Ohio is moving to clamp down on some of the ugliest corners of artificial intelligence, as lawmakers weigh S.B. 163, a three-part bill that targets AI-generated child sexual abuse material and deepfake voice-and-likeness scams. The proposal would make it a crime to create or share AI child sexual abuse imagery, tighten rules on identity fraud, and force certain AI products to carry permanent, machine-readable watermarks. Supporters say predators are exploiting gaps in current law; tech industry groups say the bill, as written, could sweep too broadly.
What S.B. 163 Would Do
S.B. 163, sponsored by Sens. Louis Blessing III and Terry Johnson, would revise Ohio’s obscenity and fraud statutes to explicitly cover “artificially generated depictions” of minors and to regulate how AI output is labeled, according to the Ohio Legislature. The bill would require a permanent watermark on AI-generated products and make it unlawful to remove that watermark if the goal is to hide where the content came from.
The proposal would give the attorney general authority to seek injunctions and civil penalties, while preserving a private right of action for people harmed when watermarks are stripped away. The introduced language treats both production and dissemination of artificially generated depictions of minors as criminal offenses, setting out a range of felony penalties and civil remedies keyed to the different violations.
Support From Prosecutors and Advocates
Attorney General Dave Yost has thrown his weight behind the measure, calling the surge in AI-generated child sexual abuse material a legal “gray area” that legislators need to clear up, according to a statement from the attorney general's office. The Ohio Prosecuting Attorneys Association has listed S.B. 163 among the bills it supports, and both the Center for Christian Virtue and the Ohio Domestic Violence Network have submitted supporting testimony in committee.
Backers argue that combining criminal penalties with civil remedies and takedown authority will finally give prosecutors the tools to get the worst content off the internet quickly, instead of watching cases fall through the cracks while courts debate how existing law applies to AI.
Tech Industry Pushback
On the other side, TechNet, which represents a range of major technology companies, has filed formal comments opposing S.B. 163 in its current form. The group says its members “support the goal” of protecting children but warns that sweeping watermark mandates and broad private rights of action could create legal uncertainty and chill innovation.
In written testimony, TechNet urged lawmakers to put the legal focus squarely on “bad actors” who misuse AI tools, not on the companies that build those tools and cannot realistically monitor every post on their platforms. The group recommends relying on evolving provenance metadata rather than a rigid, across-the-board watermark requirement.
Why Now
The legislative push accelerated after Gov. Mike DeWine used his March 10 State of the State address to call for AI “seat belts,” urging lawmakers to ban AI-generated child sexual abuse material and to hold companies accountable when their products encourage self-harm, as reported by the Statehouse News Bureau via WOUB. Related measures are advancing in the House, including H.B. 524, which would impose penalties on AI platforms that recommend self-harm content, reflecting a broader anxiety about generative AI and children’s safety.
Activity in Columbus fits into a nationwide wave of AI policymaking. Legislatures around the country are exploring disclosure rules, provenance requirements, and new criminal penalties for specific AI uses, a trend tracked by NCSL.
What Comes Next
S.B. 163 has been parked in the Senate Judiciary Committee since April 2025, logging multiple hearings but no final vote. That lag time has opened space for behind-the-scenes negotiations on technical fixes and carveouts for journalism, research, and other forms of protected expression.
TechNet and other stakeholders have already submitted proposed edits, and the bill’s sponsors say they are open to amendments as long as the final product stays targeted and constitutional. If the committee signs off on major changes, the bill could head to the full Senate. Longer term, judges or Congress may end up deciding how far states can go in policing AI.
Whether Ohio ultimately lands on a strict watermark mandate, a more flexible provenance system, or a narrower criminal statute will matter far beyond Columbus. Other legislatures, along with the companies building generative AI tools, are watching closely to see where S.B. 163 comes down. For now, the bill has pushed AI and child safety to the center of this session’s policy fights.
