[blitzcrieg1]

docs/compliance/ai-act-deployer-checklist.md headed its transparency section "Art. 52". In the final text, Art. 52 is the procedure for classifying general-purpose models with systemic risk, and transparency is Art. 50.

• The heading now says Art. 50.
• A short note says Art. 50 binds providers, and deployers only in the cases it lists, so the steps below it are good practice and not a list of what it requires. It also points to Art. 26(7) for informing workers about a high-risk system at work.

This was the only "Art. 52" citation in the docs, README, tests or package. All three articles were read on the Commission's AI Act Service Desk on 2026-11-06. This is not legal advice and counsel has not reviewed it.

Not changed, worth a separate look before this checklist is used in a sale

• The Art. 9, 10, 12 and 14 sections sit in a deployer checklist, but those articles are requirements for high-risk systems that providers must meet. Deployer duties are in Art. 26.
• Coding assistance is not automatically high-risk, and the checklist does not say so.
• The other section headings still use em-dashes.

🤖 Generated with Claude Code