On May 11, 2026, Georgia Governor Brian Kemp signed SB 540, the state's AI Chatbot Disclosure and Child Safety Act, into law. The bill — sponsored by Senate President Pro Tem Larry Walker (R-District 20) — requires AI chatbot operators to verify user age, provide parental controls, disclose to users that they are interacting with AI rather than a human, and implement specific safety protocols for conversations involving suicidal ideation, self-harm, and sexually explicit content. The law carries civil penalties of up to $10,000 per knowing violation, enforced by the Georgia Attorney General's office. SB 540 takes effect July 1, 2027.

Why It Matters

The no-carve-out structure is the most important detail. Most state AI-chatbot laws to date have been narrow strikes at specifically problematic standalone apps (Replika, Character.AI's pre-2025 content) while leaving the chatbot features inside Meta, Google, Microsoft, X, and Snap effectively unregulated. Georgia's bill is the first that would apply meaningfully to those embedded features at scale. The July 2027 effective date gives operators 14 months of runway, but for the AI-companion subset that already specializes in romantic or sexual interaction, this is another tile in a fast-multiplying compliance mosaic — Oregon, Washington, Georgia, the pending Colorado HB 1263, and the federal GUARD Act all impose overlapping but slightly different requirements that operators will need to harmonize or geo-fence around.

What makes Georgia's law nationally distinctive is the absence of a carve-out for chatbots embedded inside larger platforms. Oregon's SB 1546 (March 5), Washington's HB 2225 (March 12), and California's SB 243 each exempt or treat differently chatbots that are sub-features of broader services — a structural concession that effectively narrows scope to standalone AI-companion apps like Replika, Character.AI, Candy.AI, and DreamGF. SB 540 has no such carve-out: Meta's AI chatbot inside Instagram, Google's Gemini inside Workspace, Snapchat's My AI, and any future LLM-feature inside a social platform would be subject to the same disclosure, age verification, and content-safety requirements as the standalone companion apps. Compliance attorneys at Crowell & Moring and the California Lawyers Association have flagged this as the model that, if adopted by larger states, could meaningfully reshape how big-tech firms integrate generative AI into consumer products.

SB 540 cleared the Georgia Senate on March 6, the House on March 25, and was reconciled on March 27 — Kemp held the signed bill for roughly six weeks before formalizing on May 11, an unusually long pre-enactment runway likely reflecting the volume of industry-comment correspondence the governor's office received about the no-carve-out provision. Walker's stated motivation aligned with the broader AI-companion-regulation wave: "We wanted to make sure that we have safety around topics with AI regarding suicidal ideation, self-harm and any kind of sexually explicit content."

Georgia joins Oregon, Washington, and (pending federal advancement) Connecticut and Pennsylvania as the most active state-level enforcers of AI-companion safety regulation. The federal GUARD Act (S. 3062, Hawley/Blumenthal) advanced unanimously out of Senate Judiciary on April 30 and is awaiting a floor vote — a federal preemption fight may be looming if state laws and the federal framework diverge on definition of "covered platforms" or content moderation specifics.

Sources

Update — 2026-05-17

Initial entry — story first created.