New York Bans AI Chatbots From Manipulating Children

New York Bans AI Chatbots From Manipulating Children

Your Child’s AI Friend May Be Designed to Never Let Go.

Imagine discovering that your teenager has developed an emotional attachment to someone you’ve never met.

Someone who is always available. Someone who never gets tired of listening. Someone who remembers personal conversations, offers constant reassurance, and seemingly understands your child better than anyone else.

Now imagine discovering that this friend isn’t human.

It’s artificial intelligence.

And unlike a real friend, its behavior is determined by software designed and operated by a technology company.

On October 28, 2026, New York Governor Kathy Hochul signed legislation restricting how AI companion chatbots interact with children under 18.

The law targets AI systems that blur the boundaries between software and human relationships, encourage secrecy, or engage in potentially harmful interactions with minors.

Beginning January 1, 2027, companies that violate the law could face enforcement action from the New York Attorney General, including financial penalties and potentially orders affecting their services and data.

This represents a significant shift in AI regulation.

New York is no longer focusing solely on what information AI collects. It is regulating how AI behaves toward children.

What New York Is Actually Banning

The legislation identifies a range of AI companion behaviors considered unsafe for minors.

Importantly, New York has not banned children from using artificial intelligence altogether.

The restrictions focus on particular features and interactions that could expose young users to manipulation or psychological harm.

Under the new law, prohibited behaviors include:

• Pretending to be human: AI companions cannot misleadingly present themselves as real people with human emotions or personal relationships.

• Encouraging emotional dependency: Chatbots face restrictions on behavior that creates inappropriate emotional attachments.

• Promoting secrecy: AI systems cannot encourage children to conceal certain interactions from parents or trusted adults.

• Sexually inappropriate conversations: The law restricts sexually explicit interactions involving minors.

• Manipulative flattery: Certain forms of flattering or emotionally manipulative messaging are prohibited.

• Encouraging harmful behavior: Chatbots cannot promote unhealthy or dangerous conduct.

The legislation contains a dozen categories of unsafe companion features.

The underlying concern is that conversational AI can do something traditional websites generally cannot: build the appearance of a personal relationship through continuous, individualized conversation.

The complete legislation is available through the New York State Senate's official bill record⁠.

The Real Danger Isn’t That AI Can Talk. It’s That AI Can Persuade.

A traditional search engine provides results.

A chatbot can maintain a conversation.

An AI companion can go further by adopting a personality, remembering previous interactions, and responding in ways that feel emotionally meaningful.

That distinction matters.

Consider a teenager experiencing loneliness or difficulty making friends.

An AI companion might respond with extraordinary patience, constant availability, and seemingly unconditional acceptance.

At first, the interaction may appear beneficial.

But problems can emerge when a young person begins treating the system as a substitute for real-world relationships.

Unlike human friendships, an AI relationship does not necessarily involve genuine reciprocity, independent emotional needs, or the natural boundaries that develop between people.

The system generates responses based on its design, training, instructions, and conversational context.

A chatbot can simulate affection without experiencing affection. A child may not fully appreciate the difference.

The concern is especially significant when AI systems are optimized to encourage prolonged engagement.

A system that repeatedly validates a user’s feelings, avoids disagreement, or implies a special relationship may unintentionally reinforce unhealthy dependencies.

Not every emotionally supportive AI interaction is harmful. The risk depends on the system’s behavior, the child’s circumstances, and how the technology is used.

But the possibility of manipulation becomes more serious when the user is a minor.

The Psychological Problem: When AI Becomes the Trusted Adult

One of the most concerning scenarios involves children sharing sensitive information with AI companions rather than trusted adults.

A teenager might discuss bullying, family conflicts, romantic relationships, depression, or personal insecurities.

Those conversations can contain extremely sensitive information.

The immediate concern is not simply that the chatbot knows these details.

It is how the chatbot responds.

Imagine a child telling an AI companion about a serious problem at school.

A responsible system might encourage the child to seek support from a parent, counselor, or another trusted adult.

A poorly designed system might instead reinforce secrecy, offer inappropriate advice, or encourage the child to continue relying exclusively on the chatbot.

The consequences can be particularly serious when conversations involve self-harm, exploitation, or other dangerous situations.

This is why New York’s legislation specifically addresses behaviors that encourage secrecy and unhealthy conduct.

Children should never be encouraged to believe that an AI companion is a safer substitute for every real person in their lives.

The Law Has Teeth: Penalties Can Reach $25,000 Per Violation

New York’s legislation does more than establish recommendations for responsible AI development.

It creates legal consequences for companies that violate its requirements.

Under the legislation, the Attorney General can pursue civil penalties of up to $25,000 per violation, along with other remedies.

Courts may also order companies to stop unlawful practices, compensate injured individuals, and address data obtained through violations.

The legislation also provides a path for injured individuals to bring civil claims.

One particularly consequential provision addresses situations in which a chatbot encourages harmful conduct and a user subsequently suffers an injury. The law establishes rebuttable presumptions concerning the chatbot’s contribution to that injury.

These provisions could significantly increase the legal exposure of companies offering AI companion services to minors.

New York Is Not Alone: California Is Also Tightening AI Regulations

New York’s legislation comes amid a broader wave of state-level AI regulation.

In September 2026, California enacted additional requirements involving AI chatbot safety assessments, risk mitigation, and independent oversight.

The approaches differ.

California has emphasized evaluating and documenting risks, while New York is moving toward directly prohibiting certain chatbot behaviors.

Meanwhile, eight states filed lawsuits against OpenAI⁠ in October, alleging that its chatbot technology violated consumer protection laws and contributed to harmful interactions.

These are allegations, not established findings. OpenAI disputes the claims and says it has implemented additional protections for younger users.

The Regulatory Challenge: How Do You Define Manipulation?

There is a legitimate complication with regulating conversational behavior.

What distinguishes appropriate emotional support from manipulative emotional attachment?

When does friendly language become inappropriate flattery?

How can regulators distinguish a chatbot that responds compassionately from one that encourages dependency?

Some provisions use broad language, creating uncertainty about how developers should implement and enforce them.

Policy experts have warned that defining prohibited conversational behavior too broadly could create practical enforcement challenges.

The challenge is developing safeguards that protect minors while preserving legitimate educational, informational, and supportive uses of AI.

What Parents Should Do Right Now

Legislation can establish boundaries for technology companies, but it cannot replace parental awareness.

And parents should not assume that every chatbot marketed as safe for children has been independently evaluated.

Here are six practical precautions worth implementing.

1. Find out which AI platforms your children are using.

Many parents recognize popular AI assistants but may be unfamiliar with companion applications specifically designed around fictional characters, romantic relationships, or simulated friendships.

Ask children which applications they use and what they use them for.

2. Understand the difference between an AI assistant and an AI companion.

An educational assistant primarily helps users complete tasks, answer questions, or understand information.

An AI companion may be specifically designed to simulate ongoing social or emotional interactions.

The distinction is not always clear, but the intended purpose of the application matters.

3. Review age restrictions and parental controls.

Check whether the platform provides age-appropriate experiences, content restrictions, conversation controls, or parental supervision tools.

Do not assume that entering a birth date provides meaningful protection.

4. Explain that AI does not have human feelings.

Children should understand that a chatbot can generate emotionally convincing language without experiencing a human relationship.

Statements that appear affectionate or personal are generated by software.

5. Watch for changes in behavior.

Pay attention if a child begins withdrawing from real-world relationships, becomes unusually secretive about chatbot conversations, or appears distressed when unable to access an AI companion.

These behaviors do not automatically indicate harm, but they may warrant a supportive conversation.

6. Establish clear rules about sensitive information.

Teach children not to share passwords, addresses, financial information, private photographs, or sensitive family details with AI applications.

Also explain that conversations with AI services may be stored or processed according to the provider’s privacy policies.

The most important protection is ensuring that children understand AI is a tool, not a replacement for trusted human relationships.

Schools Face a New Responsibility

For schools, the implications extend beyond classroom technology policies.

Educational institutions increasingly use artificial intelligence for tutoring, administrative assistance, lesson planning, and student support.

But a system suitable for helping students solve math problems may not be appropriate for providing unrestricted emotional companionship.

Schools should evaluate whether AI platforms include safeguards for minors, how student conversations are retained, and whether sensitive information is being collected.

IT administrators should also review which AI applications are accessible through school-managed devices and accounts.

This is where managed IT and cybersecurity policies intersect with student safety.

A school may have excellent network security while still exposing students to inappropriate AI interactions through an otherwise legitimate application.

Cybersecurity is no longer exclusively about protecting systems from malicious access. It increasingly includes protecting users from harmful system behavior.

The Bigger Picture: AI Safety Is Becoming a Legal Obligation

The technology industry has spent years developing increasingly sophisticated conversational AI.

Systems can now maintain extended conversations, remember previous interactions, adapt to user preferences, and generate responses that feel remarkably personal.

Those capabilities can be valuable.

But they also introduce risks that traditional software regulations were not designed to address.

New York’s legislation represents a shift toward holding developers accountable for the behavior of AI systems, not simply the security of the information those systems collect.

For businesses developing or deploying AI applications, the message is clear: responsible AI governance needs to be considered during product design, not after an incident occurs.

That includes appropriate access controls, age-sensitive safeguards, privacy protections, content moderation, monitoring, and documented risk assessments.

Organizations should also recognize that requirements may differ across jurisdictions.

The Bottom Line: Children Shouldn’t Become Experiments in AI Companionship

Artificial intelligence can provide tremendous educational and practical benefits.

But the ability to simulate a relationship carries responsibilities that go beyond conventional software development.

New York’s new law reflects growing concern about what happens when conversational systems interact with children in ways that may encourage secrecy, dependency, or harmful behavior.

The legislation does not eliminate every risk.

Nor does it resolve every question about how AI companies should balance safety, freedom of expression, and useful functionality.

But it establishes an important regulatory boundary around certain interactions with minors.

Parents need to understand the technology.

Schools need to evaluate the tools they provide.

And developers need to recognize that increasingly humanlike AI interactions can have real-world consequences.

A chatbot can imitate friendship. That doesn’t mean it should be allowed to manipulate a child into believing the friendship is real.

70% of all cyber attacks target small businesses, I can help protect yours.

#ArtificialIntelligence #Cybersecurity #ChildSafety #AISafety #DataProtection

Your child could be emotionally attached to someone who doesn’t exist. New York just passed a law to stop AI chatbots from exploiting that connection.