Two signatory task forces on Marking & Detection and Labelling ensure the application of the Code of Practice on Transparency of AI-generated Content.
On 30 September 2026, over 90 signatories of the Code of Practice on Transparency of AI-generated Content have established 2 signatory task forces to support the implementation of the code. The signatories agreed on the rules and procedures (Vademecum) of the tasks forces. They will meet regularly to carry out the objectives of the task forces in ad-hoc working groups – set up to conduct technical activities. The AI Office chairs these task forces and will facilitate the discussions.
The 2 signatory task forces – envisioned under the 2 sections of the code – aim to facilitate exchanges between signatories to implement the code and advance the state of the art. The task forces will also be a forum to seek and discuss research, independent expert input and up-to-date evidence – relevant to the code’s commitments and measures – and to enable a dialogue and collaboration with stakeholders.
Experts and stakeholders with relevant expertise may be invited to participate to the task forces to maintain an open dialogue and to conduct activities in a fair and open manner. Interested experts and stakeholders may express their interest by sending an email with a summary of their expertise and intended contribution to: CNECT-AIOFFICE-CODE-OF-PRACTICE-TRANSPARENCY@ec.europa.eu
Below you can find further information about the 2 task forces.
Task Force 1 – Marking & Detection
Task force 1 gathers providers of both AI systems and AI models and providers of marking and detection technologies. Together, they discuss matters related to implementing measures for marking and detecting AI-generated content, detailed in section 1 of the code. The objectives of the task force include:
- advancing the state of the art of existing marking techniques and detection mechanisms,
- advancing existing solutions towards a higher degree of interoperability,
- standardising the interfaces of marking and detection solutions, and the underlying techniques themselves, when possible,
- developing new marking or detection solutions that are neutral to the AI provider,
- establishing industry best practices and/or public benchmarks for evaluating marking and detection solutions/mechanisms,
- organising and/or participating in red-teaming exercises to assess the limits of state-of-the-art marking and detection solutions/mechanisms.
Plenary meetings
- 30 September 2026
- 1 October 2026
Minutes will be published soon.
Ad-Hoc Working Group on Interoperability
The objective of this working group is to advance interoperability across marking and detection solutions through various mechanisms. The goal is to allow laypersons to easily assess whether content is AI-generated or manipulated, without having to try several detectors independently.
Meetings
This working group will be launched in October-November 2026.
Task Force 2 – Labelling
Task Force 2 gathers deployers of AI systems to discuss matters related to the implementation of the measures for labelling of deepfakes and certain text publications, detailed in section 2 of the code. The objectives of the task force include:
- promoting the uptake and further development of the EU icon (including an audio-only version),
- updating the EU icon to ensure state-of-the-art disclosure,
- developing an interactive second layer to be linked to the EU icon or equivalent label,
- working towards common methods towards disclosing specifically the nature of AI involvement,
- exchanging good practices and developing sectoral practices for disclosure.
Plenary meetings
- 30 September 2026
- 1 October 2026
Minutes will be published soon.
Ad-Hoc Working Group on Interactive Second Layer
The objective of this working group is to define an additional interactive element that can be triggered when the user hovers over or clicks on the primary EU icon. The working group will:
- define what information should be conveyed in this second layer,
- define how it can be implemented,
- define effective design elements,
- explore interoperability with machine-readable markings.
Meetings
This working group will be launched in January 2027
Ad-Hoc Working Group on EU Audio Earcon
The objective of this working group is to develop an EU earcon (short tune, combined with spoken language or on its own). It could convey the AI-generated or manipulated nature of deepfakes and published text, whether for audio-only content or as accessible disclosure for the visually impaired.
Meetings
This working group will be launched in January 2027
Ad-Hoc Working Group on Sectorial Practices
This working group will provide a forum for cross-sector exchanges and for developing sectoral good practices regarding the disclosure of deepfakes and published text, in line with section 2 of the code.
Meetings
This working group will be launched in October-November 2026
Latest News
On 9 October 2026, Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy, and Seán Ó hÉigeartaigh, Chair of the AI Scientific Panel, convened a special meeting.
The Commission holds a special meeting of the Scientific Panel on artificial intelligence (AI).
On 12 October, the technology roadmaps of the European Connected and Autonomous Vehicle Alliance (ECAVA) will be presented at a session in Paris, held alongside the Paris Motor Show.
From 14 October to 17 November 2026, this initiative led by the European Commission in partnership with the upcoming Irish Presidency of the Council will promote a series of online and in-person events showcasing AI innovation and adoption across Europe.
Related Content
Big Picture
Last update
12 October 2026