(David Espejo/stock photo via Getty Images)
Ohio lawmakers continue their efforts to regulate artificial intelligence, seeking to crack down on explicit and consent-based content.
One hint is enough to turn a staunch Republican into a staunch Democrat.
Artificial intelligence can create and edit photos and videos. Fake photos or videos of people, called deepfakes, are becoming more common on the Internet.
“You should be able to tell someone, ‘Please take this away, a photo like this has never happened,'” Mathews said.
Having been a victim of an AI scam himself and allowing me to counterfeit it for a story earlier, Mathews explained that lawmakers and celebrities aren’t the only targets of counterfeit images.
“Everyone should have their own dignity and a way to present themselves to the world,” he said.
Law enforcement officials in California and Arizona have launched an investigation into X’s Grok AI tool for explicit content obtained without their consent. Users could ask Grok to strip people, including children, and place them in compromising outfits or positions.
Mathews is working to stop this type of misinformation. He and state Rep. Ty Mathews (R-Findlay) proposed House Bill 185that would regulate the content of artificial intelligence.
“It’s based on protecting the name, image and likeness of ordinary Ohioans, whether it involves artificial intelligence, whether it involves the use of Photoshop, whether it involves any image manipulation with or without the use of artificial intelligence,” Mathews said.
The legislation would allow anyone to own their own image, and creating harmful content without someone’s consent would constitute trademark infringement. Civil penalties of tens of thousands of dollars may be imposed.
The bill also makes it a third-degree crime to create or distribute malicious content based on artificial intelligence in order to defraud a person who has been subjected to a deepfake.
Any pornographic deepfakes or deepfakes depicting a child are completely prohibited, with or without disclaimer.
Malicious deepfakes, i.e. artificial intelligence intended to damage someone’s image, will only be allowed if they have a clear watermark indicating that they have been artificially manipulated.
Let’s show some examples:
Original photo
This is the original photo of Mathews wearing a black shirt and standing in front of a barn.
Deepfake 1 – Needs a disclaimer
It’s a counterfeit I made about him with his permission – one that says “Joe Biden” on his shirt. This image, created with Grok, I would require reservations.
Deepfake 2 – No disclaimer required
This is an expertly drawn “deepfake” I made about him. This wouldn’t require a disclaimer.
“Political cartoons are completely protected; no one thinks the cartoon actually happened,” Mathews said
Reaction
The Motion Picture Association of America opposes it, saying it goes against free speech. Earlier, Case Western Reserve University law professor Eric Chaffee told us they were right.
“In fact, requiring these disclaimers could spoil the art or force people to say things they don’t want to say,” Chaffee said. “There are certainly First Amendment concerns.”
However, CWRU technology professor Erman Ayday pointed out that the real challenge with this bill will be enforcement.
“People will keep trying and you will keep changing the prompts,” Ayday said. “I don’t think there’s a 100% guarantee you’ll be able to stop Grok from stripping people.”
Mathews said he hopes his bill will also inspire tech companies to put in place more guardrails for artificial intelligence. In recent days, Company X announced that it has added modern measures to prevent counterfeit, non-consensual pornographic photos.
“Artificial intelligence will continue to evolve and we need to create regulations that do not chase the current level of technology,” Mathews said.
This legislation would also bypass President Donald Trump’s executive order on artificial intelligence. In December, Trump signed an executive order that penalizes states for creating their own artificial intelligence laws. Since this bill covers NIL, attorney Mathews doesn’t think Trump’s AI regulations would apply.
Follow WEWS statehouse reporter Morgan Trau X AND Facebook.
This article was originally published on News5Cleveland.com and are published in the Ohio Capital Journal under a content sharing agreement. Unlike other OCJ articles, it is not available for free republication on other news outlets because it is owned by WEWS in Cleveland.
YOU MAKE OUR WORK POSSIBLE.
