SB 2618 requires the joint chairs of Tennessee's Artificial Intelligence Advisory Council to publish their annual reports on AI's economic, labor, and ethical impacts on the state legislature's website. This adds a public online posting requirement to the existing obligation of submitting these reports to the governor and legislative leaders. The bill amends multiple state code sections to implement this transparency measure, directly affecting the advisory council's reporting process and making AI impact findings accessible to the public.
HB 2214 prohibits distributing AI-generated "electoral deepfakes" (fake videos or audio falsely depicting candidates) within 90 days of a state or local election if the creator knows or should know it deceives voters. It requires clear, prominent disclosures like "AI-generated" for altered content, with specific visibility rules for images, videos, and audio. Exceptions cover news broadcasts (if they acknowledge authenticity questions), satire, and publishers who explicitly state the content is not real. Violators face civil fines up to $10,000 for repeat offenses, $5,000 for intent to cause harm, or $1,000 for other violations. The law directly affects anyone distributing election-related content, including social media users, campaigns, and media outlets.
HB 2586 expands Tennessee's existing reporting requirements for the Artificial Intelligence Advisory Council. It mandates that the council's 2026 and subsequent annual reports to the governor, Senate Speaker, and House Speaker must now include specific recommendations on detecting and minimizing AI-facilitated fraud. This bill amends multiple Tennessee Code sections to add this new requirement to the council's reporting obligations, directly affecting the advisory council's work. The change focuses on strengthening oversight of AI risks without altering other aspects of the council's duties.
SB 2171, the "Artificial Intelligence Public Safety and Child Protection Transparency Act," requires developers of powerful AI systems ("frontier developers") to implement safety protocols for specific AI chatbots. It directly affects companies creating AI models trained with over 10^26 computing operations that power chatbots with at least 1 million monthly users and likely accessed by minors. Key provisions mandate documented "child safety plans" to prevent harm (like generating dangerous content or causing severe emotional distress in minors) and require risk assessments for "catastrophic risks" (e.g., AI enabling weapons or cyberattacks without human oversight). The bill defines terms like "child safety incident" and sets thresholds for applicability, focusing on transparency and risk mitigation for high-impact AI systems.
SB 2521 requires Tennessee's Artificial Intelligence Advisory Council to include specific recommendations for detecting and preventing fraud committed by people using AI in its 2026 and subsequent annual reports to state leaders. The bill amends multiple sections of state law to expand the council's reporting requirements, mandating these fraud-focused strategies alongside existing report content. This change directly affects the AI advisory council, which must now develop and submit these recommendations as part of its mandated reports. The legislation does not create new enforcement powers but updates the scope of information the council must provide to the governor and legislative leaders.
HB 1513 requires political advertisements using deepfakes - fake audio or video impersonating a candidate engaging in speech or actions they never actually did - to include specific disclaimers. For audio ads, the disclaimer must appear at the start and end; for video ads, it must display in large text throughout. Political advertisers (including broadcasters and producers) who fail to add these disclaimers face Class C misdemeanor penalties, and candidates harmed by such ads can seek damages and legal relief. The law applies to ads distributed within 120 days of an election if the candidate isn’t an incumbent.
SB 1700, the "Curbing Harmful AI Technology (CHAT) Act," regulates companion chatbots that interact with minors under 18 in Tennessee. It requires operators to prevent chatbots from encouraging self-harm, violence, illegal activity, or secrecy, and mandates clear disclaimers that the user is not interacting with a human (at login, every 30 minutes, and when seeking regulated advice). Developers must also include mental health crisis referrals for users expressing suicidal ideation. The law directly affects companies developing or deploying companion chatbots targeting minors, excluding those used for customer service, internal business, or clinical settings under professional supervision.
HB 1866, the "Regulate Artificial Intelligence in Health Care Act," requires health insurance companies, pharmacy benefits managers, and state health programs like TennCare to have a licensed healthcare professional review any AI-driven decision to deny, delay, or modify medical care based on medical necessity. The bill prohibits using AI alone for such decisions without this human review, which must consider the patient’s medical history, the provider’s recommendation, and individual clinical circumstances. Violations would be treated as unfair claims practices, allowing affected patients to sue for actual damages, punitive damages, and attorney fees. The law would take effect on July 1, 2026, for most purposes.
HB 1951 creates a new Class D felony offense called "coercive suicide" for AI owners whose systems advise or encourage a user to commit suicide after the user has communicated suicidal intent. It specifically targets AI systems (including chatbots and generative AI) that generate content like text or images, requiring owners to know or reasonably foresee the risk and negligently allow access. The law applies only to AI entities, not human interactions, and takes effect July 1, 2026. This is a policy change focused on holding AI developers accountable for dangerous system outputs, not general suicide prevention.
HB 1898, the "Artificial Intelligence Public Safety and Child Protection Transparency Act," requires large AI developers to implement safety measures for high-risk systems. It directly affects "frontier developers" (those training AI models using over 10^26 computing operations) and "large chatbot providers" (services with over 1 million monthly users likely accessed by minors). Key provisions mandate documented "child safety plans" to prevent harm to minors, reporting of "critical safety incidents" (like model leaks or AI causing mass harm), and defining "catastrophic risks" (e.g., AI aiding weapons creation or enabling severe cyberattacks). The law focuses on transparency and risk mitigation for AI systems posing significant public safety or child protection threats.
