Why the Deepfake Challenge Goes Far Beyond Technology

Artificial intelligence is rapidly transforming public administration, business, education, and many other sectors. The rise of generative AI has created new opportunities for producing digital content. At the same time, it has introduced new risks.

One of the most significant challenges of recent years is the emergence of deepfakes—images, audio recordings, and videos created or manipulated by artificial intelligence in ways that make them difficult to distinguish from authentic content.

However, the primary danger of deepfakes is not limited to deceiving individuals. They are gradually undermining the reliability of the traditional methods we use to verify information.

For decades, photographs, audio recordings, and videos have been regarded as evidence of real events. Today, AI technologies can generate content so realistic that even experts may struggle to determine its authenticity.

As a result, people are beginning to question not only fake content but also genuine content. Any photograph, video, or voice message can now be dismissed as fabricated. This creates risks for citizens, businesses, public institutions, and the digital environment as a whole.

The consequences of deepfakes are increasingly visible in real life.

In 2019, fraudsters in the United Kingdom used an AI-generated imitation of a company executive’s voice to persuade an employee to transfer a substantial amount of money.

In 2024, criminals in Hong Kong conducted a video conference using digital replicas of company executives. An employee failed to detect the deception and followed the instructions, resulting in multimillion-dollar losses.

Another growing threat is the use of synthetic voices and deepfake videos through mobile communications, messaging platforms, and video conferencing services.

In some cases, just a few publicly available photographs or a short voice recording are enough to create a convincing digital replica of a person. As a result, traditional methods of identity verification are becoming less reliable.

This issue is also becoming increasingly relevant for Kazakhstan. Criminals are already using modern digital technologies to deceive citizens and gain access to their financial resources and personal data.

In response to these challenges, governments, technology companies, international organizations, and professional communities are developing AI Governance frameworks—systems designed to ensure the responsible, safe, and trustworthy use of artificial intelligence—and implementing measures to counter deepfakes.

Among the most promising approaches are:

​•​labeling AI-generated content;
​•​digital watermarking;
​•​tools for detecting synthetic media;
​•​stronger safeguards against misuse;
​•​improving digital literacy.

Technologies that verify the origin and authenticity of digital content are becoming particularly important. They help identify the creator of content, determine whether it has been altered, and strengthen trust in digital information. One example of this approach is Content Credentials.

What Can Be Done Today?

Addressing deepfakes cannot rely solely on improving detection technologies. A comprehensive approach is needed—one that brings together governments, businesses, and society.

For governments, key priorities include strengthening AI Governance frameworks, improving legislation, enhancing cybersecurity, and introducing tools that help verify the authenticity of digital content.

For businesses, reviewing identity verification procedures and approval processes for important decisions is becoming increasingly important. A voice, photograph, or even a video call can no longer be considered absolute proof of identity. Organizations should therefore implement additional verification mechanisms.

For citizens, digital literacy remains one of the most effective forms of protection. In an environment where deepfakes are becoming more widespread, it is important to approach unexpected information critically, especially when it involves financial transactions, personal data, or urgent requests.

A simple rule should not be forgotten: important information should always be verified through additional communication channels, even when it appears convincing and seems to come from a trusted person.

Kazakhstan should continue to draw upon international best practices and actively contribute to the development of a safe and trusted digital environment.

In the age of artificial intelligence, the key challenge is no longer merely obtaining information, but verifying its authenticity.