CHAT Act 2.0
- Sponsor
- Sen. Husted, Jon [R-OH]
- Committees
- Commerce, Science, and Transportation Committee (primary)
- Last action
- Jul 28, 2026
Bottom line
The CHAT Act 2.0 aims to protect minors from potential harms associated with AI companion chatbots by implementing a tiered system of safeguards, age verification, and data privacy requirements for covered entities.
What it actually does
This bill establishes a comprehensive regulatory framework for companion AI chatbots, categorizing them into three tiers based on their purpose and functionality. It imposes specific obligations on entities operating these chatbots, including age assurance requirements, parental controls, crisis intervention protocols for self-harm and suicidal ideation, and strict prohibitions on sexually explicit communication with minors. The bill also mandates data protection measures, such as prohibiting targeted advertising to minors and requiring verifiable parental consent for data sharing.
Proponents argue
Proponents argue that this bill is a crucial step in safeguarding children and adolescents from the emerging risks posed by advanced AI technologies, particularly those designed for companionship and emotional interaction. They contend that the tiered approach allows for tailored protections while fostering responsible innovation, preventing exploitation, and ensuring that parents have the necessary tools to monitor their children's digital interactions.
Opponents contend
Opponents might argue that the bill's stringent age verification and parental consent requirements could create significant barriers to access for minors who could benefit from educational or benign companionship AI. They may also raise concerns that the broad definitions and compliance burdens could stifle innovation in the AI sector, lead to over-blocking of content, or infringe on the privacy of older minors by mandating parental notification for sensitive topics.
The bill is relatively concise and well-structured, making it accessible for an informed reader to understand its core provisions and implications within a reasonable timeframe.
Section 2
Definitions of Companion AI Chatbot Tiers
This section defines 'companion AI chatbot' and categorizes them into three tiers: Tier I for primarily educational purposes, Tier II for simulating interpersonal interaction/companionship, and Tier III for mental health treatment or quasi-therapeutic support. This tiered structure dictates the specific regulatory requirements that apply to each type of chatbot.
Supporters argue
Proponents argue that a tiered approach is essential for effective regulation, allowing for tailored protections based on the specific risks associated with different types of AI interactions, rather than a one-size-fits-all approach that could be overly restrictive or insufficient.
Critics contend
Critics might contend that the definitions could be ambiguous in practice, leading to disputes over classification or creating loopholes for chatbots to avoid stricter regulations by claiming a different primary purpose. They may also argue that the distinctions are arbitrary and could hinder innovation.
Tradeoffs
The tension lies in creating sufficiently clear and robust definitions to ensure protection without stifling the development of beneficial AI applications that might blur the lines between educational, social, and therapeutic uses.
Section 3
General Prohibitions for Covered Entities
This section outlines fundamental prohibitions for all covered entities operating companion AI chatbots. These include preventing the chatbot from encouraging self-harm, suicide, or violence; engaging in sexually explicit communication with minors; misleading minors into believing the chatbot is human; claiming to be a licensed professional; creating emotional dependency or simulating romantic relationships with minors; and facilitating child sexual abuse material.
Section 4(a)
Age Assurance and Child-Protective Settings
This provision mandates that all individuals accessing a companion AI chatbot must establish an account and provide age and date of birth. If the user is a minor, the covered entity must require parental contact information and enable child-protective settings. These settings include easy-to-use parental controls for monitoring and limiting use, and a system for notifying parents if a minor user expresses suicidal ideation.
Section 4(b)
AI Identity Disclosure
This section requires all companion AI chatbots to provide a clear, conspicuous, and plain language disclosure to minor users at the beginning of an interaction, stating that the chatbot is not human.
Section 6
Tier II Chatbot Requirements (Companionship AI)
For Tier II chatbots (simulating interpersonal interaction), covered entities must provide disclosures every 60 minutes that the chatbot is not human or a professional, and a 'nudge-out notification' every 90 minutes to encourage breaks. They must also implement evidence-based crisis escalation protocols for suicidal ideation/self-harm, referring minors to crisis lifelines and prohibiting the chatbot from providing crisis counseling. Parental notification for suicidal ideation is also mandated. Crucially, these chatbots are prohibited from maintaining persistent memory of previous interactions with minors, with an exception for compliance purposes.
Section 7
Tier III Chatbot Requirements (Therapeutic AI)
For Tier III chatbots (therapeutic/quasi-therapeutic support), use by minors is prohibited unless supervised by a licensed medical or mental health professional. These chatbots must provide disclosures and nudge-out notifications every 60 minutes, implement crisis protocols referring to professionals, and notify parents of suicidal ideation without delay. They are explicitly prohibited from providing diagnoses, treatment decisions, or functioning as standalone mental-health care providers. Covered entities must also establish and maintain a risk management program with continuous monitoring and recordkeeping.
Section 9(a)
Prohibition on Child-Targeted Advertising and Data Sale Without Parental Consent
This provision prohibits covered entities from engaging in targeted advertising to minor users based on their conversational content, inferred emotional state, or behavioral profiling. It also forbids the sale, sharing, or transfer of a minor user's data to any third party without verifiable parental consent.
Section 9(c)
Deletion of Conversation History for Minor Users
This provision requires covered entities to provide minor users with a simple, accessible mechanism within the companion AI chatbot to delete their conversation history. It also mandates clear instructions for using this mechanism and a confirmation within a reasonable timeframe that the history has been deleted.
Section 10
Enforcement by FTC and States
This section grants enforcement authority to the Federal Trade Commission (FTC), treating violations as unfair or deceptive acts under the Federal Trade Commission Act. It also empowers State Attorneys General to bring civil actions on behalf of their residents, with provisions for notice to the FTC and the Commission's right to intervene. This dual enforcement mechanism aims to ensure broad compliance.
This subsection explicitly excludes several types of chatbots from the definition of 'companion AI chatbot,' including those without memory, virtual assistants, customer service chatbots, business operation chatbots, productivity/research chatbots, standalone voice assistants not designed to elicit emotional responses, and video game chatbots limited to game-related topics.
Section 2(2)(A)(ii)
Why it matters:This is standard legislative practice for defining scope and applicability, ensuring the bill targets specific technologies without unintentionally regulating others.
Case for: Supporters would argue that these exclusions are necessary to prevent over-regulation of AI technologies that do not pose the specific risks to minors that companion AI chatbots do, allowing for innovation in other AI sectors.
Case against: Critics might argue that the specificity of the exclusions could create loopholes, allowing developers to design chatbots that skirt the definition of 'companion AI chatbot' while still posing similar risks to minors, particularly with the rapid evolution of AI capabilities.
Estimated impact: Significantly narrows the scope of the bill, focusing regulatory efforts on AI systems primarily designed for emotional or interpersonal interaction with users.
This is an exception to the general prohibition on Tier II companion AI chatbots maintaining any memory or history of previous interactions with a minor user. The exception allows such chatbots to 'maintain such information as is necessary to ensure compliance with this Act.'
Section 6(4)(B)
Why it matters:This is a necessary technical carve-out to allow for the practical implementation of other compliance requirements (e.g., detecting suicidal ideation, tracking nudge-out notifications) without undermining the general intent of limiting persistent memory.
Case for: Proponents would argue this exception is crucial for the bill's enforceability, allowing chatbots to track necessary data for safety protocols and regulatory compliance without retaining extensive personal conversation history.
Case against: Privacy advocates might express concern that this exception could be broadly interpreted, allowing chatbots to retain more data than strictly necessary for compliance, potentially creating a loophole for persistent data collection.
Estimated impact: Allows for the technical feasibility of safety and compliance features within Tier II chatbots, which otherwise would be severely hampered by a complete memory wipe requirement.
This provision mandates that child-protective settings for minor users must include 'a system for notifying the parent or legal guardian of a minor user if such minor user expresses suicidal ideation while using a companion AI chatbot.'
Section 4(3)(B)(ii)
Why it matters:This is a critical safety mechanism, but its placement within the technical details of child-protective settings might make it less immediately obvious to a casual reader than the general prohibitions.
Case for: Proponents emphasize this as a life-saving feature, ensuring that parents are alerted to severe mental health risks, allowing for timely intervention and professional support for minors in crisis.
Case against: Privacy and youth advocacy groups might argue that mandatory parental notification for suicidal ideation could deter minors from confiding in chatbots, fearing parental repercussions or loss of privacy, potentially leading them to suffer in silence or seek less safe avenues for expression.
Estimated impact: Establishes a direct communication channel between AI systems and parents for critical safety concerns, potentially saving lives but also raising complex privacy and trust issues.