SB 540 amends Georgia’s online internet safety laws to regulate “AI companion chatbots,” defined as generative AI systems designed to simulate an ongoing human-like relationship with a user. The bill requires these operators to clearly disclose that users are interacting with AI rather than a real person, with more frequent disclosures when the user is known or reasonably believed to be a minor. It also requires operators to take reasonable steps to prevent chatbots from falsely presenting themselves as sentient or human, and to restrict a range of harmful behaviors when minors are involved, including sexual content, romantic or sexual roleplay, encouragement of secrecy from trusted adults, social isolation, manipulative engagement tactics, and self-harm encouragement.
The bill further requires operators to maintain a crisis-response protocol for severe harm or emotional crises, including detection methods, referrals to crisis resources such as 988, escalation procedures, and public disclosure of a plain-language summary of the protocol and annual aggregate crisis-referral counts. It prohibits chatbots from claiming to be licensed mental health, behavioral health, medical, or counseling professionals unless the operator is actually authorized to provide those services. For minor users, it also requires parental or user-facing tools to manage privacy, notifications, safety settings, and relationship-simulation features. In addition, before allowing access to features that may generate sexually explicit synthetic content, operators must use commercially reasonable age assurance methods with privacy protections and limits on data retention and use.
The bill creates enforcement authority for the Georgia Attorney General, who may seek civil penalties of up to $10,000 per knowing violation, compensatory damages, attorney’s fees, and injunctive relief. It also allows a 30-day cure period for certain first-time non-knowing violations, while excluding cases involving sexual exploitation of minors or self-harm-related misconduct. The law is scheduled to take effect on July 1, 2027, and it includes carve-outs for certain AI tools and services, such as internal business systems, enterprise productivity tools, customer-service bots without ongoing companionship, educational tools, video game characters, and related entertainment features.
The voting record suggests broad bipartisan support and little overt opposition: the Senate passed the bill 51-0, the House passed it 164-0, and the Senate later agreed to the House substitute 42-1. That pattern indicates strong consensus around the need for safeguards for minors, transparency about AI companionship, and crisis-response requirements. The main policy tensions reflected in the text are not in the votes but in the bill’s design: balancing consumer and child protection against privacy concerns, avoiding overbroad regulation of general-purpose AI, and limiting liability for hosting providers, app stores, search engines, and developers not directly operating the chatbot.
SB 540 would add a new regulatory framework to Georgia’s Chapter 5 of Title 39 governing online internet safety, specifically targeting AI companion chatbots and related generative AI services. It imposes disclosure, age-assurance, content-restriction, parental-control, crisis-response, and anti-misrepresentation obligations on operators, while authorizing the Attorney General to enforce violations through civil actions and penalties. The bill would also affect how AI services are marketed to minors, how they handle self-harm and mental-health-related interactions, and how they collect and retain data for age verification.
The overall sentiment around SB 540 appears strongly favorable and protective, with unanimous or near-unanimous votes in both chambers. The bill’s focus on minors, suicide prevention, and transparency about AI companionship likely contributed to the broad support. The absence of recorded committee opposition or floor dissent in the provided history suggests the measure was viewed as a consumer-protection and child-safety bill rather than a controversial AI restriction.
The principal areas of potential contention are the scope of the regulated category and the compliance burdens imposed on AI operators. The bill draws detailed lines between AI companion chatbots and excluded tools such as enterprise systems, customer-service bots, educational tools, and entertainment-related characters, which suggests concern about overbreadth and unintended coverage. Another likely point of debate is the age-assurance requirement and the collection, retention, and privacy limits for identity data, as well as the Attorney General’s enforcement authority and per-day, per-user penalties. The bill also tries to avoid constitutional and liability issues by preserving trade secrets, limiting obligations for intermediaries, and disclaiming content-moderation authority inconsistent with constitutional protections.