Summary
June 2026 marks a major inflection point for the EU AI Act. On 1 June 2026, the European Commission formally appointed a 60-member Scientific Panel and an Advisory Forum — giving the EU AI Office the independent expert infrastructure it needs to actively enforce the regulation. Separately, a May 2026 “Digital Omnibus” political agreement introduced the first amendments to the AI Act since its adoption, including staggered deadline extensions for High-Risk AI Systems (HRAIS) and a new prohibition on AI-generated intimate content (“nudifier” apps). Full enforcement of most provisions remains on track for 2 August 2026.
Key Developments
- Scientific Panel appointed (1 June 2026): 60 independent experts covering frontier AI, technical auditing, and societal impact will advise the AI Office on GPAI model compliance and systemic risk.
- Digital Omnibus amendments (7 May 2026): Compliance deadlines for Annex III use-based HRAIS pushed from August 2026 to December 2027. New prohibition on nudifier applications effective December 2026.
- Code of Practice on AI-generated content: Final version due June 2026, with enforcement beginning August 2026. Covers watermarking and labelling of AI-generated material.
- August 2, 2026 deadline: Full Article 50 transparency obligations take effect. CE marking, conformity assessments, and EU database registration for high-risk systems must be completed.
Implications for Businesses
With the enforcement machinery now operational, EU AI Act compliance is no longer theoretical. Organisations that have delayed action face real regulatory exposure — particularly around transparency, GPAI model obligations, and prohibited practices. Fines range from €7.5M (1.5% of global turnover) for transparency breaches up to €35M (7% of global turnover) for prohibited AI practices.
The deadline extension for HRAIS is welcome relief, but does not apply universally — check which Annex your AI systems fall under. GDPR enforcement in an AI context is already active, with data protection authorities issuing fines for biometric and emotion-recognition AI.
Compliance Checklist
- ☐ Classify all AI systems in use against EU AI Act risk categories (prohibited, high-risk, limited-risk, minimal-risk)
- ☐ Confirm whether your HRAIS falls under Annex I or Annex III — deadline extensions only apply to Annex III
- ☐ Complete conformity assessments and technical documentation for high-risk systems by August 2026
- ☐ Implement AI-generated content watermarking/labelling ahead of August 2026 enforcement
- ☐ Review any use of biometric, emotion recognition, or workplace surveillance AI for GDPR exposure
- ☐ Register high-risk AI systems in the EU database
- ☐ Ensure workforce AI literacy obligations are being met (in force since February 2026)
IT & Admin Impact
- Documentation: Technical documentation and audit trails must be maintained for all high-risk AI systems — build this into your AI procurement and deployment processes now.
- Vendor management: Verify that AI vendors (especially GPAI providers) have signed or are complying with the GPAI Code of Practice. Meta has refused and faces enhanced scrutiny.
- Content systems: Any system generating text, images, or audio for public consumption will need watermarking/labelling capability by August 2026.
- Incident reporting: Establish processes for reporting AI incidents to authorities promptly — this becomes mandatory post-August 2026.
Sources: European Commission (1 June 2026), Covington Global Policy Watch (2 June 2026), Latham & Watkins (May 2026). Last updated: June 2026.