On 10 September 2020, the SHERPA project provided feedback to the European Commission as part of a public consultation on its Inception Impact Assessment for regulation of artificial intelligence. The inception impact assessment is a follow-up to the Commission’s White Paper on AI, which we provided feedback on in June 2020.
In the response, SHERPA reiterated the need for an EU-level mandatory framework that prioritises protection of human rights and recognises both the benefits and harms of AI. The protection of human rights and safety is a societal priority, and not only a means to create (or the outcome of creating) trustworthy AI.Â
SHERPA advocated for a regulatory framework of ex-ante and ex-post enforcement mechanisms at the EU-level that establishes a baseline, encourages high standards of protection of fundamental rights and societal values, and minimizes inconsistency and fragmentation at the Member State-level. An EU approach should build on existing legislation and enforcement mechanisms. A complementary voluntary labelling scheme would be valuable, but only if robust, clearly articulated, and not symbolic.Â
Other recommendations included the creation (or promotion) of effective complaint and redress mechanisms accessible to stakeholders, and comprehensive assessment of AI applications in a manner that captures unforeseen and/or unintended impacts. 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.
