Deepfake: a fun trend or an offense?

In the cult movie Forest Gump, there was a scene where Tom Hanks’ character shook hands with the fake Kennedy. This is how deepfakes were made in the pre-deepfake era. Now, deepfakes have flooded social media, and the technology itself is in the news almost every day. And today, Tom Hanks has to warn his fans that he has nothing to do with the advertisement for a certain dental clinic in which his fake clone was featured.

Deepfakes can be images, videos, or audio that reproduce artificially intelligent (AI) generated but fairly realistic content. Most deepfakes still have blurring, uneven edges, unnatural skin tone, disproportionate body parts, unnatural movements, etc. However, the technology is evolving, and it is likely that over time, high-quality deepfakes will become harder to detect.

Forbid impossible allow

But is it legal to use deepfakes? The approach to regulation will differ depending on the country we are in and the purpose for which we use the technology. Deepfakes are created for fun and parodies but also for fraud, PSYOP, and other dirty purposes. In countries such as China and the United States, the legislature has already restricted certain ways of using this technology. It is prohibited to use deepfakes if it is done for the purpose of discrediting a person, committing crimes, for political purposes, or for creating pornography with images of people who did not consent to it.

Ukraine does not yet have a specialized law on the use of deepfakes, but there are a set of related aspects whose regulation can lead to a conclusion about the commission of a violation or its absence.

The legality of using a deepfake will depend on the answers to the following questions:

• Whose image is used in the content? Did this person give their consent? If not, then the deepfake will violate the person’s non-property right to their own image.
• Does the generated content contain the works of other persons? If the copyright holder’s permission to use them has not been obtained, this may be a violation of intellectual property rights.
• Does the content contain any information that disparages another person, presents their seamy side, or attributes questionable statements or actions to them? Such content may violate an individual’s right to respect for honor, dignity, and business reputation.
• In general, what is the purpose of the content? Is it a harmless joke, entertainment, parody or revenge, misleading, fraud? In the latter cases, it may be an offense.

Deepfake and intellectual property

Sometimes, a piece of digital content can be considered a copyright: this is possible if the generated content is an original intellectual creation of the author. For example, if the author took the photo of themself, chose the music for the video, came up with the staging, or shot the video. In this case, the deepfake is protected from the moment it is created, regardless of the genre, purpose, or completeness.

Although a deepfake is close to such copyrights as a photograph, videogram or audiovisual work, the peculiarity of the technology is that AI creates the final object based on other images, videos, or audio recordings. In this sense, a deepfake is a derivative work, the result of the creative processing of another work. But let’s not forget that the rights to the works that have been processed must be respected.

In practice

It’s clear that deepfakes are not just for jokes. For example, the Ukrainian music label PAPA music actively uses deepfakes with the image of the artist Anton Velboi (Wellboy is a project of the PAPA music label and Anton Velboi is the artist’s pseudonym on the Velboi Family label), since the artist himself no longer works with the label. Is this legal? We would like to remind you that the legality of a deepfake may be affected by the permission of the person depicted, the purity of the copyright, the purpose and modality of use, and other factors. A label relies on a contract signed with an artist, but how far can such imitation go and can a contract take away an artist’s identity? Sooner or later, it seems, the court will have to resolve all these ambiguous issues.

In addition to legal aspects, technology itself entails a number of dangers, as it is a fairly simple tool for influencing the minds of naive audiences. If nowadays fakes can be recognized in most cases (at times they are so low-grade that it is even funny, as was the case with the Russian diplomatic phone with Volodymyr Zelenskyi’s alleged speech on Ukraine’s surrender), we are waiting with eagerness and a little fear to see what heights technological development will reach in a year or two.

Meanwhile

The use of deepfakes in Ukraine is not directly regulated, but it touches on a number of related issues. Analyzing these issues will help us understand whether everything is okay or whether it is time to prepare for court battles.

Author: Mariia Yesypovych, Associate at Axon Partners