reference: European Commission, AI Act. The operational recommendations above are my analysis.
