Every development we track — guidance, standards, deadlines, enforcement — and what it means for your tier. Checked on a two-week cycle.
5
Tracked entries
4
Source-linked
3
Template changes
Draft high-risk guidelines and postponed Annex III timeline reviewed
High-risk
The Commission is consulting on draft guidelines for classifying high-risk AI systems. The current implementation material points to 2 December 2027 for Annex III stand-alone high-risk systems and 2 August 2028 for product-embedded Annex I systems.
What changed: The guidance focus is classification: intended purpose, Annex III mapping, Article 6(3), product-embedded systems and overlapping operator roles. It is interpretive guidance, not a replacement for the AI Act text.
Template status: Template review: classification memo and roadmap wording already separate Article 50, Annex III and Annex I dates. No regeneration is required solely because of this consultation unless your use-case mapping changes.
Consultation deadline: 23 July 2026
Customer action: High-risk teams should compare their intended purpose against the draft examples and consider submitting feedback if the examples affect their product category.
Transparency Code opinion published for Article 50 AI-generated content duties
Limited-riskGPAI
The Commission published its opinion on the Code of Practice on Transparency of AI-generated content. The opinion says the code can help implement Article 50(2), 50(4) and 50(5), while adherence is not conclusive evidence of compliance.
What changed: Article 50 customers now have a clearer voluntary reference point for AI disclosure, synthetic-content marking and deepfake labelling evidence.
Template status: Template change: the Compliance Roadmap now prompts Article 50-triggered systems to evaluate the Transparency Code as a voluntary evidence path and to record the decision.
Customer action: If your system uses chatbots, synthetic media or deepfake-style outputs, record whether you will follow the code, another equivalent control set or a documented alternative.
AI Act harmonised-standard work remains in progress
High-risk
CEN-CENELEC work is still underway across AI Act areas such as risk management, data governance, logging, human oversight, accuracy, robustness, cybersecurity, quality management and post-market monitoring.
What changed: The Commission standardisation page identifies active work and highlights prEN 18286 on quality management as a first harmonised-standard candidate entering public enquiry, but high-risk providers should not assume final OJEU-backed standards are already available.
Template status: No template schema change yet. TrustPacket will update Annex IV, QMS and post-market-monitoring prompts when final harmonised standards or official support tools materially change the evidence expected.
Customer action: Map current controls to the expected standard areas now; wait for final standards or official support tools before treating a draft standard as a presumption-of-conformity route.
A sample public compliance page is now available so buyers can see the Monitoring output before subscribing: tier, maintained documents, last-reviewed date, template version and recent update history.
What changed: Monitoring now has a concrete public example instead of only a feature description.
Template status: No customer document template change. This is a product-demo and sales-proof update.
Customer action: Review the demo page if you want to see what a vendor-questionnaire link looks like.
From today we track EU AI Act implementation on a two-week cycle: Commission guidance, harmonised standards progress (CEN-CENELEC JTC 21), member-state enforcement setup, and deadline changes. Every relevant development lands on this page.
What changed: Monitoring moved from a product promise to a public, timestamped changelog.
Template status: No template change in this entry. Future entries will state explicitly whether templates changed.
No action needed.
Get these as email alerts — with your tier called out
Monitoring subscribers get each update by email with a concrete action for their system, plus free document regeneration when templates change.