24-04-2025
Ohio senators want to crack down on child pornography created with AI
Ohio lawmakers want to regulate images and sounds made by artificial intelligence, including making simulated child pornography made with AI illegal.
This is the second time state Sens. Bill Blessing, R-Colerain Twp., and Terry Johnson, R-McDermott, have introduced this legislation. Along with Senate Bill 163's rules against child pornography, it would require images and sounds made with AI to include a watermark and it would also prohibit identity fraud with AI.
More: Ohio teen tapped to help draft legislation on health care and AI
Blessing said current statues about child pornography leave loopholes for people to create AI-generated child porn as long as it is not explicitly based on a real person. Current law requires a real photo of a child to prosecute someone for generating or possessing child sexual assault material. Because AI does not generate a "real image," there is an exploitable gap that AI fits into.
Blessing said Ohio Attorney General Dave Yost reached out to the General Assembly to fill this gap in Ohio law.
"If you're peddling in child pornography, even if it's artificially intelligence generated it's effectively the same thing," he said.
Under SB 163, creating or distributing simulated obscene material, including child pornography, would become a third-degree felony and buying or possessing it would be considered a fourth-degree felony.
The bill would also prohibit users from generating sexual images of adults without that person's consent.
A provision in the bill would require AI programs to include a distinctive watermark that informs the user that what they are viewing or hearing was made by AI.
Blessing said the watermark would not be intrusive and easy for companies to implement. Instead of something visible in an AI-generated image or something a listener could hear in AI-generated audio, this watermark would not be detectable by humans but would instead be detected by another machine.
It would exist in the file's metadata and be used to determine the origins of AI images and audio. SB 163 would allow citizens and the attorney general to sue AI generators for not including these watermarks and anyone who intentionally removes them for damages.
A group of six companies and interest groups including Technet, a trade association for technology companies, signed onto opponent testimony from the previous version of the bill last year. They acknowledged the threat AI poses, but said companies are doing enough to regulate nefarious uses, including child sexual assault material.
They also said the requirement for a watermark might not be feasible with current technology. The regulations proposed by the bill, the interest groups argued, could stifle innovation.
If the bill becomes law, citizens could sue people who use AI to replicate someone's persona to, harm their reputation or defraud them. They would also be able to take action if someone used their persona to convince someone to make a financial decision that would require that person's approval.
"There's going to be a lot of legislation in this space going forward, and rightfully so, because this is not something that should be left up to the markets as my colleagues like to say." Blessing said. "I'd rather have a regulated environment in this space, rather than just leaving it up to the courts or, you know, God knows what."
Donovan Hunt is a fellow in the E.W. Scripps School of Journalism's Statehouse News Bureau.
This article originally appeared on The Columbus Dispatch: State lawmakers introduce bill to prohibit AI identity fraud
