Deepfake. A Powerful New Weapon in The Information War

illustration: bing.com/create

Deepfake technology uses machine learning algorithms to manipulate images and sounds. These algorithms are trained on massive data sets, such as videos and audio recordings, enabling them to imitate specific individuals with great accuracy.

Deepfake can be used for various purposes, both positive and negative. On one hand, it can be used for entertainment, such as creating comedy videos or parodies. On the other hand, deepfake can be used to manipulate public opinion or harm the reputations of public figures.

Deepfake. Real-life examples

The manipulative capabilities of this new technology are increasingly powerful, and the results are becoming more credible and harder to distinguish from real recordings or actual statements. Some of the most well-known examples used to spread misinformation and fuel conflicts include:

• Mark Zuckerberg, explaining that he has control over the world and our personal data;
• Barack Obama, making derogatory remarks about Donald Trump;
• Kim Kardashian, responding to her critics;
• Nancy Pelosi, allegedly intoxicated during a public appearance;
• Donald Trump, lecturing Belgians;
• Volodymyr Zelensky urging Ukrainian soldiers to lay down their arms.

New examples are rapidly emerging, increasingly well-crafted. In recent years, deepfake has been frequently used on social media, where videos spread virally. The danger lies in the fact that the information about their false origin cannot reach all viewers. The growing use of deepfake in political campaigns raises concerns and opposition, including in Poland.

Deepfake as a threat to democracy

Deepfake can pose a serious threat to democracy. This technology can be used to spread misinformation and manipulate public opinion.

For instance, deepfake could be used to create a fake video in which a politician says or does something they never actually said or did. Such a recording could damage their reputation or influence election outcomes.

Deepfake can also be used to intimidate or blackmail public figures. For example, it could be used to create a fake video in which a public figure appears as an adult film actor. Such a recording could then be used to blackmail them into specific actions.

How to recognize deepfake?

There are several ways to recognize deepfake. One is to pay attention to details in the image or sound. In the case of deepfake, certain irregularities may appear, such as:

• improper proportions of the face or body
• unnatural facial or mouth movements
• unnatural sounds

Another way to recognize deepfake is to check if the recording comes from a credible source. If the recording comes from an unknown source or is shared by unknown people, it is likely to be a deepfake.

How to combat deepfake?

There are several ways to combat deepfake. One is to educate the public about this technology. People should be aware of the threat posed by deepfake and should learn to recognize fake materials.

However, currently, only one in four EU residents reports having verified the authenticity of information found on social media or other internet sources. In Poland, this figure is significantly below the EU average—just 16%.

In our country, we also face another problem: a crisis of trust in traditional media. These should help verify the authenticity of information but are increasingly seen as sources of fake news themselves. This is due to various reasons, mainly political bias but also occasional lack of thorough research and... falling victim to fake news.

We have a problem assessing information credibility

A quarter of information consumers rely on sources where verifying credibility is a challenge. The real problem remains the issue of fake news. This is noted by 77% of respondents, with 51% admitting difficulty distinguishing between true and false information, according to the Deloitte Digital Consumer Trends survey.

Digital Poland`s February 2023 report reveals how effective social media is as a channel for spreading fakes:

• Fake news is 70% more likely to be shared.
• The result is that false information spreads six times faster than the truth.

Therefore, another—and perhaps the only effective—way to combat deepfake is to develop technology capable of detecting false materials. Currently, research is underway on artificial intelligence that could be used to recognize deepfake.