As state-level legislation regarding deepfake continues to be introduced in the United States, ..

Amid increasing state-level legislation on deepfake in the United States, it is notable that California, home to Silicon Valley, the “home of big tech companies,” has pulled the trigger.

The deepfake regulation bill passed by the California legislature is significant because it will be the first comprehensive regulation of artificial intelligence (AI) deepfakes in the United States if Governor Gary Newsom signs it. The bill could punish those who create AI-generated sexual exploitation of children, not just real children, and could punish those who distribute and possess deepfake sexual exploitation. According to ABC, 35 of the world’s 50 largest companies are based in California, and they have launched super-tight regulations for companies based in the state.

AI is seen as a key factor that will determine national competitiveness in the future.

On the surface, the world says it is in favor of “safe AI,” but behind the scenes, the serious reality is that they are engaged in a fierce war of nerves to protect their own industries. The European Union (EU) is introducing the Digital Services Act (DSA) and ensuring that platform companies take responsibility and have a system in place to respond to illegal and harmful content, fake news, services and products. Platform companies must take responsibility and quickly remove illegal and harmful content or fake news so that it does not spread.

However, many say that it was a decision by the EU, which has no domestic companies comparable to Big Tech, to promote its own companies and keep Big Tech in check. This is the background of the “Sovereign AI” that calls for promoting its own models without being subservient to Big Tech. In other words, you can say that EU sanctions against Big Tech and US sanctions are fundamentally different in nature. Analysts say that the United States, which has the most AI companies, is a leader in protecting its own industry, but recently, when sexual exploitation with deepfakes flooded in, has decided to tighten regulations in this area.

According to a report by Multistate, a government policy advisory group, as of the end of June, there were 27 Maguk state governments that had passed laws regarding deepfake sexual exploitation. Legislation has been moving at a rapid pace in recent years, with 20 places passing legislation since December last year.

According to the New York Times (NYT), a case in October last year, in which boys created, shared and distributed deepfakes with real photos of 10th graders (14 years old) in the same class, caused a lot of commotion in the community. Similar incidents have occurred in the United States, including at Beverly Hills High School in California. When the deepfake incident shocked the entire United States, the state started to create similar laws. In the past, there were many forms of inclusion of content created by deepfakes in the law on rebenziforno (distributing sexual photos or videos of former lovers) or sexual exploitation of children.

The Deepfake Child Sex Offenses Act, recently passed by the California House of Representatives, is seen as a stronger bill because it bans images of children created by AI, not real people. Unlike SB1047, a sweeping AI regulation bill also being pursued in California, the Deepfake Regulation Act was passed as quickly as it wreaked havoc on everyday people.

There are calls for Korea to benefit from this trend and actively regulate the big tech companies that are a distribution channel for deepfakes.

Until now, South Korea has been unable to impose tough sanctions on big tech companies due to concerns such as diplomatic disputes with the United States. However, analysts say the country should use the recent global sanctions on deepfakes to respond quickly.

“The US no longer favors self-regulation, but more strongly increases the level of business regulation,” said Kim Yoon-myung, a professor at Hallym University of International Studies. “This is a conditioned move where only ‘safety-guaranteed AI’ can establish itself in the market with sustainability, and Korea is relatively slow in terms of regulatory maintenance, which only increases the uncertainty in the market.” “Instead of creating a new law on harmful deepfakes, priority should be given to clearly diagnosing what is lacking in the boundaries of the existing legal network and jointly making up for it in a rapid manner,” he said.