The European Commission has initiated a consultation on a draft implementing act aimed at establishing rules for the scientific panel of independent experts, which is integral to the enforcement of the AI Act. This panel will assist the AI Office and national market surveillance authorities in enforcing the AI Act. The panel, along with the AI Board, is empowered to act independently by requesting documentation from general-purpose AI model providers and alerting the Commission about potential risks.
The draft limits the panel to 60 experts selected for two-year terms through an open call for expressions of interest. These experts must be independent from any AI systems or general-purpose AI model providers. Expertise in areas such as fundamental rights and equality will be considered during selection.
A Secretariat comprising the AI Office and Joint Research Centre will support the panel. Specific rapporteurs or contributors will be appointed from among its members to carry out tasks as required by the AI Act. Members will receive remuneration only when performing these roles at the request of the AI Office.
The draft implementing act requires that alerts issued by the scientific panel regarding general-purpose AI models be endorsed by a qualified majority of its members. Similarly, requests to access documentation held by a general-purpose AI model provider must have support from one-third of the panel's membership.
Responses to this consultation are due by November 15, 2024.
In related developments, Member States face a deadline of November 2 to nominate fundamental rights authorities under the Act. These authorities will have powers to request and access documentation through national market surveillance authorities.
On October 23, a workshop was held exclusively for general-purpose AI model providers under the Code of Practice process. The EU’s Joint Research Centre has outlined harmonized standards facilitating compliance with obligations on high-risk AI systems under preparation by European Standardisation Organisations.
Henna Virkkunen, Commissioner-designate for Tech Sovereignty, Security and Democracy, expressed an industry-friendly approach in her written responses ahead of confirmation hearings. She advocates for reducing compliance burdens on businesses adopting AI while proposing innovation-friendly legislation when necessary.
Michael McGrath, Commissioner-designate for Democracy, Justice and Rule of Law, emphasized actions needed for applying the AI Act in justice areas including guidance on risk categories and funding support for judicial authorities using AI tools.
A statement from G7 Data Protection Authorities highlighted their role in examining AI issues through data protection lenses and called for their inclusion in governance frameworks. The European Data Protection Board is set to hold a closed stakeholder event on November 5 concerning artificial intelligence models.
Additionally, a ruling by the Court of Justice of the European Union recognized commercial interests as legitimate under GDPR, relevant to processing personal data for AI training purposes. The EDPB released guidelines on this matter open for public consultation until November 20.
