What Bermuda Should Do Next on AI
Bermuda has the foundations for good AI governance: a regulator that consults seriously, a modern privacy law and a Government with an AI policy. What it lacks is a joined-up agenda. Here is the one I would propose, in the order I would do it.
1. Finalise the BMA Guidance Note, With an Evidence Standard
The BMA's consultation closed on 30 October 2026. The final Guidance Note should keep its principles-based design, but say plainly what evidence is sufficient for material and higher-impact AI, with an illustrative minimum evidence set. Firms want to comply; tell them what compliance looks like.
2. Publish PIPA Guidance on AI
PIPA has been fully in force since 1 January 2025 and already governs any AI that uses personal information. But the Privacy Commissioner has not yet published AI-specific guidance. Short guidance on training data, generative AI tools, overseas AI vendors and AI-related breaches would prevent a great deal of accidental non-compliance.
3. A Public Register of Government AI Use
The Government's AI Policy requires human review of AI decisions that affect people. That commitment would be far stronger with a public register of government AI use cases, their owners and their oversight arrangements, starting with the NEAR AI deployment announced in May 2026.
4. Measure AI's Effect on Bermudian Jobs
62% of residents believe AI puts jobs at risk, and published estimates range from 2% to 20%. Bermuda should stop arguing from borrowed numbers. A short annual survey of employers through the Department of Statistics would show which roles are changing and how fast.
5. Make AI Governance a Bermudian Profession
Every BMA-regulated firm will need people who can build AI inventories, assess materiality and evidence controls. Bermuda College's AI courses and the Government's AI Education Series are a good start; the next step is a governance track tied to the BMA's expectations, so the jobs AI creates go to Bermudians.
6. Court Rules for AI
The judiciary's committee on an AI practice direction should finish its work. Clear rules on disclosure of AI use and verification of citations protect litigants and lawyers alike, and build on the Supreme Court's forward-looking 2025 decision in Fourworld v Enstar.
7. Legal Clarity for AI Agents, Done Carefully
The Premier has said the Government will explore legal clarity for AI agents in financial services. That is worth doing, but in step with the BMA: accountability must always trace back to a person or firm, and agents' boundaries must be enforced by systems outside the model.
One Agenda, Many Owners
None of this requires a sweeping AI Act. It requires each institution to finish the job it has already started, and to do it in the open. Laurenceo will track every step on its Bermuda AI Tracker.
About the Author
Jorai Laurenceo is the Founder and Executive Director of Laurenceo, Bermuda's AI governance think tank, based in Hamilton, Bermuda. He wrote Laurenceo's public response to the Bermuda Monetary Authority's AI consultation and leads Laurenceo Intelligence. More about Jorai.