From 2 August, the European Commission’s AI Office, in collaboration with national authorities, will implement the legislation on artificial intelligence (AI Act), the Commission recalled on Friday 31 July.
The Office will be able to enforce the rules of the AI Act on providers of general-purpose AI models (GPAI). These providers must document certain information and provide it to the competent authorities or to downstream providers. They must also put in place a copyright policy and publish a sufficiently detailed summary of the content used to train their models.
On 2 August, new transparency rules will also enter into force. Under these new rules, chatbots and other interactive AI systems will have to inform users that they are interacting with AI and not with a human. Deepfakes (images, videos or audio files altered or generated by AI) will have to be labelled. Content generated or altered by AI will also have to include machine-readable markings so that it can be detected more easily.
In addition, on 31 July the Commission published an initial list of more than 180 organisations that have signed the Code of Practice on Transparency of AI-generated Content, which implements the transparency rules for such content.
See the list: https://aeur.eu/f/n35
Responsibility for enforcing the transparency rules and prohibited practices is shared between three bodies: - the AI Office, which enforces the rules applicable to AI systems offered by the same provider as the underlying general-purpose AI model; - the competent national authorities, which enforce the rules applicable to other AI systems; - the European Data Protection Supervisor, which enforces the rules relating to AI systems used by the institutions, bodies and agencies of the European Union.
The AI Office and the competent national authorities will be supported in their enforcement work by the Scientific Panel, an advisory body made up of 60 independent experts, and by Professor Alessandro Abate, appointed Lead Scientific Adviser.
To facilitate enforcement of the AI Act, the Office has set up a Complaints Tool to report alleged infringements of the AI legislation committed by providers of AI systems supervised by the Office (see the tool: https://aeur.eu/f/n36 ). People working with providers of AI systems or general-purpose AI models will be able to use the Whistleblower Tool to safely report potential violations of the legislation (see the tool: https://aeur.eu/f/n37 ). Lastly, downstream providers developing AI systems from general-purpose models will also be able to report alleged infringements committed by the providers of those models (see the tool: https://aeur.eu/f/n38 ).
In the event of an intentional or negligent breach of the AI Act, the Commission may adopt a decision imposing penalties on the provider of the AI system or general-purpose AI model concerned. The amount will depend on the nature, seriousness and duration of the infringement, and may be as much as €35 million or 7% of the offender’s total worldwide annual turnover, whichever is higher. (Original version in French by Camille-Cerise Gessant)