European Commission prepares to launch Office for AI

 The European Commission is ready to adopt a decision creating the European Office for Artificial Intelligence (or AI Office). The news is reported by Euractiv on January 23, 2024





The idea comes from the European Parliament

The idea of ​​an Office for AI to centralize the application of all the rules on artificial intelligence came from the European Parliament. However, during the negotiations, this amount was revised downwards. a little less of an agency to be integrated into the Commission, but with a separate budget line.

However, the question of the degree of autonomy that the Office will enjoy remains sensitive within the Commission, especially since it is not clear whether it will become an entity with its own political objectives or an extension of the the unit responsible for AI law.

Euractiv reports that the resulting draft decision was amended following an internal consultation to include wording clarifying that the Office must not interfere with the competences of Commission services.

According to the document, the decision should enter into force urgently on February 21, 2024, before the formal adoption of the European AI law . Euractiv estimates that the decision should be taken on Wednesday January 24.

The Office will have the mission of supervising AI

The Office for AI will mainly play a supporting role with regard to the application of rules on artificial intelligence systems, and most competences will fall to national authorities . However, the Bureau has been tasked with monitoring general purpose AI (GPAI) models and systems, the most powerful types of AI to date.

Recent advances in computing power, data collection, and algorithmic techniques have led to the development of powerful models, such as OpenAI's GPT-4, which powers ChatGPT, the world's most popular chatbot.

AI Act, a progressive approach to GPAI models

The AI ​​Act includes a phased approach to GPAI models to distinguish those that could pose a systemic risk to society from others. The AI ​​Office is responsible for developing methodologies and benchmarks to evaluate the capabilities of GPAI models.

The Office should be able to distinguish itself in monitoring the application of the rules on GPAI models and systems, in particular when they are developed by the same supplier, and the emergence of unexpected risks arising from such models, on the based on the reports of a scientific group of independent experts. .

The new EU entity will also have wide latitude to investigate possible violations of GPAI-related rules by collecting complaints and reports, issuing document requests, conducting assessments and requesting mitigation or other enforcement measures.

The Office will also coordinate the application of the AI ​​Act to AI systems already regulated by other EU legislation, such as recommendation systems on social networks under the Services Act digital markets and search engine positioning algorithms under the Digital Markets Act.

The EU AI Office, a support and coordination role

The Office for AI will carry out a supporting role in the preparation of secondary legislation implementing the AI ​​Act, in the uniform application of the Regulation, in the publication of guidance and supporting tools such as standardized protocols, in the preparation of standardization requests, in the implementation of regulatory experiments, in the development of codes of good practice and conduct at EU level.

The entity will also provide secretariat to the IA Board of Directors and administrative support to the Stakeholder-Led Advisory Forum and Scientific Panel. The draft decision explicitly refers to the obligation to regularly consult scientific and civil society stakeholders.

In particular, the Office of AI must “establish a forum for cooperation with the open source community to identify and develop best practices for the development and safe use of open source AI models and systems.”

Furthermore, the new entity is responsible for promoting innovation ecosystems and working with public and private actors and the start-up community. As revealed by Euractiv, the AI ​​Office will be responsible for monitoring the progress of GenAI4EU, an initiative aimed at promoting the adoption of generative AI in strategic sectors.

The Office is also responsible for cooperating with relevant EU bodies, such as the European Data Protection Supervisor. Collaboration with other Commission services, in particular the European Center for Algorithmic Transparency, is also necessary to test GPAI models and systems and facilitate the adoption of AI tools in relevant EU policies .

Internationally, the Office will promote the EU approach to AI, contribute to AI governance initiatives and support the implementation of international agreements.

European AI Office, funding

The financial aspect of the AI ​​Office has been a sore point from the start. The inflexibility of EU budget allocations and the lack of willingness of member states to put more resources on the table mean that new tasks always face tight budgetary constraints.

The Commission's digital policy service, DG CNECT, will allocate human resources. The recruitment of temporary staff and operational expenses will be financed by the reallocation of the Digital Europe program budget.

Two new Edic consortia

The European Commission is also establishing, with certain Member States, two European Digital Infrastructure Consortia (EDIC). These are the Alliance for Language Technologies (Alt-Edic) in support of development of large European language models, and CitiVerse for the application of artificial intelligence tools to develop and improve local digital twins for smart communities and smart cities . Member States can create the European digital infrastructure consortia Alt-Edic and CitiVerse Edic with the support of the Commission.

Communication Ai@Ec

Finally, the European Commission adopted a communication which describes the strategic approach to the use of artificial intelligence and with which the European executive prepares internally the implementation of the ' AI Act . The Communication includes concrete actions aimed at strengthening the institutional and operational capacities necessary to ensure the development and use of reliable, safe and ethical AI. The Commission is organizing itself to support EU public administrations in the adoption and use of artificial intelligence.

Comments