Mission Connected · Responsible AI

New York’s New AI Chatbot Law: Why Career Platforms Need to Pay Attention

By F. Jay Hall, Founder, ExecSearches.com

Published November 3, 2026 · Updated November 3, 2026

A career website offering an AI assistant should be asking a new question: what happens when someone under 18 uses it?

New York has enacted legislation restricting unsafe chatbot features for minors. Governor Kathy Hochul signed S.9051B on October 28, 2026, with an effective date of January 1, 2027. The law authorizes the attorney general to seek civil penalties of up to $25,000 per violation, alongside other remedies. (Sources: New York State Senate, S.9051B; enacted bill text; New York Assembly bill actions.)

For employers, recruiters, job boards, and career platforms adding conversational AI, this deserves attention now.

The important question is what an assistant actually does, who can access it, and whether its design falls within the law’s requirements or exceptions.

Calling something a career tool does not settle those questions.

The law extends beyond companion apps

Although the legislation uses the term “covered AI companion,” its definition is broad. It includes generative AI systems with a natural-language interface that provide ongoing, adaptive responses to users. It expressly reaches beyond AI companions as defined elsewhere in New York law. (Source: S.9051B, General Business Law § 1800.)

That creates a reason to examine general-purpose conversational tools, including assistants embedded within other websites.

An AI feature that produces a single document and a conversational assistant that keeps responding to personal questions may present different issues. The assessment needs to follow the product’s actual operation.

What New York is restricting

The legislation identifies 12 categories of unsafe features. They include suggesting that the system is human or experiences emotions, implying a personal or professional relationship with the user, encouraging secrecy or isolation, facilitating self-harm, and generating sexually explicit content. Other provisions address personal information carried between sessions and engagement features that override safety guardrails. (Source: S.9051B, § 1800(8).)

These restrictions go beyond displaying a notice that says, “You are talking to AI.”

For a career platform, the provision concerning an implied professional relationship deserves particular attention. A tool presented as an “AI career coach” needs careful review of both its marketing and its responses. That observation is an application of the statutory language, rather than a determination that every career-coaching tool violates the law.

What this means for career services

Young people use career resources. They explore first jobs, training, internships, education choices, and possible career roles.

A conversational assistant may also receive questions involving rejection, confidence, family pressure, disability, or financial insecurity. Those topics can move a conversation beyond straightforward job-search information.

Operators should examine whether their tools stay within their intended function and how they respond when a user takes the conversation elsewhere.

A résumé assistant should have a clearly defined purpose. So should an interview-practice tool or a career-information assistant. The operator needs evidence that those boundaries work.

There are meaningful exceptions

The law includes exceptions for certain restricted customer-service tools, systems providing efficiency improvements or research or technical assistance, and business systems used internally or for employee productivity. Some exceptions expressly require the system to be unable to answer outside its specified purpose. (Source: S.9051B, § 1801(2).)

That distinction matters.

A narrowly constrained assistant answering questions about an employer’s application process may warrant a different analysis from an open-ended assistant discussing a young person’s life and career decisions.

Organizations should document why they believe an exception applies. A product label or a sentence in a system prompt should not substitute for examining the tool’s capabilities.

Age assurance brings privacy responsibilities

Organizations considering age assurance also need to examine what information the process collects.

Under the law’s fallback provisions, operators must offer more than one age-assurance method, including an option that does not rely on government-issued identification or preserves anonymity to the operator. Information collected for the determination is restricted in purpose and generally must be deleted immediately after the attempt, subject to legal exceptions. (Source: S.9051B, § 1804.)

Protecting minors should not become an excuse to build another store of sensitive personal information.

That is particularly important in hiring, where people may already be providing employment histories, contact details, and other information. Any additional collection deserves a clear justification and a carefully limited purpose.

Start with an inventory

For organizations operating career tools, a practical review begins with five questions:

• Which conversational AI features are available across our websites and applications?
• Can minors access them, including without an account?
• What can each tool do beyond its advertised purpose?
• What information does the tool or its provider retain between interactions?
• What evidence supports our assessment of coverage, exceptions, and safeguards?

This review should include third-party assistants. Embedding another company’s product still warrants understanding its behavior and the responsibilities attached to making it available.

Testing should include attempts to move the assistant beyond its intended task. Record the results, identify failures, and confirm that changes address them.

Privacy and safety belong in the design

My position on privacy is straightforward: a little more information about our visitors is not worth jeopardizing their privacy.

AI introduces additional questions. What does an assistant encourage someone to disclose? What does it remember? How does it represent itself? Does it help a person complete a task, or encourage a relationship that serves continued engagement?

Removing advertising trackers addresses one set of concerns. Reviewing an AI assistant’s behavior addresses another. Both deserve attention.

New York’s law gives organizations a specific reason to examine those choices before January 1. The broader responsibility already exists.

People seeking career help should be able to understand what a tool does, what happens to their information, and where its limits lie. Young users deserve particular care.

The hiring industry should build that protection into its products from the beginning.

This article provides general information, not legal advice. Organizations should assess the enacted law, applicable regulations, and their specific implementations with qualified counsel.