Komron Tursunmurodov

Image source: www.eweek.com

In 2023, the Ministry of Internal Affairs alerted citizens of Uzbekistan to a case of cyber fraud involving a deepfake video featuring Prime Minister Abdulla Aripov. Attackers used artificial intelligence to process video sequences with the participation of the head of government and added an audio track, falsely informing about the organization of a committee to refund the funds to victims of cyberbullying with promises of tenfold payments as compensation.

Regrettably, as AI technologies continue to advance, despicable people have been taking an advantage of deepfake technology to spread misinformation, commit fraud, create sexually explicit images, and engage in other harmful activities. This underlines the current need for Uzbekistan to establish regulations around deepfakes in order to avoid its potential terrifying consequences faced by other countries that have been impacted by this technology.

What is a deepfake?

The Oxford English Dictionary defines a deepfake as “any of various media, especially a video, that has been digitally manipulated to replace one person’s likeness convincingly with that of another, often used maliciously to show someone doing something that he or she did not do.” In other words, deepfakes are altered and manipulated video and audio materials, and images that have a potential to harm reputations, fabricate evidences, undermine trust in democratic institutions, and facilitate criminal activities.

Even though deepfakes are primarily associated with negative implications, Caroline Quirk highlights their beneficial uses. For example, deepfake technology can be utilized in a myriad of ways: in education to create interactive content, in the film industry to seamlessly substitute an actor with their stunt double or align dubbing in a foreign language film, and in the retail sector to enhance a prospective buyer’s experience.

Despite the positive applications, deepfakes pose negative impacts onto society. This blog post addresses the potential risks associated with deepfake technology and suggests that the government should take measures to improve the current legislation, implement and refine the regulations concerning deepfakes.

Do we need to regulate AI-generated deepfakes to prevent crimes?

During a parliamentary meeting in Uzbekistan on October 24, 2023, the proposal for the Law “On Amendments and Additions to the Criminal, Criminal Procedure Code of the Republic of Uzbekistan and the Code of the Republic of Uzbekistan on Administrative Liability” was preliminarily reviewed. Oliy Majlis (Uzbek Parliament) highlighted: “In cyberspace, there are cases of falsification of photographs, voices and other biometric data of individuals using special programs.” It also acknowledged the urgent need to strengthen liability for cybercriminals who scam people by demanding money for not distributing deepfakes.

Despite the recent introduction of a new chapter into the Criminal Code titled “Information Technology Crimes,” which establishes criminal liability for nine specific offenses, the current legislation does not encompass provisions for the creation, dissemination, and use of deepfakes for malicious intents. This in turn makes evident the urgency to regulate the utilization of deepfake technology in our country.

Multiple reasons support the regulation of deepfakes in Uzbekistan. Of particularly significant importance are the following:

I. Deepfakes have become a primary tool for spreading misinformation. The deepfake video of Abdulla Aripov promising tenfold payments as compensation serves as a notable example of this reason. The circulation of such deceptive content not only damages specific individuals’ reputation, but also undermines the credibility of government institutions. For instance, a potential deepfake video involving a minister of internal affairs allegedly accusing a general prosecutor’s office in state capture would harm the reputation of both law enforcement agencies.

II. Deepfake technology poses a threat to women’s safety. In many countries women’s dignity, safety and identity are being compromised by malicious individuals who create pornographic content and sexually explicit images. Instances of AI-generated explicit content targeting public figures like Taylor Swift, sparked discussions on improving the legislation and criminalizing the creation of deepfakes in the US. The Uzbek government must ensure women’s security, as guaranteed by the Constitution and other legal frameworks.

III. Deepfakes facilitate fraudulent activities. According to the Director of Global Policy at Onfido, Aled Owen, fraudsters are increasingly using deepfakes as a way to attempt to dupe identity verification systems. In Uzbekistan, perpetrators are exploiting deepfakes to engage in fraudulent schemes, such as impersonating directors of firms or family members to extort money from stakeholders or relatives. Recent examples of such cases include businessman Zafar Khashimov and General Director of the Navoi Mining and Metallurgical Combinat Sanakulov Kuvondik. The ability of deepfakes to imitate gestures, facial expressions and mimic voices flawlessly and convincingly complicates the detection of fraudulent activities, leaving ordinary citizens vulnerable to deception.

Accepting these concerns, the Uzbek Parliament should enact comprehensive regulations addressing the creation, distribution, and malicious use of deepfakes to combat misinformation, safeguard women’s rights, and decrease fraud crimes in the country. We recommend that these regulatory measures are incorporated into a dedicated legislation, such as a specific Law “On Artificial Intelligence” or “On the Regulation of Artificial Intelligence”. Such regulatory measures will preserve trust, security, and integrity in a digital sphere and protect individuals from the harmful consequences of deepfake manipulation.

Interestingly, most countries have already developed their laws which specifically regulate AI deepfakes. In the United States, the Congress has passed a bill that focuses on safeguarding national security against the threats posed by deepfake technology and ensuring legal recourse for victims of detrimental deepfakes. Similarly, the Chinese government has implemented stringent regulations, known as the Deep Synthesis Provisions, which restrict the creation of deepfakes without user consent and mandate verification of AI-generated content. In the United Kingdom, the adoption of the Online Safety Act in 2023 has made it illegal to share digitally manipulated explicit images or videos, thereby addressing concerns related to the dissemination of misleading content online. Furthermore, in European Union, deepfakes will be regulated under the AI Act, which stands as the world’s first comprehensive AI law.

What improvements in law shall be made?

Uzbekistan has signed onto the UN Resolution 78/265 “Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development” adopted by the General Assembly on March 21, 2024. This resolution emphasizes the importance of establishing global consensus on safe, secure and trustworthy artificial intelligence systems that are consistent with international law, in particular the Charter of the United Nations, the Universal Declaration of Human Rights, and the 2030 Agenda for Sustainable Development. This shows our countries political commitment to regulating AI innovative technologies, including AI-generated deepfakes.

Based on this and taking into account the harmful impacts of deepfakes on Uzbek society, it is recommended that our parliament and government take the following steps:

Firstly, the Oliy Majlis should enact a law on regulation of AI. This law should set forth fundamental principles for the use of AI technologies, define key terms, such as AI or deepfakes, designate the government institutions that will be responsible for overseeing AI applications in various sectors, and outline other regulatory frameworks for the use of AI.

Secondly, Uzbek legislators should consider criminalizing the creation, dissemination and utilization of deepfakes for malicious purposes. A new article should be added to Chapter XX1 of the Criminal Code to ensure criminal liability for offenses involving deepfakes.

Finally, the legislative body should introduce the use of AI in the commission of crimes as an aggravating factor in sentencing. Article 56 of the Criminal Code should be amended and include the use of AI in committing crimes as a new circumstance aggravating punishment. The main reason of this amendment lays in the fact that an offender leveraging AI technologies may cause greater harm to society, or economic, social and political institutions of the government, in comparison with  traditional forms of crime.

Conclusion

This blog post aimed at proposing our government to establish a regulatory framework for AI technologies, specifically establishing criminal liability for creating, sharing and using deepfakes in order to commit crimes. The key findings illustrate that the deepfakes as significant tools for spreading misinformation, pose risks to women’s safety, dignity and identity, and facilitate criminal activities.

It is essential for Uzbekistan to introduce a new law that regulates AI technologies, specifically criminalizing malicious uses of deepfakes, and amend other legislative acts. These changes shall be accompanied by drawing on the experience of developed countries such as the US, the UK, and China. Uzbekistan should prioritize these measures in order to prevent potential serious repercussions of deepfakes and make sure that the perpetrators are held accountable and are brought to justice.

Cite as:  Komron Tursunmurodov, “The Deepfake Dilemma: Safeguarding Society Through Regulation”, Uzbekistan Law Blog, 09.07.2024.