AI Regulation Tracker  /  Regulator guidance issued

Jersey's regulator tells finance firms that AI accountability stops at the board

The JFSC published AI guidance on July 17, 2026. It creates no new regime. It confirms that existing regulatory obligations already apply to AI, and states that responsibility cannot be delegated to technology. Boards and senior management remain ultimately accountable for AI-assisted decisions.

By Anthony Guerriero  ·  Published July 31, 2026  ·  Last updated July 31, 2026

Offshore financial centres have been slower than the EU to publish AI expectations, and the assumption in some quarters has been that a jurisdiction without an AI act is a jurisdiction without AI obligations. Jersey's regulator has now said the opposite in the clearest terms available to it.

On July 17, 2026 the Jersey Financial Services Commission published guidance on the use of artificial intelligence in the island's financial services sector. The document is guidance rather than a rulebook amendment, and the JFSC states plainly that it does not introduce a new AI-specific regulatory regime. What it does is confirm that firms are expected to apply their existing regulatory obligations to the use of AI.

The accountability sentence

The guidance's central proposition is short enough to quote and blunt enough to act on. Responsibility cannot be delegated to technology. Boards and senior management remain ultimately accountable.

That sentence closes a gap firms have occasionally tried to open. Where an AI system contributes to a regulated decision, the accountable person is not the vendor, not the model, and not the analyst who accepted the output. It is the board and senior management, under the accountability structures the firm already has.

In practice, firms may choose to assign a clear internal owner for each AI deployment, similar to how they structure responsibility for outsourced functions.

What the JFSC expects firms to address

The guidance identifies a set of areas firms should work through. Governance and accountability structures for AI systems come first. Risk identification follows, and the guidance names accuracy, bias, and explainability specifically. Human oversight of AI-driven decision-making is treated as a distinct expectation rather than folded into governance. Data governance, cybersecurity, and operational resilience are grouped together. Ongoing monitoring and lifecycle testing close the list.

The inclusion of explainability alongside accuracy and bias is the item most likely to bite in practice. A firm can evidence accuracy through testing and can address bias through data review, but explainability is a property of the model and the deployment together. Firms using third-party systems may find that the vendor cannot supply what the regulator expects the firm to be able to describe.

Lifecycle testing is the other demanding item. It points toward AI assurance as an ongoing exercise rather than a one-off pre-deployment test, which may go beyond what many firms currently do for a purchased tool.

The outsourcing angle

The guidance asks firms to review third-party AI provider arrangements, including contractual protections for data use and confidentiality.

This is the provision with the most immediate work attached, because it is retrospective. Firms that adopted AI tooling over the past two years generally did so under standard vendor terms, and those terms frequently permit input data to be used in ways a trust company or fund administrator would not accept for client data. The guidance implies that firms may need to revisit contracts they have already signed.

For firms holding client confidential information under Jersey trust and fiduciary obligations, the confidentiality point is not merely a data protection question. It touches the duties owed to beneficiaries and clients directly.

Where Jersey sits among the offshore centres

The offshore jurisdictions are moving at different speeds and with different instruments. The Bermuda Monetary Authority published a discussion paper on the responsible use of AI in Bermuda's financial services sector in July 2025, taking a principles-based approach. The Cayman Islands civil service adopted a first internal AI policy in March 2026, and the premier has indicated a legislative framework could arrive by mid-2027 following recommendations from a national digital taskforce. In the meantime, Cayman businesses are directed to existing data protection, financial services, and anti-money laundering requirements.

Jersey's guidance is therefore a relatively detailed regulator-facing statement from a Crown Dependency or Overseas Territory, and it arrives without waiting for legislation.

The pattern across all three is the same and is worth naming, because it is the opposite of the assumption that offshore means unregulated. None of these jurisdictions has an AI act. All three have indicated that existing obligations are relevant when firms use AI. A firm that treated the absence of an AI statute as the absence of an AI obligation has been reading the position incorrectly.

Browse the full AI Regulation Tracker →

Drafted with AI assistance and verified against the published guidance by a human editor. Informational only, not legal advice.