Navigating Multiple State AI Laws in Wake of US Deregulation

The Trump administration’s AI action plan signals a major deregulatory shift by removing a 10-year freeze on the ability for states to regulate AI. States are moving quickly, setting up a potential patchwork of regulations that could affect AI deployments.

See Also: AI Agents Demand Scalable Identity Security Frameworks

Company policies must be “flexible enough to meet the strictest state without rewriting them every few months,” said Avani Desai, CEO at compliance consulting firm Schellman. While clear standards, better transparency in model training and strong accountability measures will help, the reputational stakes are high.

“Once public trust is lost, I think it’s incredibly hard to get back,” Desai said.

Desai advises firms to perform AI-specific risk assessments, test models before deployment, monitor outputs for drift, check training data sources and conduct AI red-teaming exercises to find weaknesses before attackers do. “Security teams need to treat AI as part of their core threat surface, regardless of what’s happening externally,” she said.

In this video interview with Information Security Media Group, Desai discussed:

• How conflicting state regulations could add compliance complexity to AI deployments;
• Strategies for integrating governance into AI innovation pipelines;
• The role of global standards in mitigating liability and trust risks.

Desai leads Schellman, the largest niche cybersecurity assessment firm in the world that focuses on technology assessments. She is an accomplished executive with domestic and international experience in information security, operations, P&L, oversight and marketing involving both start-up and growth organizations. She has been featured in Forbes, CIO.com and the Wall Street Journal, and is a sought-after speaker as a voice on a variety of emerging topics including security, privacy, information security, future technology trends and the expansion of young women involved in technology.