Deepfake Blurs Truth and Falsehood. Human Perception Research

illustration: DALL-E

Researchers Qiang Liu, Lin Wang, and Mengyu Luo from the University of Science and Technology in Shanghai set out to examine how deepfake technology affects the perception of information and users` trust in media. In their article When Seeing Is Not Believing: Self-efficacy and Cynicism in the Era of Intelligent Media, published in Humanities and Social Sciences Communications, they conducted two experiments involving 1,826 participants, analyzing how cynicism toward information changes depending on users` ability to recognize AI-generated content.

The experiments revealed that:

• Individuals with low self-assessment in recognizing AI content are more likely to question the authenticity of news that is personally significant to them.
• Low-risk content paradoxically raises more skepticism than content considered high-risk.
• Users who repeatedly struggle to assess the authenticity of deepfakes lose confidence in their abilities and abandon efforts to verify information.

This phenomenon leads to the so-called "apathetic reality," where audiences choose indifference over critical thinking regarding media consumption.

Why Are We Losing Trust in Media?

The growing cynicism toward AI-generated information is not just a technological issue but also a matter of how audiences process content. Research suggests that users engage more in content analysis when they feel the topic directly affects them.

Data shows that:

Factor	Impact on AI Self-Assessment	Impact on Cynicism
High content relevance	Increased confidence	Reduced cynicism
Low content relevance	Decreased confidence	Increased cynicism
High-risk news	Greater inclination to verify	Lower level of cynicism
Low-risk news	Less interest in verification	Higher level of cynicism

The study`s authors highlight the effect of cognitive fatigue. When users repeatedly encounter situations where they cannot distinguish deepfakes from real content, they stop making an effort to check. Ultimately, instead of verifying authenticity, they begin treating all information as potentially unreliable.

What Can We Do?

Experts emphasize that the solution lies not only in developing deepfake detection technologies but also in shaping new media literacy models.

• Social media platforms should implement more advanced mechanisms for labeling AI-generated content.
• Users should be trained not only in recognizing fake news but also in consciously processing information in an era of informational chaos.
• Journalists should make greater use of content verification tools and build communication strategies based on source transparency.

Liu, Wang, and Luo`s study found that even a small increase in users` self-confidence regarding AI leads to a significant reduction in cynicism and greater engagement in content analysis. This means that education and tools supporting content verification can help audiences regain control over what they consider true.

The Future of Trust in Information

Deepfake news is a challenge we will face for years to come. In the age of artificial intelligence, it is not just technology that determines what we believe but also our ability to recognize and critically analyze content. If we do not begin developing skills to navigate the world of synthetic information, we may find ourselves in a reality where we cannot even trust what we see with our own eyes.

* * *
Article by Liu, Q., Wang, L., Luo, M. (2025) When Seeing Is Not Believing: Self-efficacy and Cynicism in the Era of Intelligent Media, published in Nature Humanities and Social Sciences Communications, available at
https://www.nature.com/articles/s41599-025-04594-5