Artificial intelligence is advancing at breakneck speed, reshaping industries and decision-making processes. Yet, regulatory frameworks and compliance systems are struggling to adapt in real time. This widening gap poses risks not just to governance, but to trust and accountability. How can organizations and regulators close it without stifling innovation?
Stacey English, director of regulatory intelligence at Theta Lake, believes that while AI tools are rapidly reshaping compliance, companies can’t afford to lose sight of how these technologies align with tomorrow’s regulatory expectations.
She said, “Industry research of 500 compliance and IT leaders shows that nearly two-thirds of firms are already using AI in supervision—but 62% face data and implementation challenges. This highlights the concerns about whether current tools are flexible and transparent enough to meet the compliance demands of tomorrow. The best-in-class solutions are designed with the future in mind.
“They can already accommodate multiple purposes such as enabling keyword-based detections in parallel with AI classifiers to detect compliance risks. This helps firms to detect both known red flags as well as less obvious clues where contextual information like emojis and reactions are used.”
Despite this, adopting AI isn’t a ‘set it and forget it’ in the view of English. “Continuous maintenance is essential. Models must be regularly updated to reflect changing regulatory requirements, emerging risks, and shifts in communication norms—for instance, our classifiers have evolved to identify conversations that may signal off-channel activity or determine whether an AI notetaker bot or meeting assistant is present in a meeting.”
Without regular calibration, English underlines, firms risk over-relying on outdated models that may fail to detect relevant issues or generate false positives that erode trust in the tools.
“The AI compliance tools that will endure are those that invest not just in AI capabilities, but in the governance and model management to ensure they accommodate evolving risks and regulatory expectations,” she said.
In order to strike the right balance between AI innovation and regulatory safety, transparency is key, states English.
She detailed, “Compliance teams and regulators alike need to understand how decisions are made by AI tools. Features such as Theta Lake’s classifier audit reports and detection explanations and annotations are now essential, not optional features, that enable auditability and confidence in machine-driven decisions. As the use of AI for compliance continues to grow it will be matched by growing demand for explainable AI. The future of compliance isn’t just about embracing AI’s efficiencies —it’s about making sure fast-evolving AI capabilities keep pace with shifting compliance risks and regulatory demands.”
Fast-paced
It’s undeniable that AI is advancing at a pace that few could have predicted even a couple of years. For John Byrne, CEO and founder at Corlytics, said that while this acceleration creates enormous opportunities for firms to improve efficiency, decision-making and customer outcomes, whilst also exposing a fundamental tension – the gap between AI capabilities and the pace at which regulatory frameworks evolve.
He said, “The deployment of AI is not regulated, through the AI Directive and similar frameworks. Many large, regulated firms already have their own security and safety standards regarding the deployment of AI; many of these go beyond the scope of any current regulatory requirement. It is rare for large, regulated firms to accept latest models deployed by vendors that are untested. This requires proactive engagement with regulators, investment in explainable and transparent AI systems, building internal governance structures that can adapt as standards evolve.”
However, Byrne believes there is a risk that AI will outpace regulation, and in many ways, already has.
“Rather than viewing this as a constraint, it should be seen as a call to lead,“ said Byrne. “The latest models can do things not possible six months ago, however, there needs to be a more rigorous science-based approach, as many reliable models are already producing accurate results. Firms that embed compliance into their AI design processes will not only avoid regulatory pitfalls but will also build trust with customers, partners and society at large. The benefits are greater accuracy, consistency, immediate up-to-date traceability in compliance systems and greater efficiency.”
