On October 26 the Federal Trade Commission and 20 states sued Dollar General for charging shoppers more at the register than the price on the shelf. By then the company had already settled seven state cases over the same problem, and its own records showed tens of thousands of customer complaints. The lawsuit raises a broader question: why federal consumer protection so often arrives years after the harm is obvious, and what that means for regulating artificial intelligence.
The complaint, filed in the U.S. District Court for the Middle District of Tennessee, makes a simple accusation. Dollar General, which runs more than 21,000 stores, “has, for years, repeatedly mislabeled prices, overcharged consumers and failed to correct the issue,” according to the FTC’s announcement. The Commission voted 2-0 to file. “Dollar General has betrayed the trust of millions of American consumers,” said Christopher Mufarrige, director of the FTC’s Bureau of Consumer Protection.
The allegations are not yet proven. A court will decide them. RSN found no public response from Dollar General to the lawsuit at the time of writing. In a separate national class-action settlement in December 2025, the company denied wrongdoing and said it settled “to avoid further burdensome and costly litigation,” as reported at the time.
One summary of the case measures the delay from the earliest complaint posted online, arriving at about 19 years, starting with a 2007 post on a consumer-complaint website. That method does not work. The 2007 post was about rude clerks, not prices, and a count of posts on a review site does not measure law-breaking. The documented record is shorter. It is also, in some ways, more damning, because most of it was produced by government agencies.
What the complaint says regulators already knew
The FTC complaint builds its case largely on public enforcement records:
- Seven state settlements since 2019. “Since 2019, Dollar General has settled seven state actions with six different state Attorney General Offices and consumer protection agencies regarding its pricing inaccuracies,” the complaint says (para. 82). The settlements were Wisconsin (2019, $10,586), Vermont (2019, $1.75 million), Wisconsin again (2023, $850,000), New Jersey (2023, $1.2 million), Ohio (2023, $1 million), Colorado (2025, $400,000) and Pennsylvania (2025, $1.55 million). Payments “exceed $6 million” (complaint, para. 83).
- Failed inspections, state after state. State and local weights-and-measures inspectors repeatedly recorded overcharges. Missouri found failures in 92 of 147 inspections over 18 months ending in 2023. New Mexico failed 53 percent of 73 inspections. Iowa failed all eight. In Kansas, one store failed 44 inspections. In Wisconsin in 2023, 662 of 7,344 items checked were overcharged (complaint, paras. 67–77).
- Undercover purchases in 2026. The complaint cites undercover purchases at four stores between January and April 2026. At a Cleveland store, 13 of 58 items (22.4 percent) rang up wrong (complaint, para. 53).
- The company’s own complaint log. “From January 1, 2023, through March 31, 2026, Dollar General received at least 29,265 complaints categorized as ‘Did Not Honor Marked Price,'” from all 48 states where it operates (complaint, para. 9). That is roughly 750 a month.
- Internal reporting. The company “produces weekly and quarterly reports showing that many stores do not update pricing labels on time or at all” (complaint, para. 63).
The complaint’s own summary is blunt: “Despite repeated red flags, Dollar General has failed to correct the issue and continues to make false pricing statements” (complaint, para. 63).
One detail is easy to miss. The 29,265 complaints were made to Dollar General, and the FTC obtained them in its investigation. They are not complaints the agency had sitting in its own files. The inspection failures and state settlements were public, however, and some of them go back seven years.
The case that the FTC was not simply asleep
Before blaming the agency, it is worth looking at the reasons the FTC is structurally slow. Some of them are good reasons.
Price accuracy has long been a state job. Checking scanner prices against shelf tags is the work of state and local weights-and-measures offices, and 33 states use a model price-checking procedure published by the National Institute of Standards and Technology (NIST), under which a store fails if more than 2 percent of the items checked ring up wrong (complaint, para. 65). The federal role has been to write the model procedure, coordinate and backstop. The FTC first studied the problem with NIST and officials in 37 jurisdictions in 1996. Its 1998 follow-up found that 71 percent of store inspections in 37 jurisdictions met the 98-percent-accuracy standard. Against that baseline, the inspection results in the Dollar General complaint are far worse. But the inspections themselves were the states’ work, and so were the first lawsuits.
The Supreme Court took away the FTC’s main refund tool in 2021. For decades the FTC used Section 13(b) of the FTC Act to win refunds for consumers in federal court. In AMG Capital Management v. FTC (April 2021), the Supreme Court unanimously held that Section 13(b) does not authorize monetary relief. Then-Chair Lina Khan told Congress in 2022 that “future Commission distributions will decrease due to the loss of the ability to obtain monetary relief.” Since then, a deceptive-pricing case that ends with refunds has needed a separate legal basis.
The Dollar General complaint shows how the agency now works around that. Alongside the FTC Act, it charges the company under Section 521 of the Gramm-Leach-Bliley Act, a 1999 banking-privacy law that bans obtaining a customer’s financial information “by making a false, fictitious, or fraudulent statement.” The FTC’s theory is that a false shelf price is a false statement made to get shoppers’ card and bank details at checkout (complaint, para. 114). Because violations of that statute can be enforced like violations of an FTC rule, they open the door to court-ordered refunds under Section 19 (complaint, paras. 112–113). A federal judge in New York, Jed Rakoff, accepted this reading in 2023 in FTC v. RCG Advances. Whether courts will accept it for a nationwide retailer is an open question. That a pricing case has to rely on a 1999 anti-pretexting law shows how limited the agency’s direct remedies have become.
Congress could fix this. The House passed the Consumer Protection and Recovery Act to restore the FTC’s refund authority on July 20, 2021, by a 221–205 vote. Republicans objected to its 10-year limitations period and its breadth, according to Kelley Drye’s account of the debate, and the bill stalled in the Senate. Democrats reintroduced it in 2026, and its prospects are described as a “steep climb.”
The agency is small, and getting smaller. The FTC’s budget justification for fiscal 2027 lists 1,224 full-time staff in fiscal 2025 and an estimated 1,183 in fiscal 2026. It assigns 583 of them to consumer protection. In 1980 the agency had 1,719 staff, according to Khan’s 2022 testimony. Congress funded the FTC at $383.6 million for fiscal 2026, down from $425.7 million the year before. Meanwhile, consumers filed 3 million fraud reports in 2025 and reported $15.9 billion in losses, according to FTC testimony to Congress in March 2026. A consumer-protection staff of about 583 full-time positions cannot investigate every pattern that shows up in that data.
Big cases take years even when the agency moves. The FTC began investigating Amazon’s Prime sign-up and cancellation practices in 2021, sued in 2023, and settled in September 2025 for $2.5 billion, including $1.5 billion in refunds and a $1 billion civil penalty, as reported at the time. That case had a statute behind it, the 2010 Restore Online Shoppers’ Confidence Act, that allows monetary relief. It still took four years.
The case that it was too slow anyway
None of this explains why the problem continued for so long after it was well documented.
The state settlements did not work. Ohio collected $1 million from Dollar General in 2023, yet inspections in the state still showed double-digit error rates afterward, Axios reported in December 2025. The FTC complaint confirms the pattern: Lake County, Ohio, failed half its inspections in early 2026, and more than 36 percent of re-inspections also failed (complaint, para. 73). Tennessee’s failure rate rose from 5 percent in 2025 to more than 10 percent in early 2026 (complaint, para. 78). Each state settled on its own. Each settlement was a cost of doing business, not a reason to change.
The amounts make the point. Seven years of state payments, a little over $6 million, compare with net sales of $42.7 billion in 2025 alone (complaint, para. 43). That is about 0.014 percent of one year’s sales. When penalties are that small and arrive one state at a time, the cheaper course for a company is to keep paying them.
This is the gap a national regulator exists to close. One federal case, with injunctive relief that binds every store, can do what seven state settlements did not. By the FTC’s own account, the warning signs were “repeated” and public. The agency filed suit in October 2026, roughly ten months after the Guardian published a national investigation of dollar-store overcharging in December 2025. According to the complaint, an email from Dollar General’s chief executive about pricing accuracy followed that article (para. 87).
Political texts: a loophole that regulators did not create and cannot close
A common question is why Washington cannot stop the flood of political texts and calls, when the Do Not Call Registry was supposed to fix this two decades ago. The answer is that the registry was never designed to cover politics, and the law that does reach some political messages was narrowed by the Supreme Court.
The Do Not Call Registry is enforced through the FTC’s Telemarketing Sales Rule, which covers sales calls. The FTC’s own guidance says “political solicitations are not covered by the TSR at all,” and neither are charities calling on their own behalf or pure surveys. Signing up for the registry was never going to stop campaign texts.
The other law, the 1991 Telephone Consumer Protection Act, does apply to political messages. It bans autodialed or prerecorded calls and texts to cell phones without consent, regardless of who sends them. When political consultants challenged that ban in 2020, they lost. In Barr v. American Association of Political Consultants, the plurality opinion announcing the Court’s judgment said that “plaintiffs still may not make political robocalls to cell phones.”
Texts are where the protection broke down, in two steps:
- In June 2020, the FCC ruled that texting platforms requiring a person to click “send” for each message are not autodialers, and that high volume alone does not make a system one. This is the “peer-to-peer” model most campaigns now use.
- In April 2021, in Facebook v. Duguid, the Supreme Court held that an autodialer must have the capacity to use “a random or sequential number generator.” Campaigns text from curated voter lists, not random numbers. Justice Sotomayor wrote that the plaintiff’s “quarrel is with Congress, which did not define an autodialer as malleably as he would have liked” (opinion).
When the FCC tried to tighten consent rules for commercial calls and texts, the Eleventh Circuit vacated the rule in January 2025, holding that the agency “impermissibly exceeded its statutory authority.” The volume shows the result. Industry estimates compiled by Fox 13 put political texts at roughly 350 million in 2018, and RoboKiller counted about 15 billion in 2022.
So could the government end this easily? The regulators cannot. They would need a new statute, and the courts have recently been striking down agency attempts to stretch old ones. Congress could. It could update the TCPA’s definition of an autodialer, or create an opt-out list for political texts, as at least one 2026 congressional candidate has proposed. Any such law would have to survive First Amendment review. Barr shows that the Court will strike down rules that favor some kinds of speech over others, but it also upheld a neutral restriction on robocalls. The more obvious obstacle is incentive, and this part is RSN’s analysis rather than established fact: the people who would have to write the law are the main users of the loophole, in both parties.
What this means for regulating artificial intelligence
The Dollar General case and the texting loophole show the same pattern. Federal regulators act quickly when an existing statute clearly fits the conduct. They act slowly, or not at all, when it does not, and they then wait for Congress.
The clearest AI example proves the point. Days before the January 2024 New Hampshire primary, voters received robocalls using an AI-generated imitation of President Biden’s voice. In February 2024 the FCC ruled that AI-generated voices count as “artificial” voices under the 1991 TCPA. It later fined the political consultant behind the calls $6 million, and the carrier that transmitted them agreed to pay $1 million, according to New Hampshire Public Radio. The federal response was fast because a 33-year-old statute already covered “artificial” voices. The criminal case went differently: in June 2025 a New Hampshire jury acquitted the consultant of all state voter-suppression and impersonation charges.
Most AI harms will not fit an existing statute so neatly. Current federal policy also leans toward limiting the regulators that act first:
- Executive Order 14365, signed December 11, 2025, created an AI Litigation Task Force at the Justice Department to challenge state AI laws. It also directed the FTC to explain when state laws requiring changes to AI outputs are preempted (Federal Register).
- The White House’s March 2026 National Policy Framework asks Congress to preempt state AI laws that impose “undue burdens,” while preserving state child-protection enforcement and calling for stronger action against AI-enabled fraud. It is not binding. Congress declined in 2025 to adopt a temporary moratorium on state AI laws and has not passed a comprehensive federal AI law.
- In July 2026 the FTC proposed a policy statement that would treat undisclosed, deliberate steering of AI outputs as a potential deceptive practice, and that names Colorado’s AI law as an example of a state law that may be preempted. It is still a proposal.
- In December 2025 the FTC set aside a 2024 order against Rytr, an AI writing service accused of enabling fake reviews. The Commission said the original complaint “failed to satisfy the legal requirements of the FTC Act.” It said it “will continue to hold accountable actors that use AI to violate the law or deceive consumers.”
- In September 2025 the FTC opened a study of AI “companion” chatbots and their effects on children, sending orders to seven companies. It is a study, not an enforcement action.
There is a reasonable conservative argument for this restraint. A patchwork of 50 state AI codes would burden small developers. An agency that condemns a product “simply because it potentially could be used in a problematic manner,” as the FTC’s consumer protection director put it in the Rytr matter, is regulating hypothetical harm rather than actual fraud. RSN has made similar arguments about regulators that overreach, from a 2016 bill to expand the FTC’s role over contact-lens sellers to the CFPB’s 2015 auto-lending cases built on guessed borrower race.
The Dollar General case suggests a caution that conservatives, who usually defend federalism, should weigh. In that case the states moved first. State inspectors found the overcharges, and state attorneys general secured seven settlements before the federal government sued. If AI follows the same path, with federal action arriving years after the harm and only when an old statute happens to fit, then the states are the early-warning system. Preempting state AI enforcement before Congress has written a federal standard would remove the only layer that has reliably acted first. It would not replace it with anything. The forged documents RSN reported this week in Britain’s immigration tribunals, which the tribunal warned AI will make easier to produce, were caught by ordinary verification checks, not new AI laws. That is a reminder that enforcing general fraud law well matters more than writing new AI-specific codes.
Where the federal government actually is
The federal government is present on consumer protection, but its reach is narrower and its timing later than most people assume. It wins large cases when Congress has given it a statute with remedies, as with Amazon. It struggles when it must improvise, as the Gramm-Leach-Bliley count against Dollar General shows. When it tries to write broad rules, courts often stop it: the Eighth Circuit vacated the FTC’s “click-to-cancel” subscription rule in July 2025 on procedural grounds, days before it was to take effect. In June 2026 the Supreme Court decided in Trump v. Slaughter that presidents may remove FTC commissioners at will, ending the agency’s formal independence. The FTC’s priorities will now follow the administration in office more closely.
None of that is a scandal by itself. It is how the system was built: Congress writes narrow statutes, courts read them narrowly, agencies are funded below the scale of the problem, and the states fill gaps. The cost of that design falls on the people who notice a small overcharge at a dollar store and lack the time to fight it. For a family on a fixed income buying most of its groceries there, that cost is not trivial.
What is established, disputed and unknown
Established: Seven state settlements over Dollar General’s pricing since 2019, totaling more than $6 million. Repeated public inspection failures across many states. The FTC and 20 states filed suit on October 26, 2026. The Supreme Court ended the FTC’s 13(b) refund authority in 2021, and Congress has not restored it. Political calls and texts fall outside the Do Not Call Registry, and Duguid narrowed the TCPA’s reach over campaign texting platforms.
Alleged, not proven: That Dollar General knowingly overcharged consumers nationwide and failed to fix the problem. These are the FTC’s and the states’ allegations, and the company has not yet answered in court.
Not established: That consumers complained about pricing for 19 years before regulators acted. How much consumers were overcharged in total; the complaint says “millions of dollars” and gives no figure. When the FTC opened its investigation. Whether courts will accept the Gramm-Leach-Bliley refund theory in a retail pricing case.
What accountability would look like
- Restore refund authority with guardrails. Congress should give the FTC back the power to return money to defrauded consumers, with a shorter limitations period and a focus on deliberate deception, which answers the main Republican objections of 2021. The current workaround of stretching a banking-privacy law is not a durable basis for enforcement.
- Make repeat state findings trigger federal review. When a national chain settles with multiple states over the same practice, the FTC should be required to review the pattern and say publicly whether it will act.
- Close the texting loophole by statute. Congress should update the TCPA’s definition of an autodialer for modern texting platforms and give voters a viewpoint-neutral way to opt out of campaign texts.
- Do not preempt before you legislate. Any federal AI law that displaces state enforcement should first give consumers equivalent federal protection against AI-enabled fraud, and should keep states’ authority to enforce their general consumer-protection laws.
Dollar General’s shoppers did not need a new law. They needed an existing law enforced before the seventh settlement. If that is how regulation works for shelf prices, readers should not expect regulators to keep pace with AI unless Congress gives them the tools to do so.
Sources: FTC et al. v. Dollar General Corp., complaint (M.D. Tenn., filed Oct. 26, 2026); FTC press release (Oct. 26, 2026); FTC, Price Check II (Dec. 16, 1998); FTC FY 2027 Congressional Budget Justification; FTC testimony to the Joint Economic Committee (Mar. 25, 2026); Chair Khan testimony (May 18, 2022); FTC Do Not Call business guidance; Facebook v. Duguid (2021); Barr v. AAPC (2020); Executive Order 14365; plus the news and legal-analysis sources linked in the text.
Featured image: The Federal Trade Commission Building (Apex Building), Washington, D.C. Photograph by Carol M. Highsmith, Library of Congress, Prints and Photographs Division (LC-DIG-pplot-13734-01526), public domain, via Wikimedia Commons.
How this article was made: Right Side News researched and drafted this article with the help of AI tools. Sources are linked so readers can check them. Spot an error? Tell us in the comments and we will correct it.