The autonomy industry should stop asking for regulatory certainty
What innovators call certainty, incumbents use as a moat. The sandbox model — provisional permission, dense evidence, iterated rules — is serving the sector far better than a premature rulebook would.
Daniel Carter — Policy Editor
At every autonomy conference this year, the same lament from the same podiums: 'We just need regulatory certainty.' It is a reasonable-sounding request that deserves more scepticism than it gets.
Certainty, delivered early, means rules written before evidence exists — which in practice means rules written around the capabilities and lobbying positions of whoever was largest at the time of drafting. The history of technology regulation is a museum of prematurely certain frameworks that took decades to unwind.
The sandbox record
Compare the jurisdictions running evidence-first regimes. Norway's maritime sandbox produced twelve commercially operating autonomous vessels and a data trail regulators trust. The UK's BVLOS pathfinder programmes generated the safety case that EASA's new framework now generalises. Provisional permission plus dense evidence has consistently outperformed comprehensive rulebooks drafted in advance.
The mature ask is not certainty but process: clear evidence thresholds, defined review clocks, and a commitment that rules follow data. Companies that can live with that are the ones building things that work.
Cite this article
Daniel Carter. "The autonomy industry should stop asking for regulatory certainty." Autonomous Systems Review, 8 Jul 2026. https://autonomoussystemsreview.com/articles/opinion-regulation-pacing-problem.
