As regulatory change accelerates worldwide, businesses face growing pressure to keep up. In this shifting landscape, generative AI has emerged as a powerful ally, offering automation, analysis, and the ability to turn regulatory chaos into clarity. But the question remains: can GenAI truly transform compliance without introducing new risks?
According to Alex Mercer, Head of Innovation at Zeidler Group, many firms are cautiously embracing GenAI to automate existing processes. “Fully autonomous systems are rare, but steps like quarterly filings are increasingly automated,” he said. Still, Mercer notes GenAI struggles with legal nuance and interpretation. “Models may sound convincing but can lack consistency.” He emphasizes that this can be improved by fine-tuning models, implementing retrieval-augmented generation (RAG), and reinforcing prompt structures.
Regulators, Mercer adds, are watching closely but haven’t stepped in forcefully yet. “They see the potential for firms to remain compliant and reduce risk exposure, as long as responsibility and audit trails are maintained.”
Yet risks remain. Mercer cautions that most large language models (LLMs) are biased toward U.S. laws and may be outdated. Without layers of support and localized data, interpretations can be misleading. Moreover, the nondeterministic nature of LLMs means responses vary, raising issues for consistency.
Meanwhile, UK-based RegTech FullCircl sees a future where AI, machine learning, and RPA fully reshape compliance. From AML to credit screening, they argue AI is best used on complex workflows. Natural language processing (NLP) excels in summarizing legal documents, though oversight remains critical for more intricate tasks. FullCircl believes regulators support AI adoption—provided it protects consumers and upholds market integrity.
Firms like 4CRisk.ai highlight how AI accelerates regulatory scanning, rulebook creation, and impact analysis. However, co-founder Supradeep Appikonda warns that domain-specific models are essential to avoid misinterpretation and bias.
Flagright’s Madhu Nadig agrees, noting that while GenAI drastically speeds up analysis, it must remain a drafting tool, not a decision-maker. Traceability, explainability, and human oversight remain vital.
AscentAI, too, sees GenAI as a game changer—enabling fast, focused summaries of regulatory obligations that help firms act quickly. But across the board, experts stress the same thing: AI is a powerful aid, not a replacement for human judgment.
