H.829 requires businesses or organizations making robocalls with artificial voices (generated or altered using AI) to first play a natural, unrecorded voice announcement before delivering any prerecorded message. This announcement must state the call’s purpose, the caller’s name/address/phone number, ask if the recipient consents to hear the message, and disclose that an artificial voice will be used. The law applies to all robocallers in Vermont using AI-generated voices and imposes civil penalties of $500 for a first violation (rising to $1,000 for repeat offenses), plus potential criminal penalties of up to 90 days in jail or $1,000 per call for willful violations. It directly affects telemarketers, debt collectors, and other entities using automated calls with artificial voices within Vermont.
This Vermont bill (S.207) prohibits businesses from using "surveillance pricing" to set different prices for consumer products based on data collected through electronic surveillance technology (like AI or tracking tools). It directly affects retailers, online sellers, and service providers in Vermont that sell consumer products (e.g., groceries, electronics, subscriptions). The law bans price discrimination using personally identifiable or aggregated consumer data, except for price differences based solely on actual costs, equal discounts for all customers (with clear notice and consent if personal data is used), or licensed insurance companies following existing rate rules. The bill takes effect July 1, 2026.
H 752 requires Vermont's Agency of Digital Services to annually review all state government computer systems that make decisions (like benefits or permits) for cybersecurity risks and personal data vulnerabilities. The bill amends state law (sections 3303 and 3305) to mandate that the Agency's annual inventory of these systems include assessments of privacy risks and cybersecurity safeguards. If a system is found to pose significant risks, the Agency can formally request that state agencies terminate its use. This directly affects all Vermont state departments using automated decision systems that process personal data or make government decisions.
This bill requires chatbot providers operating in Vermont to protect users' data privacy and security, clearly disclose when users are interacting with an AI (not a human), and allow users to access their own chat logs. It defines key terms like "chatbot" (AI systems simulating conversation) and "affirmative consent" (clear permission, not hidden terms of use), and prohibits deceptive "dark patterns." The law applies to all chatbot providers within Vermont, directly affecting users who interact with these services. Key provisions mandate transparent data handling, user access to their input/output data, and strict rules against misleading consent practices.
This Vermont bill regulates developers and deployers of automated decision systems (ADS) used in consequential decisions - such as hiring, housing, loan approvals, or healthcare access - to prevent algorithmic discrimination. It requires businesses to clearly inform consumers before using ADS in such decisions, explaining what data is measured, how it is used, and how it influences outcomes. The bill prohibits the use of ADS that result in discriminatory treatment based on protected characteristics like race, gender, age, or disability. These requirements apply to businesses operating in Vermont that deploy or develop ADS for decisions materially affecting residents' rights, liberties, or welfare.
H 365 requires social media platforms and artificial intelligence (AI) providers operating in Vermont to annually register with the Secretary of State by January 31, pay a $100 fee, and submit privacy policies. The bill directly affects companies using platforms or AI systems with Vermont residents, mandating they avoid harmful design features for minors (like excessive use prompts), set high default privacy settings, block unknown adult contact without minor-initiated interaction, and prohibit using minors' data for targeted advertising. Providers must also comply with "product safety and privacy terms" focused on preventing algorithmic discrimination and "heightened risk of harm" to minors, including unfair data processing. Violations can trigger daily $50 fines (capped at $10,000 annually) and civil penalties enforced by the Attorney General.
This bill sets rules for when companies can use digital replicas (like AI-generated versions of a person) in service contracts. It requires clear consent from individuals and specific terms about how these digital replicas can be used. The law directly affects people whose digital likeness might be replicated and the businesses creating or using such replicas.
This bill establishes safety standards and oversight for developers and deployers of inherently dangerous artificial intelligence systems in Vermont. It requires these entities to submit safety and impact assessments to a new Division of Artificial Intelligence within the Agency of Digital Services. The Attorney General can enforce the law by taking legal action against noncompliant developers or deployers, including seeking court orders to stop violations. The bill specifically targets high-risk AI systems, dual-use foundational models, and generative AI systems that pose serious risks to public safety, economic security, or individual rights.
S 23 prohibits the distribution of deceptive synthetic media (AI-generated content falsely representing political candidates) within 90 days of Vermont elections without clear disclosure. It requires visible or audible disclaimers stating the media was "created or manipulated by digital technology," with specific formatting rules for different formats. The law exempts news media during bona fide reporting, satire, and platforms under Section 230 of federal law. Violations carry fines up to $15,000 (with higher penalties for repeat offenses or intent to cause harm), and candidates can seek court injunctions to stop distribution.
