‘The EU Artificial Intelligence Act (‘AI Act’) pursues two overarching goals. First, to protect health, safety, and fundamental rights, while, secondly, improving the internal market by promoting the uptake of trustworthy artificial intelligence and supporting innovation. To pursue these two aims, the EU has opted for a dense regulatory framework consisting of (qualified) prohibitions of certain AI practices, a detailed legal framework for high-risk AI systems that still needs to be complemented by technical standards, transparency obligations, and a novel governance regime for general-purpose AI models. The regulation’s twin goals cannot be neatly separated, as innovation is bolstered by coherent and clear legal frameworks. Whether the AI Act overall achieves this objective is a discussion for another time. Rather, this editorial aims to highlight the confusing nature of its research exemptions, which are, besides the rules in Chapter IV, among the few rules explicitly dedicated to fostering innovation.’
Link: https://academic.oup.com/grurint/advance-article/doi/10.1093/grurint/ikaf100/8238058
AI Update 