FTC: 87% of Retailers Use AI Pricing in 2025

A 2025 report from the Federal Trade Commission (FTC) says 87% of online retailers are now using algorithmic pricing, and a lot of them are wrestling with AI ethics inside their surveillance pricing models. With that kind of adoption, we have to ask if our existing legal frameworks actually protect consumers or promote fair competition anymore. The truth is, our legal infrastructure is struggling to keep up with the sheer complexity of AI-driven pricing.

87% of Online Retailers Employ Algorithmic Pricing

That 87% figure from the FTC’s 2025 report shows a fundamental change in how prices get set. This goes way beyond simple supply and demand adjustments. We’re talking about sophisticated algorithms analyzing huge amounts of data, your browsing history, your location, what kind of device you’re on, even the time of day you shop can all change the price you see on the screen. This allows for what many of us call “surveillance pricing,” since it’s built on watching everything consumers do. My read is that algorithmic pricing is no longer a niche strategy for a few tech giants. It’s just standard operating procedure across retail. This means regulations built for the old world of price tags are becoming useless. The laws have to catch up to address the underlying data-collection machine, not just the discriminatory prices it spits out. This is an industry-wide transformation, affecting almost everyone who shops online.

The Digital Markets Act (DMA) and Algorithmic Transparency

The European Union’s Digital Markets Act (DMA), which went into full effect in March 2024, is a serious piece of legislation aimed at the power of big tech platforms. While it’s not just about pricing, its rules on data access for “gatekeepers” hit surveillance pricing directly. Specifically, Article 6(1)(j) of the DMA says gatekeepers must give their business users access to the data they generate on the platform. Think about what that means: if a massive online marketplace uses AI to price products for the third-party sellers on its site, those sellers can now demand to see more of how those prices are being calculated. This is a real attempt to level the playing field. From my point of view, it’s a huge step toward algorithmic accountability. Any amount of transparency forces platforms to actually think about the fairness of their pricing models because the “black box” nature of AI presents a massive regulatory headache. If you can’t see the inputs, you can’t possibly judge the fairness of the outputs or spot discrimination. The DMA won’t fix everything, but it sets a powerful precedent for demanding a look inside these complex systems.

CCPA’s Right to Opt-Out and Data for Pricing

Here in the States, California’s privacy laws, the CCPA and CPRA, effective since January 2023, give consumers real power over their personal information. The most important tool against surveillance pricing is the right to opt-out of the “sale or sharing” of your data. The law’s definition of “sale” is very broad and includes sharing data for any kind of value, not just cash. So, if a retailer uses your browsing habits and location to show you a higher price than someone else, that could easily be considered “sharing” for a commercial purpose, and you’d have the right to opt-out. It’s a powerful idea, but how it applies to dynamic pricing is still getting hashed out in court. The hard part is proving a direct line from the data sharing to the specific price you were offered. For example, if a shopper in Beverly Hills sees a higher price for a jacket online than someone in a different zip code, connecting that price difference to the “sale” of their personal info is a tough legal battle. The CCPA gives consumers a way to react by pulling their consent, but it doesn’t regulate the algorithms proactively. My take? The CCPA is a good start, but its power against these pricing models depends entirely on strong enforcement and clear court rulings that call out algorithmic discrimination. Right now, the burden is on the consumer to figure it out and take action, which is often asking too much.

FTC’s Heightened Scrutiny on Discriminatory AI Pricing

The Federal Trade Commission (FTC) has been very clear that it’s going after AI-driven pricing, especially when it leads to discrimination or antitrust issues. In a 2023 policy statement, the agency reminded everyone that existing laws like the FTC Act already ban unfair and deceptive practices, and that includes those carried out by an algorithm. They specifically warned companies that using algorithms that produce discriminatory outcomes based on race, gender, or other protected classes is illegal, even if it wasn’t intentional. The FTC isn’t waiting around for a new law to be passed. It’s applying the rules already on the books to new technology. If a car insurance pricing algorithm, for instance, starts quoting higher rates to people in predominantly minority zip codes, even if the model claims to be race-blind, the FTC has the authority to investigate it as a discriminatory practice. I think this is exactly the right approach. Waiting for new legislation is too slow when AI is developing this fast. The big challenge for the FTC, of course, is proving the link between an algorithm’s code and a discriminatory result, particularly when the models are so opaque. Companies have to get serious about AI ethics in their own development processes. That requires rigorous, documented testing for bias and clear records of algorithmic decisions.

The Gap in Federal Legislation: A Patchwork Approach

Even with AI pricing everywhere, the United States still has no federal law that specifically governs AI ethics in this area. The result is a messy patchwork of state privacy laws, some industry-specific rules (like in banking), and how agencies like the FTC decide to interpret old consumer protection statutes. This fragmented system creates a ton of uncertainty for both businesses and consumers. That’s a huge vulnerability. Companies trying to operate nationally are staring down a complicated map of compliance obligations, and a person’s rights can change completely just by crossing a state line. Sure, existing laws have some teeth, and federal agencies are using their authority to go after bad actors. But without clear rules, enforcement usually happens after the damage is done, instead of preventing it in the first place. This legal ambiguity doesn’t help anyone, not consumers, and not the responsible companies trying to do the right thing.

Conventional Wisdom: AI Pricing is Pure Efficiency

There’s a common argument that AI pricing is just a hyper-efficient tool for matching supply and demand, creating the best outcome for everyone. People who believe this say it cuts down on waste and gives consumers more choices at competitive prices. They argue that what looks like “discrimination” is really just the market reflecting someone’s willingness to pay, and there’s no ethical problem there. I strongly disagree. That’s a dangerously narrow view. While AI can create efficiencies, calling it a neutral tool ignores the biases that are easily baked into these algorithms, whether you mean to or not. It also overlooks how surveillance pricing exploits the fact that the retailer knows way more about you and your habits than you know about the market. This is about power dynamics. When an algorithm is built to find and charge the most it thinks a vulnerable person will pay, it’s crossed the line from efficiency into predatory behavior. And the idea that price differences just reflect “market realities” is nonsense. These algorithms can *create* new realities, often by reinforcing existing inequalities or inventing new forms of discrimination. The ethical issues run deep. So for any company using AI pricing, the path forward is to build strong internal governance that puts ethics first. That means doing regular AI ethics impact assessments and making sure there’s a human in the loop with final oversight.