EDITORIAL | Regulate Generative AI Before Deep Fakes Further Harm Society

Fake news videos and other misuse of generative AI and artificial intelligence on the Internet are causing a stir.

For example, a fake video of Prime Minister Fumio Kishida has caused considerable controversy in Japan. In it, the Prime Minister appears to make an indecent remark.

A deluge of such fake news videos could disrupt society. They cannot be considered mere pranks. The Government should take immediate action to prevent the spread of false information that misleads the public.

To create the fake video of Prime Minister Kishida, the creators first used actual Nippon TV news report footage. Then, artificial intelligence learned how to sound like the Prime Minister's voice while making the lewd remark. Then they posted the fake news clip on X (formerly Twitter) and other platforms, where it has been viewed millions of times.

A fake video was also behind the creation of another purported Nippon TV news program. In that case, imitating a real female announcer, the generative AI video urged people to register on an investment information website. It caused such an uproar that the company had to call on viewers to be careful.

*The photo presents a comparison of deep fake image (left) of Ukrainian President Zelenskiy posted on X (formerly Twitter) and real image (©Kyodo)*

Call for Caution

Fake images and videos made with generative AI are called "deep fakes" and can be easily created using online apps.

In the case of the fake Kishida video, the quality was so poor that it was immediately recognizable as fake. But AI technology is rapidly evolving. Actually, it is changing so rapidly that there is concern fake videos will soon be indistinguishable from the real thing.

Overseas, generative AI videos are already causing tremendous harm. In May 2023, the New York Stock Exchange Dow Jones Industrial Average temporarily plummeted because of an AI video. In that case, it spread false images online of an alleged explosion near the United States Department of Defense.

Numerous AI-generated images and other dubious content have also been circulating in the Palestinian territory of Gaza. In that example, fighting between the Israeli army and the Islamic fundamentalist organization Hamas continues.

*Prime Minister Kishida talks about the possibilities and risks of generative AI at the United Nations conference "Internet Governance Forum" on October 9, Kyoto (© Kyodo)*

Different Strategies in Reactions to Generative AI

Against this backdrop, the European Union is moving to regulate deep fakes.

Meanwhile, the Japanese government has also set up an AI Strategy Council and is working on guidelines. Nonetheless, it has not yet moved to regulate AI, prioritizing technological innovation in the field instead.

Although regulation may conflict with freedom of expression, deep fakes differ from mere satire and parody. If they are allowed to circulate in society, people will not be able to make decisions based on normal information. In that case, the foundations of democracy could be undermined.

This is a problem that puts society at risk. There is therefore an urgent need to tackle the spread of the problem, including through regulation.

The public also needs to be on the alert. We should be concerned that foreign actors may spread false information to disrupt Japanese society, for example during elections. Net literacy is necessary to avoid being led astray by deep fakes.