EU AI Innovation Accelerator preparatory action -

Deliverables

• Harmonised frameworks, tools and infrastructure at European level to support compliance with the AI Act that have been tested in practice by under-resourced SMEs and public authorities.
• Provision of materials and training courses to support compliance with the AI Act

The overall objective of this topic is to accelerate innovation and trustworthy AI by providing a range of services and tools to providers of AI systems, in particular to SMEs, to prepare for and facilitate compliance with the upcoming AI Act.

The action to be supported under this topic should focus on developing prototypes of services, tools and infrastructure that would be supporting companies to achieve and assess the compliance of their AI systems with the AI Act. This work notably includes:

• Developing process frameworks, virtual infrastructure, tools, metrics and benchmarks for compliant development and validation of AI systems against the AI Act requirements (e.g. for data quality, accuracy, robustness, safety), including through prototype(s) of software solutions for legal compliance;
• Developing a prototype of a digital tool supporting providers of AI systems in classifying their AI system in accordance with the levels of risks envisaged in the AI Act;
• Developing a prototype of a self-assessment tool for conformity assessment, risk management and post market-monitoring, specifically adapted to the needs of SMEs;
• Developing a digital platform and infrastructure making the frameworks, tools and metrics developed under this action publicly available;
• Providing personalised technical and legal support to small and medium sized enterprises and public authorities to use the frameworks, tools and infrastructure developed under this action, including to provide ad hoc replies to inquiries in relation to the AI Act - this would include developing online training and awareness-raising material targeted at providers and users of AI systems in relation to relevant requirements and obligations under the AI Act.

The awarded proposal from this topic is expected to collaborate as appropriate with the one from section 2.3.7.2 (AI regulatory sandboxes: EU-level coordination and support). In particular, the latter should rely on draft tools, metrics and guidelines developed under this action. At the same time the testing experience in the sandboxes under 2.3.7.2 should serve to test and update the deliverables under this action. In terms of target group, this action should focus mainly on under-resourced companies, in particular SMEs and start-ups, and under-resourced public authorities developing high-risk AI services/products or general-purpose AI as defined by the AI Act, in particular for what concerns any personalised services and support actions. All final materials, frameworks and tools developed and validated under this action will be made publicly available and may be used by any provider or user of AI system. To build on synergies with other initiatives supported by the EU, this action should establish contacts with TEFs, the AI-on-Demand platform, data spaces, and EDIHs, in order to work out which existing resources can be offered jointly to AI innovators and which resources would have to be built in future. Close cooperation should also be pursued with international and European Standardisation Organisations (in particular CEN107 and CENELEC108) to take into account progress in the development of AI standards and use them as a basis for the development of the action.

News flashes