From Autonomous AI Agents to Complex Systems: On the Role of Framing in Policy and Law Making

Artificial intelligence (AI) systems are routinely characterised as autonomous agents endowed with human-like characteristics – not only in general discourse but also in expert literature and ‘fit-for-purpose’ legislative assessments. The presentation will show that this anthropomorphic framing distorts not only societal perceptions of AI technology but also policy and legal analyses. In particular, this misconception has given rise to a range of untenable normative claims – from proposals to recognise legal personhood for AI systems (including for the purposes of allocating liability), to calls for abolishing intellectual property (IP) law and imposing a moratorium for lack of any other instrument to address the control problem.

The presentation will offer an alternative account of AI technology grounded in complex systems theory, which explains the emergent capabilities of AI systems – what may appear as autonomous behaviour, creativity, or ingenuity – without mystifying them. We will then examine the implications of this reframing for foundational normative and legal concepts across several legal domains, focusing on authorship and inventorship while also outlining its ramifications for legal personhood, tort liability, and risk regulation. Overall, the analysis shows that, from a complex-systems perspective, the fundamental objectives of these legal fields, and the substantive legal requirements intended to secure them, continue to hold in the age of generative AI. The challenges arise, rather, at the level of interpreting and applying the law.