The Future Society responded to the European Commission’s call for comments on its proposed AI policies for the European Union (EU) outlined in its White Paper on AI. We welcome the European Approach on AI and we recommend to carefully build the civil service’s capabilities for experimentation, testing and audit of AI technologies. Otherwise, the EU’s great ambitions for trustworthy AI will falter during the enforcement of its policies.

You can download our contribution here and we provide a summary of our recommendations below.

World reference experimentation, testing and audit capabilities. The Future Society believes the EU should equip itself with measurement and benchmarking protocols, EU technology laboratories, risk assessment frameworks, auditing tools, test beds, etc. for new AI technologies. these new public capabilities will turn the European Approach into reality, protecting citizens’ rights while empowering innovation.

Design the evidence-based enforcement mechanisms the European Approach’s ambitions deserve. The Future Society believes the European Commission should review the available evidence and best practices for the design of experimentation, testing and audit capabilities and other policy instruments. From Member States’ national experience, U.S. NIST, and Public Administration experts, there is a lot of to learn from for designing an effective governance of AI for trust and excellence.

Enabling Excellence in Trust, and Trust in Excellence. The Future Society believes both ex-ante and ex-post compliance tools (such as pre-market testing protocols and auditing tools) are needed for a market that is both more innovative and more trustworthy. The EU should combine ex-ante and ex-post capabilities within the same coherent governance system.

Leadership is not simply about being first; it’s also about bringing the others forward with you. The Future Society believes the European Commission should facilitate access to its world reference experimentation, testing and audit capabilities. By encouraging utilization abroad without lowering the EU’s quality standards, the EU could strengthen its various partnerships with both developed and developing nations.

New capabilities for the Davids forming the backbone of European society, not only the Goliaths. The Future Society believes the new experimentation, testing and audit capabilities should be designed with and for SMEs, local governments & authorities, start-ups, self-employed individuals, NGOs, researchers, etc. without lowering the EU’s quality standards. This requires paying attention to the time-efficiency and financial impact of compliance, and considering novel technological approaches for enforcement.

We cannot always know the future, but we can learn faster. The Future Society believes agile governance is key for the legitimacy and sustained effectiveness of the European Approach on AI. Built-in opportunities for civic consultations and for regular landscape reviews would provide feedback loops for the EU’s leadership to continuously adapt to technological risks and opportunities, and citizens’ evolving preferences.

Modern technologies require modern governance tools and skills. The Future Society believes Research, Innovation & Capacity-building is crucial for the long-term success of the European Approach on AI. From the use of regulatory sandboxes to blockchain and software tools for decentralized audit, there is a whole new range of techniques to achieve compliance – and of skills to learn. The EU’s experimentation, testing and audit capabilities should include a programme for Research, Innovation & Capacity-building for trust and excellence.

Picture by Christian Lue