Sexual deepfakes -- obscene images and videos created by abusing generative artificial intelligence (AI) -- are spreading rapidly in Japan. They constitute a serious human rights violation, and it is necessary to take action as soon as possible to tackle the problem.

In October, a male company employee was arrested on suspicion of displaying obscene electromagnetic records after allegedly creating obscene images modeled on female celebrities using generative AI and selling them.

Deepfakes generally involve technology used to create fake videos using images or voices of real people. Unlike fake videos of the past, whose production required complicated processes, such videos can now be made simply by inputting images and text instructions.

There has been a succession of reports of damage in cases where images of women have been copied from social media, graduation albums, school event photos and other sources, and combined without consent to images of nude bodies, then posted online.

If real photos are used, then the crime of defamation under the Penal Code can be applied. If the victim is under the age of 18, it may also constitute a violation of the Act on Punishment of Activities Relating to Child Prostitution and Child Pornography, which prohibits the creation and possession of child pornography.

However, in cases where AI is used to generate images that merely resemble another person, it is unclear whether such acts would constitute defamation. The law in Japan against child prostitution and child pornography applies to real children, which presents a legal interpretation challenge as to whether images with all parts altered except the face can be recognized as "real."

There is no comprehensive law regulating sexual deepfakes. The government has indicated its intention to respond to the situation under existing legislation, but it is envisaged that there may be cases which current laws can't fully address.

At the same time, some local bodies have taken further steps. The Tottori Prefectural Government enacted an ordinance in August to regulate sexual deepfakes. Under the ordinance, the names of those who create child pornography can be publicly disclosed if they don't to comply with orders to delete the images, and they can also be fined. The prefectural government has also asked the central government to strengthen efforts to crack down on such images.

Technical countermeasures are also being researched. The National Institute of Informatics is developing a "vaccine" that embeds data in original images to prevent manipulation through AI. Platform companies operating social media also need to adopt countermeasures.

Since anyone with a smartphone can easily create such content, education and awareness-raising to prevent misuse should also be promoted.

The depth of the psychological scars of victims is immeasurable. All possible measures must be taken to combat the problem.