The Push for Bilingual AI Regulation in the US Senate

When the tourists thin out and the cherry blossoms disappear in late April, a certain silence descends upon Senate hearing rooms, and staff members in the back rows begin to whisper about which bills will actually move before the August recess.

That’s the silence I’ve been thinking about a lot lately because, somewhere within it, a strange new idea has started to take shape: that the most futuristic battle in Washington, artificial intelligence policy, might ultimately turn on something as archaic as language.

There isn’t yet a single bill advocating for bilingual AI regulation. It’s more of a system of pressure between state lawmakers, Spanish-speaking constituencies, and a few senators who have begun to question why all consent forms, chatbot disclosures, and “you are talking to a machine” notifications are written in English first and then, if at all, translated. It’s possible that what appears to be a translation problem is actually more significant, raising concerns about the people AI is intended to safeguard.

This gets interesting in Florida. The “AI Bill of Rights” proposed by State Senator Tom Leek mandates that tech companies notify users when they are interacting with a machine rather than a human. It sounds easy. However, the disclosure issue quickly becomes complex in a state where about 25% of people speak Spanish at home.

A notification buried in English-only legalese is merely a formality rather than a notification at all. Sen. Carlos Guillermo Smith, a Democrat from Orlando, has been focusing on the bill’s enforcement gaps. You can hear him questioning the kind of annoyance that arises when a policy created for one community is applied carelessly to another.

Then there is Trump’s executive order from last fall, which threatened broadband funding for states that violate state laws and established an AI Litigation Task Force to challenge them. One rulebook, no patchwork, and no fifty different sources of approval are all part of the administration’s pitch for unity. Tech lobbyists adore it. According to NetChoice, it’s a step toward “smart, unified federal policy.” However, 280 state lawmakers from both parties signed a letter opposing it. Liz Larson, a Democrat from South Dakota, stated unequivocally that it would be reckless to remove state authority without providing a substitute framework.

It’s difficult to ignore how rapidly the debate has divided when observing this from a distance. Industry voices contend that AI is developing too quickly for disjointed regulation. However, state lawmakers who have already enacted consumer protections, algorithmic discrimination regulations, and deepfake laws don’t think Washington can improve. The bilingual question sits somewhere in the middle, almost as an afterthought.

A few senators have begun subtly proposing language requirements in committee, such as disclosure guidelines in both Spanish and English, reports on AI safety incidents translated into both languages, and chatbot opt-ins accessible in the user’s preferred language. It hasn’t made it to the floor. Speaking with employees, however, gives me the impression that something is coming together. By default, the EU’s AI Act supports 24 official languages. So far, only one is covered by American AI law.

Whether any of this is passed into law in 2026 is still up in the air. Congress has already made two unsuccessful attempts to supersede state AI laws. Tallahassee will continue to host the Florida battle. Additionally, the bilingual push is currently more of a murmur than a movement. However, murmurs in Senate cloakrooms have a way of becoming legislation—sometimes years later, sometimes more quickly than anyone anticipated. This one doesn’t seem to be finished forming yet.