On 12 June 2020, the SHERPA project provided feedback to the European Commission as part of a public consultation on its White Paper on Artificial Intelligence – A European Approach. The feedback was based on SHERPA findings and preliminary recommendations.
In the response, SHERPA advocated for promoting and funding artificial intelligence (AI) that is ethical, protects human rights, and supports human flourishing. Specific concerns cited include the potential of AI to endanger safety, breach fundamental rights, and affect economic and power distributions.
In regard to building an ‘ecosystem of trust’, SHERPA expressed concern about the EC’s proposed framework for assessing high-risk AI applications, which seemed unlikely to capture unforeseen and/or unintended impacts. To address the challenges presented by AI, SHERPA indicated that a combination of ex-ante compliance and ex-poste enforcement mechanisms are needed at the EU and Member State-level, which should begin with a comprehensive gaps analysis to identify and prioritise required regulatory reform. SHERPA agreed that the proposed voluntary labelling system could be useful, but only if robust and carefully designed.
Other recommendations included making protection from harm the primary objective of a European approach; prioritizing the principle of non-maleficence over building ‘trust’ in AI; creating and/or promoting an effective complaint and redress mechanism accessible to stakeholders; and encouraging the use of ‘ethics by design’. The response referred to other SHERPA activities, including guidelines for AI developers and users, ongoing development of Terms of Reference for a regulator, and work on standardisation.
