Could the EU AI Act’s focus on trustworthy AI leave human trust behind?

Details

This conceptual and legal analysis revisits fiduciary relationships in digital platforms and artificial intelligence (AI), asking whether fiduciary duties remain relevant in an AI-driven socio-technical environment. It outlines core features of fiduciary law and examines how digital platforms reshape interpersonal and institutional trust, drawing on the EU Digital Services Act (DSA) of 2022 and the UK Online Safety Act (OSA) of 2023.

Building on Balkin’s 2016 theory of information fiduciaries, the paper analyzes a broader transformation of trust across platforms and AI systems. It argues that AI systems may widen the trust gap as users increasingly over-rely on AI and move further away from traditional fiduciary relationships.

The DSA and OSA seek to enhance user trust through stronger transparency and accountability duties, but the paper identifies a distinct regulatory shift in the EU AI Act of 2024. It argues that the Act’s emphasis on AI trustworthiness risks separating trust from its moral foundations, potentially fostering misplaced trust (O’Neill, 2018), distributed trust (Botsman, 2017), or lazy trust (Myskja & Steinsbekk, 2020).

The paper concludes by criticizing the EU AI Act for deprioritizing human trust and advocates stronger individual rights and citizen participation in AI governance to mitigate trust gaps and the declining role of fiduciary relationships.