Chapter 7 GOVERNANCE Flashcards
Q: What is the role of the AI Office established under the EU AI Act?
A: The AI Office is a function of the Commission that will develop Union expertise and capabilities in the field of AI to contribute to the implementation, monitoring and supervision of AI systems and governance under the Regulation. Member States shall facilitate the tasks entrusted to the AI Office.
Q: What is the composition and structure of the European Artificial Intelligence Board?
A: The Board is composed of one representative per Member State, designated for renewable 3-year terms. The EDPS participates as an observer and the AI Office attends meetings without voting. Representatives should have relevant competences, serve as a contact point, and be empowered to facilitate consistency in their Member State. The Board adopts rules of procedure and establishes sub-groups.
Q: What are the main tasks of the European Artificial Intelligence Board?
A: The Board’s tasks include: coordinating national authorities; collecting expertise and best practices; advising on Regulation implementation; contributing to harmonized practices; issuing recommendations and opinions; promoting AI literacy and awareness; facilitating common criteria; cooperating with other bodies; assisting in expertise development; supporting regulatory sandboxes; contributing to guidance; advising on international AI matters; and handling general purpose AI model alerts.
Q: What is the role of the advisory forum established under the EU AI Act?
A: The advisory forum provides balanced representation of industry, SMEs, civil society and academia to provide technical expertise, advise the Board and Commission, and contribute to their tasks. Members are appointed by the Commission for renewable 2-year terms. The forum elects co-chairs, meets at least twice a year, and can establish sub-groups. It prepares an annual activities report.
Q: What are the provisions regarding the scientific panel of independent experts?
A: The Commission will establish a scientific panel of independent experts to advise and support the AI Office in implementing and enforcing rules on general-purpose AI, supporting market surveillance and cross-border activities, and assisting with safeguard clause duties. Experts are selected based on AI expertise, independence, and ability to be diligent and objective. Detailed arrangements are set out in an implementing act.
Q: How can Member States access the pool of experts on the scientific panel?
A: Member States can call upon the scientific panel experts to support their enforcement activities. Fees may be charged for this advice and support, as set out in the implementing act. The Commission will facilitate timely access to experts and ensure support activities are efficiently organized to provide the best added value.
Q: What are the requirements for Member States in designating national competent authorities under the EU AI Act?
A: Each Member State must establish or designate at least one notifying authority and market surveillance authority to act independently, impartially and without bias. A single point of contact must be designated and notified to the Commission. Authorities must have adequate technical, financial and human resources, and take appropriate cybersecurity measures. Confidentiality obligations apply.
Q: What guidance and reporting obligations apply to national competent authorities?
A: National competent authorities may provide guidance and advice on Regulation implementation, especially for SMEs and start-ups, taking into account Board and Commission guidance. They must report to the Commission every two years on the status and adequacy of their resources. The Commission will facilitate exchange of experience between national authorities.
Q: What is the purpose and content of the EU database for high-risk AI systems?
A: The Commission will set up an EU database for high-risk AI systems listed in Annex III, in collaboration with Member States. It will contain information entered by providers on systems and by deployers who are public authorities. Publicly available information will include provider details, system descriptions and conformity status. Non-public information includes more sensitive details.
Q: What are the access rules and data protection provisions for the EU database for high-risk AI systems?
A: With some exceptions, information on systems registered by providers will be publicly accessible in a user-friendly, navigable and machine-readable manner. Information registered by prospective providers will only be accessible to market surveillance authorities and the Commission, unless consent for public access is given. The database will contain limited personal data with adequate technical and administrative support.
