5:23 PM, October 31, 2026

On October 20, 2026, the New York City Department of Consumer and Worker Protection (DCWP) and Commissioner Samuel A.A. Levine proposed a rule that would impose stricter requirements for customer service AI chatbots.

By now, we have all been a witness to some sort of AI blunder. Chipotle’s chatbot, Pepper, responds to customers in Python code instead of helping them with their burrito order. Facebook AI openly admits to users that it makes up false information. Woolworths’s AI agent references its “mother” in chats. The ways chatbots mess up is sometimes comical, but for consumers, this is turning into no laughing matter.

While it is common knowledge that AI is faulty, what the public does not know is that some of these “blunders” are attempts to manipulate consumers. Corporate sludge, which slows down business transactions with the intended result of exhausting people, has become increasingly prevalent. AI advancement only furthers the existing agenda to get consumers to “give up” on fighting refunds, cancellations, and monetary losses. Specifically, there has been an uptick in companies’ use of AI chatbots to run their customer service departments.

Although on the surface it may appear that these companies are just trying to cut costs, there are even more sinister reasons for replacing human representatives with AI.

Corporations that we all begrudgingly deal with are notorious for their chatbot customer service representatives. Cable providers, delivery companies, and subscription services are some of the biggest offenders of staffing their customer service departments with bots. Although on the surface it may appear that these companies are just trying to cut costs, there are even more sinister reasons for replacing human representatives with AI. Using chatbots has made it easier for these corporations to implement corporate sludge. Some common examples are intentionally dropped calls, long phone hold times, and multi-step subscription cancellation processes (also known as “dark patterns”). 

The chatbot “doom loop” strategy fully reveals the harmful effects of AI customer service bots. The “doom loop” is a corporate tactic that makes it difficult—or even impossible—to escape AI chatbots. Getting a human representative on the line may take hours, which many working-class individuals do not have to spare. These “doom loops” are tiring to customers, and many people are forced to accept monetary losses due to the excessive difficulty of speaking with an actual person. 

By wasting time, companies have been successful in retaining these customers against their will. It is easy to feel hopeless when it comes to these frustrating practices. The result has been general societal apathy, now bordering on acceptance. Since individual consumers feel unable to take on this fight, they need someone to stand up for them. Unfortunately, the federal government has stepped back from enforcing bans on unfair and deceptive business practices, which is why NYC’s DCWP is pushing to hold companies that use AI chatbot customer service accountable.

Because chatbot customer service is preventing customers from getting any sort of meaningful help and therefore causing financial strain, it is arguably a violation of Unfair or Deceptive Acts or Practices (UDAP). More specifically, this practice would easily fall under the “unfair” and “deceptive” prohibitions. 

There are many avenues in which this practice could be classified as a UDAP violation. First, companies frequently make claims that robust customer service is accessible but are unable to deliver due to the nature of their AI chatbots. They can face further liability because the design of their chatbots is obstructive and non-responsive. Companies’ full reliance on an AI model that causes customers to fall into “doom loops” makes this an easy case for non-responsive service. Finally, because these models are poorly built and inadequately implemented, they cannot carry out their so-called intention. Since these models are so unreliable and regularly spew out misleading information to consumers, attacking AI customer service under UDAP could be a viable way to prevent further financial harm. 

DCWP’s proposed rule would make it a deceptive and unconscionable trade practice for businesses to claim that customer service is available while at the same time responding to customers’ inquiries with unhelpful, irrelevant, or blatantly erroneous information. Specifically, the rule would:

• Require businesses to disclose when a customer is talking to an automated customer service system such as AI;
• Require businesses to provide an easy offramp to speak to a human representative;
• Prohibit businesses from evading responsibility for using AI that results in unlawful conduct;
• Require businesses to actually offer meaningful customer assistance if it is advertised.

Following the release of this proposal, stakeholders have a 30-day comment period in which they are able to provide feedback. After this comment period, DCWP will review submissions before the rule is finalized. Whether or not the rule is ultimately adopted, this proposal shows a shift toward prioritizing consumers over chatbots. 

Rachael Johnston

Rachael graduated from Randolph-Macon College with a B.A. in English and Spanish. Currently, she is a 2L at the University of North Carolina School of Law. Rachael is a Staff Member on the North Carolina Journal of Law and Technology and serves on the Executive Board for the Consumer Legal Advocacy Society.