Governor Brian Kemp signed SB 540 into law this month, making Georgia the latest state to require that AI-powered chatbots disclose their non-human nature to users. The law is deceptively simple on its face. But in the context of where AI is heading — and how much consumers are already relying on chatbots for legal advice, medical guidance, customer service disputes, and financial decisions — it matters more than the headline suggests.

What the Law Requires

SB 540 mandates that any chatbot interacting with Georgia consumers must clearly identify itself as an artificial intelligence system when asked — or proactively disclose it at the start of meaningful interactions. The disclosure must be in plain language. Burying “you may be interacting with an AI” in a terms of service page doesn’t satisfy the requirement.

The law applies broadly: customer service chatbots, AI assistants embedded in apps and websites, and automated phone systems that simulate human conversation are all covered. There are carveouts for obvious contexts — a clearly labeled “AI chat” widget on a tech company’s website doesn’t need to announce itself every message. But when there’s meaningful ambiguity, disclosure is required.

Why “Simple” Is Actually Hard

The instinct is to dismiss chatbot disclosure laws as performative — of course companies should tell you when you’re talking to a bot. But the reality of modern AI deployment is that the line is genuinely blurry.

Large language models are now good enough to pass as human in many contexts. They use first-person language, express empathy, and adapt their tone. In customer service disputes, users are already making decisions — about whether to escalate, whether to accept a settlement offer, whether to provide sensitive personal information — based partly on whether they believe they’re talking to a person who has discretion and judgment.

The deception isn’t always intentional. Companies may deploy AI tools without fully disclosing it to the user-facing team, let alone consumers. The chatbot just appears as “Support” or “Alex” or “Jordan.” SB 540 creates an obligation to be explicit.

Georgia Joins a Growing Roster

Georgia is now part of a wave of states legislating AI disclosure. California’s BOT Disclosure Act has been on the books since 2019 but was narrow in scope. More recent state laws in Illinois, Texas, and others have expanded disclosure requirements to cover a wider range of AI-driven interactions — not just political bots or social media manipulation, but commercial AI across industries.

What’s changed is the enforcement context. In 2019, regulators had limited visibility into AI deployments and limited appetite to pursue disclosure violations. In 2026, state attorneys general have both the tools and the political incentive to go after companies that deploy AI deceptively — especially in high-stakes sectors like healthcare, insurance, and legal services.

The Consumer Privacy Angle

Chatbot disclosure isn’t just about knowing whether you’re talking to a human. It’s about consent to data collection and processing.

When you interact with a human customer service representative, your conversation may be recorded and used for training purposes — but the nature of what’s captured is relatively understood. When you interact with an AI, your conversation is training data in a more direct sense. Every message you send is potentially a data point that shapes future model behavior. The privacy calculus is different.

Consumers who don’t know they’re talking to an AI can’t make informed decisions about what to share. Someone disclosing a sensitive medical situation to what they believe is a human customer service agent is making a very different privacy decision than someone who knows they’re feeding information into a language model.

Georgia’s law creates the minimal condition for informed consent: you have to know what you’re interacting with.

What Businesses Need to Do

If you operate a chatbot that serves Georgia consumers, the checklist is straightforward but requires audit:

Review every consumer-facing AI touchpoint. This includes embedded chat widgets, phone trees that use speech synthesis and NLP, mobile app assistants, and any automated messaging that a user might reasonably interpret as human-generated.

Update disclosure language. “You’re chatting with an automated assistant” is better than nothing, but review whether it meets plain-language requirements under SB 540.

Train customer service staff. Human agents who hand off to or collaborate with AI tools need to understand the disclosure requirements and not inadvertently represent AI responses as their own.

Document your AI stack. Enforcement actions tend to start with regulators asking “what AI systems do you use?” Companies that can’t answer that question clearly are already in trouble.

The Bigger Picture

SB 540 is unlikely to be the story of the year. But it’s representative of a regulatory pattern that’s accelerating: states are moving faster than Congress on AI transparency, and they’re being specific. Not “AI should be trustworthy” as a principle, but “chatbots must disclose their nature” as an enforceable rule.

The Federal government has struggled to pass comprehensive AI legislation. While the SECURE Data Act and various AI governance proposals remain stalled or underspecified, states are filling the vacuum — one disclosure requirement, one bias audit mandate, one data retention limit at a time.

For consumers, this is genuinely protective. For businesses, the compliance map is getting more complex by the month. Georgia is now on it.