Regulation
New seven-country study finds citizens want stricter AI oversight than regulators are delivering
A conjoint survey across seven countries finds a consistent public preference for safety over innovation and public governance over industry self-regulation. That gap is directly relevant to how AI-driven trading gets regulated.
A new academic study by researchers Magnus Lundgren and Jonas Tallberg surveyed citizens across seven countries with varied political and economic profiles, using a conjoint experiment to isolate what the public actually wants from AI governance rather than what they say in the abstract. The finding worth sitting with: there is a systematic gap between the regulatory approach most governments and industry bodies are converging on and the preferences citizens express when asked to trade off specific options. People consistently favour safety over innovation speed, public oversight over industry self-regulation, and international coordination over fragmented national rules.
None of this study is about trading specifically, but the implications land squarely on AI-driven financial markets. Retail-facing AI trading tools, robo-advisors and agentic execution platforms are exactly the kind of consumer-facing AI application where the study's finding, that people want public bodies rather than firms setting the rules, has direct regulatory consequences. If regulators in the UK, EU and US are currently leaning toward principles-based, industry-led frameworks for AI in finance, this research suggests that approach may be running ahead of, or against the grain of, what the public that firms serve actually wants.
The self-regulation versus public-oversight distinction is not academic for anyone running or evaluating an AI trading product. The FCA's current posture on algorithmic and AI-driven trading leans on existing SYSC and MAR frameworks rather than AI-specific rules, effectively asking firms to document and self-assess. That is precisely the model the study's respondents trust least. It does not mean the FCA approach is wrong, existing frameworks may in practice be adequate, but it does mean firms relying on light-touch self-assessment should expect political pressure toward more prescriptive oversight if public sentiment tracks what this study found.
The practical takeaway for anyone building or deploying AI trading infrastructure: treat governance documentation and independent validation as an investment in license to operate, not just a compliance checkbox. A firm that can show a model's decision logic, risk limits and validation regime to a sceptical regulator or journalist is in a stronger position than one relying on "trust our process," and this research suggests that scepticism has more public backing than industry conversations often assume.
Read the original: The Cryptonomist - New seven-country study finds citizens want stricter AI oversight than regulators are delivering. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.