Regulating emerging technologies: Asymmetric legal uncertainty and the EU General-Purpose AI Code of Practice

We studied the regulation of emerging technologies through the European Union’s General-Purpose AI Code of Practice drafting process, asking when and why do actors prefer greater legal certainty or greater legal uncertainty. Consistent with economic theories of regulation, we found that smaller firms preferred legal certainty, whereas larger and influential AI developers often preferred greater legal uncertainty. Information asymmetry also explained differences in preferences. However, economic theories assume the availability of information, whereas emerging technologies entail Knightian factual uncertainty. To account for this, we put forward a complementary sociological explanation for actors’ regulatory preferences, which we term asymmetric legal uncertainty: a situation in which all actors are equally ill-informed, but some, due to a perceived expertise, are given a de facto degree of control over the factual narrative on which legal consequences hinge. We argue that, in such a situation, legal uncertainty allows actors to speculate about future legal outcomes.