California looks to Europe to rein in AI

California, home to Silicon Valley, is keen to rein in the adoption of artificial intelligence and is looking to a tough European approach to big tech for inspiration.

California, the richest state in the US by GDP, is a hotbed of unbridled technological innovation, but lawmakers in the state capital, Sacramento, want to give the industry the industry laws and guardrails it has largely been spared in the internet age.

Brussels is rushing to pass an AI law, enacting a series of laws on the US-dominated technology after OpenAI’s Microsoft-backed ChatGPT debuted in late 2022 and sparked a global AI race.

“What we’re trying to do is actually not just learn from the Europeans, but work with them to bring regulation to AI,” said David Harris, senior policy advisor for the California Technology and Democracy Initiative. It’s about finding a way.” .

Just as they have enacted EU laws on personal data in the past, California lawmakers are taking note of recent European legislation on AI, especially given that there is little prospect of comparable domestic legislation coming out of Washington. ing.

Harris said there are at least 30 bills proposed by California lawmakers related to various aspects of AI. Harris said he has advised domestic and European officials on such laws.

Proposed bills in California range from requiring AI manufacturers to disclose what was used to train their models to banning election ads that include computer-generated features.

“One of the aspects that I think is really important is how do we deal with deepfakes and fake texts that make it look like a human being is sending a message,” Harris told AFP.

State Rep. Gail Pellerin is sponsoring a bill that would effectively ban the spread of deceptive digital content created by AI in the months leading up to and weeks after the election.

“The bad guys who are taking advantage of this really want to cause chaos in the election,” Pellerin said.

“Bad person” who breaks the law

Industry association NetChoice is adamantly opposed to importing European laws on AI and other aspects of EU technology regulations.

“They’re basically taking a European approach to artificial intelligence, which means they have to ban the technology,” said Karl Szabo, general counsel for the association, which advocates for light regulation of the technology. Stated.

“(Nothing) will stop the criminalization of AI. It’s bad when the bad guys don’t follow the law,” Szabo argued.

“That’s what makes them the bad guys.”

The US computer software giant Adobe, like many tech giants, has worked with Europe on AI legislation, said Dana Rao, Adobe’s general counsel and chief trust officer.

At the heart of EU AI law is a risk-based approach, with AI practices deemed riskier coming under greater scrutiny.

Rao said he was “satisfied with how the AI ​​Act ultimately turned out” due to the high-risk, low-risk approach.

Rao said Adobe engineers are already conducting “impact assessments” to assess the risks of AI products before making them available.

“We should think about nuclear safety, cybersecurity, and AI when making important human rights decisions,” Rao said.

“Looking at California”

In California, Rao said he expects the issue of deepfakes to fall under the new law’s jurisdiction for the first time.

Assembly Bill 602 would criminalize non-consensual deepfake pornography, and Assembly Bill 730 would ban the use of AI deepfakes during elections.

To combat this, Adobe worked with other companies to create what Rao calls “content credentials,” the equivalent of a “nutrition label” for digital content.

Congressman Pellerin hopes that the AI ​​law passed in California will be replicated in other states.

Many US states are also working on their own AI deepfake legislation, and “people are looking at California,” Pellerin said.

“We are all in this together, and we must stay ahead of those who are trying to wreak havoc in the election,” she said.