Colorado just stepped back from one of the country’s toughest proposed limits on AI-driven pricing and wage setting. The veto leaves businesses with room to personalize, but it also leaves consumers and workers waiting for clearer rules.
Jared Polis has blocked a Colorado bill that would have put sharp limits on how companies use personal data, algorithms and artificial intelligence to set individual prices and wages. That matters because this is no longer a theoretical fight about future technology. Retailers, platforms, insurers, lenders and employers already use data systems to sort people into offers, rates, discounts and opportunities.
House Bill 26-1210, known as the Prohibit Surveillance Price and Wage Setting bill, would have treated certain uses of surveillance data in price and wage algorithms as deceptive trade practices under Colorado consumer protection law. The measure passed the Legislature in May, with the Senate approving it 19 to 15 and the House later concurring on amendments. Polis vetoed it on June 2, saying the proposal was too broad and could sweep in ordinary business activity along with conduct lawmakers were trying to stop.
According to the Colorado General Assembly’s bill summary, the measure targeted price or wage setting algorithms that use statistical modeling, data analytics, artificial intelligence or other data processing techniques to analyze surveillance data when that output becomes a substantial factor in what a consumer is charged or what a worker is offered. That definition is important. It shows why the bill became such a clean test case for AI regulation: the same tools that can make markets more efficient can also make them more personal in ways people may never see.
Surveillance pricing is a simple idea with uncomfortable implications. A company does not just ask what a product is worth in the market. It asks what that product may be worth to you, based on browsing history, inferred income, location signals, health clues, family circumstances, biometrics or other personal characteristics. In wage setting, the concern is similar. A platform or employer could use data about a worker’s situation to estimate the lowest offer that person might accept.
Supporters of HB26-1210 argued that consumers and workers cannot meaningfully bargain in a market where the other side knows more than they do and may be using private signals against them. Consumer Reports backed the bill earlier this year, warning that personalized pricing can turn ordinary data collection into a tool for extracting the highest price from each shopper. Democratic sponsors, including Sen. Iman Jodeh and Sen. Mike Weissman, framed the measure as a transparency and fairness issue rather than a broad attack on technology.
Business groups saw it differently. The U.S. Chamber of Commerce urged Polis to veto the bill, arguing that the definition of surveillance data could cover common commercial practices and create uncertainty for companies using targeted discounts, loyalty programs, fraud tools and customer segmentation. The Financial Technology Association made a similar case, saying the legislation could disrupt ordinary ecommerce and financial services tools that help customers receive tailored offers.
Polis accepted that concern. His veto message said the bill did not define the bad conduct narrowly enough and could capture technology that only incidentally influences a price or wage. That is the central tension. Lawmakers want to stop companies from using intimate data to quietly charge more or pay less. Businesses want to preserve the ability to offer discounts, rewards, dynamic pricing and automated decision tools without facing a new enforcement risk every time software touches a price.
Statehouses are becoming the AI rulemakers
The veto does not end the issue. It probably accelerates it. Washington has not produced a broad federal law governing AI-driven pricing or algorithmic wage setting, so states are filling the space one bill at a time. Colorado has already been active on artificial intelligence policy, and HB26-1210 showed how quickly state lawmakers can move from general AI risk to very specific commercial practices.
For companies, the lesson is practical. Even without this bill becoming law, the political tolerance for opaque personalization is shrinking. If a business uses algorithms to set offers, wages or prices, it will need to explain what data is being used, why that data is relevant and where the guardrails sit. Saying the model decided will not be enough when regulators, workers or customers start asking whether the model used sensitive personal signals.
For consumers and workers, the veto means the burden remains mostly invisible. A shopper may not know whether a discount reflects a broad promotion or a prediction about willingness to pay. A gig worker may not know whether a pay offer reflects market demand or an inference about financial pressure. That uncertainty is exactly why advocates are likely to bring the bill back in a narrower form.
The next version will probably focus less on banning broad categories of tools and more on specific harms: sensitive data, lack of notice, inability to challenge wage data, and pricing practices that punish vulnerability. That would be harder for business groups to dismiss as a threat to ordinary discounts, and easier for lawmakers to defend as a targeted consumer protection measure.
Colorado’s veto gives companies more breathing room for now. It does not give them a blank check. The market wants personalization because it works, but the politics of personalization are changing fast. The businesses that understand that early will build systems that can survive both customer scrutiny and the next statehouse fight.
Also read: Meta scales back its employee tracking plan after AI data backlash • AI's data center footprint is becoming a country-sized constraint • Companies are learning how to game AI search through Reddit
