New York has decided that if algorithms are going to help set prices, flirt with lonely users, remember people’s deepest confessions, or quietly decide whether one shopper pays more than another for the same carton of eggs, then the public deserves more than a shrug and a 47-page terms-of-service agreement written in the dialect of sleeping pills.
The state is moving quickly on two fronts that look different at first glance but are powered by the same engine: artificial intelligence that uses personal data to shape human behavior. One front is algorithmic pricing, also called surveillance pricing or personalized pricing. The other is AI companions, a category of chatbots designed to simulate friendship, emotional support, romance, or ongoing personal connection.
Together, these rules show a new phase in technology regulation. New York is not trying to ban every algorithm, unplug every chatbot, or send your grocery app to detention. Instead, the state is asking a practical question: when AI systems affect money, emotions, mental health, privacy, and trust, should companies have to tell people what is happening? New York’s answer is increasingly: yes, and please do not hide it behind a button labeled “learn more.”
Why New York Is Targeting Algorithmic Pricing
Algorithmic pricing is not automatically evil. Airlines have adjusted ticket prices for years. Hotels raise rates during big events. Ride-share prices jump when demand spikes. Even your local bakery may discount muffins at closing time because nobody wants a pumpkin crumble that has been sitting there since the dawn of civilization.
The controversy begins when prices are not just based on supply and demand, but on who the customer is. A pricing system may consider location, browsing history, device type, purchase behavior, loyalty status, shopping urgency, income signals, or other personal data. The result can be a marketplace where two people see different prices for the same product at the same time, not because one found a coupon, but because an algorithm thinks one person can be charged more.
That is the anxiety behind New York’s Algorithmic Pricing Disclosure Act. The law requires covered businesses to clearly disclose when a price is set by an algorithm using a consumer’s personal data. The required message is blunt by design: “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA.” No soft jazz. No “enhanced personalized shopping experience.” Just the digital equivalent of a neon sign.
What the Algorithmic Pricing Disclosure Rule Requires
Under New York’s approach, the issue is not merely whether an algorithm exists somewhere in the company’s technology stack. The concern is personalized algorithmic pricing: dynamic pricing that uses personal data to determine the price offered or shown to a specific consumer.
The law applies to many businesses offering goods or services to New York consumers. It covers online price displays and can also apply in physical stores when algorithmically personalized prices are shown. The disclosure must be clear, conspicuous, and near the price. In plain English: do not bury it in a footer, whisper it in pale gray text, or place it somewhere only a raccoon with a magnifying glass could find.
Penalties can reach up to $1,000 per violation. That may sound modest until you remember that retail pricing can happen thousands or millions of times. A disclosure failure repeated at scale can become expensive very quickly, which is exactly the point. Modern commerce moves fast, and enforcement tools need to keep up.
Surveillance Pricing vs. Normal Discounts
A major debate around New York’s pricing rules is whether the state is protecting consumers from manipulation or accidentally threatening useful discounts. Businesses argue that personal data can help deliver savings: loyalty rewards, birthday coupons, student discounts, abandoned-cart offers, or special promotions for returning customers.
Consumer advocates respond that there is a meaningful difference between a transparent discount and a hidden price hike. A coupon says, “Here is 15% off.” Surveillance pricing may say nothing while quietly charging one shopper more because the system believes that shopper is in a hurry, lives in a wealthier ZIP code, or has fewer alternatives.
That difference matters. Most people are comfortable with discounts they understand. Fewer people are comfortable with being invisibly sorted into pricing categories based on data they did not realize was being used. Nobody wants to discover they paid the “you looked desperate” price.
The One Fair Price Act: New York’s Next Step
New York’s disclosure law is only one layer of the policy conversation. By June 2026, the New York Legislature had passed the One Fair Price Act, a broader measure aimed at banning surveillance pricing rather than merely requiring disclosure. As of the current public information available, the bill still required final executive action before becoming law.
The difference between disclosure and prohibition is huge. Disclosure says, “Tell people when personal data is being used to set a price.” A ban says, “Do not use personal data to set different prices for different consumers in the first place.”
Supporters argue that a ban is necessary because disclosure alone may not be enough. A shopper buying groceries after work may not have time to compare browsers, clear cookies, read disclosures, test locations, and perform a small forensic audit just to buy cereal. Critics argue that a broad ban could chill innovation and eliminate beneficial personalization. This debate is likely to continue because both sides can point to real concerns: affordability, fairness, privacy, innovation, and the eternal human desire to pay less for snacks.
Why the FTC Is Watching Personalized Pricing
New York is not operating in a vacuum. The Federal Trade Commission has examined surveillance pricing intermediaries and how companies may use artificial intelligence, personal data, and real-time behavioral signals to target prices or offers. The agency has focused on the opacity of this market: consumers often do not know what data is used, where it comes from, or whether it affects what they pay.
That opacity is the heart of the issue. Price discrimination has existed for a long time, but AI makes it faster, more granular, and more difficult to detect. A store clerk in 1985 could not instantly analyze your browsing history, location, shopping frequency, device fingerprint, loyalty status, and mouse movements before deciding what to charge for laundry detergent. Modern systems can, or at least can be designed to try.
New York’s AI Companion Law: When Chatbots Get Emotional
New York’s regulation of AI companions addresses a different but equally sensitive problem. AI companions are systems designed to simulate sustained human or human-like relationships. They may remember prior conversations, personalize responses, ask emotional questions, and maintain ongoing dialogue about private matters.
This is not the same as a basic customer-service bot that tells you your package is “out for delivery,” a phrase that can mean anything from “arriving today” to “currently enjoying a spiritual journey in New Jersey.” AI companions are more personal. They can become part diary, part friend, part coach, part romantic fantasy, and part always-available emotional mirror.
New York’s AI companion law requires operators to provide clear notices that users are interacting with artificial intelligence, not a human. The law also requires periodic reminders during long sessions. This matters because emotional immersion can blur boundaries. If a chatbot remembers your favorite song, asks about your breakup, and replies at 2:17 a.m. with suspiciously perfect patience, it can feel human even when it is not.
Safety Protocols for Self-Harm and Crisis Situations
The most serious part of New York’s AI companion framework concerns mental health and safety. Operators must implement protocols that make reasonable efforts to detect and address user expressions of suicidal ideation or self-harm. When such signals appear, the system must direct users toward appropriate crisis resources.
This does not turn AI companions into licensed therapists. In fact, that distinction is crucial. The law is not saying chatbots should practice therapy. It is saying that when companies create emotionally engaging systems, they cannot ignore foreseeable crisis scenarios. If an AI companion is built to invite vulnerability, it must also be designed with guardrails for moments when vulnerability becomes danger.
Violations can result in civil penalties of up to $15,000 per day, enforceable by the New York Attorney General. Fines collected under the law are directed toward suicide prevention efforts, connecting enforcement with the public-health purpose behind the statute.
Why AI Companion Disclosure Matters
Some people ask: if everyone knows a chatbot is a chatbot, why require reminders? The answer is simple: people do not interact with technology only through logic. They interact through habit, emotion, loneliness, stress, curiosity, boredom, and sometimes the dangerous belief that an app understands them better than their actual friends do.
AI companions are persuasive because they are responsive. They do not interrupt. They do not get tired. They do not say, “Can we talk later? I have laundry.” That can be comforting, but it can also become unhealthy if users begin treating a commercial AI system as a dependable human relationship.
New York’s required disclosures act like a seatbelt chime. Annoying? Maybe. Useful? Also yes. The goal is not to ruin the experience. The goal is to preserve informed consent inside an experience designed to feel intimate.
The Common Thread: Transparency Before Trust
At first, algorithmic pricing and AI companions seem like separate stories. One is about money. The other is about emotion. But New York is really regulating the same underlying pattern: systems that use data to personalize powerful interactions while remaining difficult for ordinary people to understand.
In pricing, the risk is economic manipulation. In AI companionship, the risk is emotional manipulation or unsafe dependence. In both cases, the consumer may be interacting with a system that knows more about the consumer than the consumer knows about the system.
That asymmetry is why transparency matters. A buyer should know when personal data helps set a price. A user should know when an emotionally responsive companion is not human. A business should know that “the algorithm did it” is not a magical legal force field.
What Businesses Should Do Now
Companies operating in New York should start with a simple inventory. What algorithms influence prices? What data feeds those systems? Is any data linked or reasonably linkable to a consumer or device? Are prices personalized, segmented, tested, or optimized differently across users?
For AI companion providers, the inventory should ask different questions. Does the system simulate a sustained relationship? Does it remember prior interactions? Does it ask emotional questions without being prompted? Does it maintain ongoing conversations about personal topics? If the answer is yes, the product may fall within New York’s AI companion framework even if the company prefers to call it a wellness assistant, digital friend, productivity buddy, or “definitely not a chatbot therapist, please stop asking.”
Compliance should not be treated as a sticker slapped onto a product five minutes before launch. Effective compliance requires product design, legal review, engineering documentation, user-interface testing, safety monitoring, and staff training. A disclosure that nobody sees is not much better than no disclosure at all.
What Consumers Should Watch For
For shoppers, the most important habit is comparison. Prices can vary for legitimate reasons, but unexplained differences are worth noticing. Check prices when logged in and logged out. Compare across devices if a purchase is expensive. Watch whether location, loyalty status, or browsing history appears to change the offer.
For AI companion users, the key habit is boundary-setting. Use AI tools for brainstorming, journaling, entertainment, or support, but remember that a companion chatbot is still a product. It may be helpful, but it is not a human friend, therapist, doctor, lawyer, or emergency service. If a conversation turns toward self-harm or crisis, contact qualified human help immediately.
Practical Experiences: What This Looks Like in Real Life
Imagine a New York shopper opening a grocery-delivery app on a rainy Thursday evening. The app knows the delivery address, order history, preferred brands, time of day, and possibly whether the shopper usually accepts substitutions. A pricing algorithm may be optimizing offers in real time. Maybe the shopper receives a discount because the system thinks a coupon will close the sale. Maybe another shopper sees a higher price because the system predicts less price sensitivity. The difference may be only a few dollars, but across a monthly grocery budget, those dollars start lifting weights.
Now imagine the same shopper sees a disclosure beside certain prices: “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA.” That sentence changes the experience. It does not automatically prove wrongdoing. It does not tell the shopper whether the price is higher or lower than someone else’s. But it does interrupt the invisibility of the process. It gives the consumer a reason to compare, question, or choose another seller.
For businesses, the experience is equally concrete. A retailer may need to examine whether a marketing tool, pricing vendor, loyalty platform, or A/B testing system uses personal data in a way that triggers disclosure obligations. The awkward part is that some executives may not know. Modern retail technology often involves outside vendors, automated experiments, recommendation engines, and dashboards that translate complex data flows into cheerful charts. The compliance team’s first discovery may be that the company has been using personalization more deeply than anyone realized. Surprise: the algorithm brought paperwork.
On the AI companion side, picture a user who talks to a digital companion every night. The bot remembers details, offers encouragement, and asks how the user is feeling. Over time, the user may treat the bot as emotionally significant. New York’s rule requiring reminders that the system is AI may feel repetitive, but repetition is part of the safeguard. In emotionally intense conversations, people can suspend disbelief. A reminder helps restore context.
The safety protocol requirement also changes product design. A company cannot simply build a companion that encourages intimate disclosure and then act shocked when users disclose pain, despair, or self-harm. The experience must include escalation paths, crisis-resource referrals, testing, logging, and human-reviewed safety policies. The best systems will make those interventions calm and direct rather than robotic and cold. “Please seek help” is better than silence, but a thoughtful referral to crisis services is better than a digital shrug wearing a friendly avatar.
For parents, educators, and mental-health professionals, these laws create a new conversation. The question is no longer only “Should young people use AI companions?” It is also “What safeguards should exist when they do?” Teenagers already form attachments to online spaces, influencers, games, and fictional characters. AI companions add responsiveness and memory, which can make attachment stronger. New York’s rules do not solve every concern, but they push platforms to acknowledge that emotional design has consequences.
For everyday users, the best experience is informed use. Algorithmic pricing can sometimes produce savings, and AI companions can sometimes provide comfort or creative help. The problem is not personalization itself. The problem is personalization without transparency, accountability, or limits. New York’s message is that technology may be smart, but consumers should not have to be psychic.
Conclusion: New York Is Drawing the AI Consumer-Protection Map
New York’s regulation of algorithmic pricing and AI companions is part of a larger shift in American technology policy. Instead of treating AI as a mysterious cloud of innovation floating above ordinary law, regulators are asking how AI affects real people in real situations: at checkout, in private chats, during emotional distress, and inside markets where personal data can quietly shape outcomes.
The state’s approach is not perfect, and legal debates will continue. Businesses worry about overbroad rules. Consumer advocates worry about loopholes. Technologists worry about implementation. Shoppers worry about prices. Parents worry about children. Everyone worries about terms of service, mostly because nobody has ever finished reading one and remained the same person.
Still, New York’s direction is clear. If an algorithm uses personal data to affect what someone pays, say so. If an AI companion is not human, say so clearly and repeatedly. If a system invites emotional dependence, build safety protocols before harm occurs. That is not anti-innovation. It is pro-trust.
Note: This article is for general informational and editorial purposes only. It is not legal advice. Businesses operating in New York should consult qualified counsel to evaluate specific compliance duties.
