AI Transparency: Vermont Leads the Way
Vermont’s new law regulating synthetic media in elections is a critical step in setting standards for AI transparency in civic life.
Artificial Intelligence has made it easier than ever to create hyper-realistic media in a matter of seconds. While this capability is great for entertainment or artistic projects, it causes massive headaches for civic engagement and voting. Vermont recently took a decisive step forward, as Governor Phil Scott signed bill S 23 into law. This legislation marks a new era for AI transparency in political campaigns.
The bill targets the rise of "synthetic media" and deepfakes within election materials. Lawmakers in Vermont decided that voters deserve to know when media has been artificially generated, especially when that content aims to sway an election. By mandating clear disclosure for manipulated material, Vermont is positioning itself as a leader in state-level AI governance.
Why does this struggle between technology and policy feel so urgent right now? Because our current information ecosystem makes it difficult to verify reality at scale. When voters see a video of a candidate speaking, they naturally assume it captures a real event. But when that footage is synthetic, the barrier to mass deception is lowered significantly.
Governance around AI often feels like it lags years behind the technology’s actual progress. Vermont’s law is interesting because it’s proactive rather than reactive. It provides a clear framework that sets expectations for what counts as transparent information, rather than waiting for a crisis to define the rules.
This also forces us to consider where our trust actually lives. Do we trust the candidate, or do we trust the platform that hosts the video, or perhaps the legal framework that demands the disclosure of truth? As these disclosures become mandatory, we’re essentially building a "nutrition label" for political media. It helps audiences better navigate what they are seeing in their social feeds.
The challenge ahead is simple but difficult to enforce: how do you catch high-quality deepfakes if they arent labeled? Disclosures help keep the conversation honest for those who follow the rules, but the law wont stop a bad actor who intentionally ignores a label.
Clear laws like this are an important baseline. They provide a clear standard for political parties and creators. It’s a necessary evolution for our modern digital democracy.
Source: https://www.transparencycoalition.ai/news/ai-legislative-update-march6-2026