2024 was a big year for AI regulation, especially in California, where Governor Gavin Newsom signed 18 new AI-related laws while vetoing a high-profile bill that would have imposed stricter rules on AI companies. But the action isn’t slowing down—2025 could see even more movement, particularly at the state level.

Experts suggest that states are likely to outpace the federal government when it comes to AI rules. Mark Weatherford, a cybersecurity veteran who’s worked with both state and federal governments, believes states are better positioned to act quickly. “AI is changing daily,” he says. “The federal government will get there, but it’s going to take time.”

California is already setting the tone. In the past year alone, the state introduced over 400 pieces of AI-related legislation. While some measures were narrowly focused—like rules around AI-generated pornography or data training—others aimed for broader oversight. Newsom’s veto of a major AI bill, which would have required extensive testing and slowed innovation, sparked debate. Still, Weatherford expects California to push for stricter regulations in 2025.

One challenge? Harmonizing state laws. With so many states crafting their own rules, companies face a patchwork of regulations. “Harmonization is key,” Weatherford explains. “Otherwise, businesses have to navigate a maze of conflicting requirements.” But getting states to align their laws isn’t easy. While some states are borrowing language from each other, there’s no formal system to ensure consistency.

On the federal level, progress is slower. A recent 230-page report from a House task force on AI highlights the complexity of crafting national rules. Partisan divides and bureaucratic delays mean federal action could take years. Meanwhile, states like California are filling the gap, often setting trends that other states follow.

Weatherford also sees synthetic data as a game-changer for AI development. “Without quality data, there’s no AI,” he says. Synthetic data—artificially generated information that mimics real data—could help address privacy concerns and reduce bias. Companies like Gretel, where Weatherford now works, are betting big on this technology.

As AI continues to evolve, so will the debate over how to regulate it. While some fear overregulation could stifle innovation, others argue that guardrails are essential to protect privacy and security. One thing’s clear: 2025 will be another busy year for AI policy, with states leading the charge.