AI Accountability – Why “I Didn’t Understand It” Is No Defence

UK regulators are sending a clear message to senior managers. If artificial intelligence causes harm on your watch, you will still be held accountable.

In January, Parliament’s Treasury Committee warned that the FCA expects senior managers to understand any AI systems used within their area of responsibility. The difficulty is that many AI tools are opaque, adaptive and not fully explainable, yet the Senior Managers and Certification Regime offers little practical guidance on how firms are meant to govern them.

The FCA has already made its position clear. A lack of understanding will not be accepted as a defence.

Lawyers are increasingly concerned that senior managers may face enforcement action after the event rather than being guided by clear expectations in advance. This is not a theoretical risk.

Examples already being discussed include:

• • • AI-driven suitability and risk profiling in retail investment and crypto product sales
    • Automated distribution, marketing and product targeting systems
    • Automated distribution, marketing and product targeting systems
    • Model-driven product risk classification and customer segmentation

Where AI wrongly classifies customers, firms could sell unsuitable products to thousands of people before the problem is detected. Regulators could frame this as a failure of systems and controls, exposing the senior manager responsible to personal enforcement action.

We have seen this approach before. The PRA has already fined senior managers for failures linked to large-scale IT systems. Many lawyers now expect the same framework to be applied to AI.

The key risk is not that firms use AI, but that they do so without clear governance, challenge and oversight.

How Objectivus can help

Objectivus supports boards, senior managers and compliance teams to:

• Map accountability for AI under the SMCR
  • Assess governance, validation and model risk controls
  • Stress test automated decision frameworks for regulatory defensibility
  • Build audit-ready documentation for FCA and PRA scrutiny

AI in Risk and Compliance | Roundtable | 5 March | 16:00 to 19:00

Objectivus is hosting a closed roundtable, AI in Risk and Compliance, on 5 March from 16:00 to 19:00 to explore how firms can align AI deployment with regulatory accountability under the Senior Managers Regime.

Registration is available via the Axiol website.

If AI is now part of your control environment, it must also be part of your regulatory strategy.