New Mexico Artificial Intelligence Accountability Act (AI2A)

New Mexico just proposed a pretty decisive step on synthetic media: the Artificial Intelligence Accountability Act (AI2A) would require AI-generated images, audio, and video to include latent digital watermarks identifying the content as synthetic, and it would also require free provenance/verification tools so people (and law enforcement) can check origin and authenticity.

This is the direction I keep seeing globally: policymakers are moving from “please disclose” to “make it technically verifiable.”

My take: this isn’t an “AI industry” preference, it’s what the general public wants when deception becomes cheap and scalable. Even if parts of Big Tech and some investors push back, it’s hard to see this trend reversing: state by state, more of these laws will pass.

Where DeepMark fits: this is exactly what we build, robust, real-time provenance IDs embedded directly into content that stay detectable after compression, re-uploads, and common edits. The goal is simple: if content is synthetic, there should be a reliable way to prove it, and if it’s misused, there should be a way to trace it back to the source.