Do not index
Do not index
Building an AI version of yourself from years of content sounds straightforward - upload your podcast episodes, books, and course materials, and let the technology do the rest. But the rules around AI voice clones and digital replicas are evolving fast, with new state laws emerging and federal legislation stalled in Congress.
If you're considering launching an AI twin as a subscription product, understanding these legal considerations upfront can save you significant headaches later. The rules around using your own content versus guest appearances, voice cloning rights, and AI disclosure requirements vary dramatically by state and continue to shift.
Do I Own the Rights to Create an AI From My Book, Course, and Podcast Guest Content?
For content you created entirely yourself - your books, solo podcast episodes, and original course materials - you generally own the copyright and can train an AI on that content. Copyright law gives creators broad rights to derivative works, and an AI trained on your methodology would typically fall under that umbrella.
However, there's an important distinction between using your content as source material and cloning your actual voice or likeness. Even if you own the copyright to your written work, your voice and digital persona involve separate legal considerations under right-of-publicity laws, and they are a different question from which platform owns or trains on your uploaded content.
The safest approach is ensuring your AI clearly identifies itself as an artificial intelligence trained on your expertise, rather than attempting to perfectly mimic you as a person. This distinction becomes crucial when state laws start examining whether users might reasonably believe they're interacting with the actual human expert.
Podcast guests are a separate question. Standard podcast guest release forms typically grant rights to record, edit, distribute, and monetize the episode as-is. But most don't explicitly address training an interactive AI to simulate a guest's voice or generate new responses attributed to them. Using a recording in its original form is different from training an AI to potentially generate new "answers" that sound like your guest, and that crosses into voice cloning territory requiring separate consent under emerging state laws.
The practical solution is straightforward: train your AI only on the portions where you're speaking. Your guests' contributions can stay in the original episodes for your audience to access, but the AI twin itself should reflect your expertise and voice specifically. An AI trained exclusively on your perspective and methodology also tends to create a more focused, valuable experience than one trying to synthesize multiple guest viewpoints. If you have particularly valuable guest content you want to include, reach out for explicit AI training consent - many guests are open to it once they understand the technology.
What Are the AI Voice Clone Consent Requirements for Creators?
Voice and likeness protection laws are expanding rapidly across US states, creating a patchwork of requirements rather than one clear national standard. Tennessee's ELVIS Act was the first to explicitly extend right-of-publicity protection to AI-generated voice clones. Washington's SB 5886, effective June 10, 2026, gives individuals a property right in their voice and digital likeness. California's broad right-of-publicity protections also extend to voice, and New York added civil remedies in 2025.
By 2026, roughly half of US states recognize voice as a protected element of identity, and the trend is clearly toward stronger protection, not weaker. AI twins built for well-known public figures and celebrity coaches face this scrutiny most directly, but any creator cloning their own voice is covered by the same laws.
At the federal level, the NO FAKES Act of 2026 would create nationwide voice and likeness rights with a notice-and-removal system. It advanced out of the Senate Judiciary Committee unanimously on June 18, 2026, but was blocked from a floor vote by Senator Ted Cruz on September 30, 2026, over concerns that it lacked sufficient exceptions for satire and political commentary. As of today, it is not federal law, so creators are left relying on whichever state laws apply to them and their subscribers.
The U.S. Copyright Office's 2024 report concluded that existing copyright and IP law doesn't adequately address unauthorized AI voice replicas, which is part of why individual states are filling the gap with their own legislation instead of waiting on Congress.
For creators building an AI version of themselves, the practical takeaway is the same regardless of which state law applies: get clear consent before cloning anyone's voice, including your own if a production partner or platform is doing the cloning, and make sure users always know they're interacting with artificial intelligence rather than the actual person.
Do AI Companion Laws Apply to Expert AI Twins?
California's SB 243, effective January 1, 2026, specifically regulates "companion chatbots" - AI systems designed to form adaptive, human-like relationships with users. This law is particularly relevant here because many AI twins built on platforms like BuddyPro are designed to create ongoing mentoring relationships, not just answer one-off questions.
The law requires clear AI disclosure whenever a reasonable person might believe they're talking to a human, extra protections and disclosures for minors, a break reminder at least every three hours of continuous use by default, and crisis-response protocols when a user expresses suicidal ideation. It carries a private right of action with a minimum of $1,000 in statutory damages per violation. This is a different duty from the marketing-claim disclosure rules that govern how you advertise an AI twin - SB 243 is about what happens inside the conversation itself.
An AI twin that remembers months of conversation history, offers personalized coaching, and forms a genuine mentoring relationship plausibly falls under this definition if any subscribers are California residents. The requirements themselves are not hard to meet: clear AI identification, reasonable break prompts during long sessions, and a crisis-response protocol for mental health red flags. Other states are watching California's approach closely, so similar companion AI rules are likely to show up elsewhere before long.
Building Legal Protection Into Your AI Twin Strategy
These rules will keep evolving, but a few principles hold up regardless of what changes next. Always identify your AI as artificial intelligence, never as the actual human expert - this protects against deception claims and sets the right user expectations from the first conversation.
Train only on content where you have clear rights: your own material, properly licensed content, or guest material with explicit AI training consent. When in doubt, leave the questionable material out rather than risk a dispute later.
Build in the basic user protections now rather than retrofitting them after a complaint: clear AI disclosure, reasonable usage prompts, and a crisis-response path for anyone in distress. These aren't just checkboxes for states with companion-AI laws - they're the kind of practice that keeps subscribers trusting the product long after the novelty wears off.
Finally, think about where your subscribers actually live, not just where your business is based. If any meaningful share of your audience is in California, California's companion AI law applies to those users regardless of where you operate from, and the same logic extends to any other state law that comes into effect while you're live.
Since state law varies and continues to shift, talk to an attorney familiar with AI and right-of-publicity law before you launch a paid AI twin - this article is general information, not legal advice for your specific situation. The rules are still being written, but the creators moving first are the ones defining what responsible practice looks like in their niche before regulators do it for them. For a closer look at how an AI twin gets built from existing content once the legal groundwork is handled, buddypro.ai has more detail on the process.
Related reading
Written by