The swift acceptance and uptake of artificial intelligence (AI) models—plus ample buzz around the potential for AI to take over many human tasks—means a variety of vendors are jumping into the space. But experts warn compliance professionals that AI use poses some risks for organizations, and they need to be prepared to mitigate the risks before adopting AI-based tools.

Many people are using general-audience tools like ChatGPT, which is based on a large language AI model, Anna Gressel, an attorney with Debervoise & Plimpton, said at a webinar sponsored by the Society of Corporate Compliance and Ethics May 10.^[1] But that’s set to change, with multiple, very specific applications in the pipeline, Gressel explained.

“We’re really seeing the emergence of a whole ecosystem or supply chain of AI models—everything from what we might call foundation models developed by big tech companies to all kinds of startups, and really established companies that are trying to tailor those models for specific use cases,” Gressel said.

“Some of those might be trained on your own data within your company. Some of those might be highly tailored to specific purposes based on industry data or industry insights. And so, we’re going to start seeing a world in which you can really tailor your experience based on the kind of insights you want, the kind of data you need, and the kind of protection around the environment that you might need for the types of data,” she said.

AI-based tools that work to detect fraud already are in use, said attorney Avi Gesser, with Debervoise & Plimpton LLP. But AI tools offer the promise of customized compliance tools that could collect a data set, consider certain risk factors like roles or geographic location and combine it all in a customized program that could recommend specific compliance training for specific team members, Gesser explained.