Deepfake technology poses an existential threat to trust

What's this about?

People disagree about whether deepfake tools could destroy trust across all of society. They can harm some kinds of trust, but we are not sure they could ruin trust everywhere.

What supporters say

• Deepfakes can make people trust videos, voices, and other proof less.
• People accused of wrongdoing may call real recordings fake, using doubt to avoid blame.
• Deepfakes can make scams, false claims, and fake acts by well-known people easier to spread.

What critics say

• Current facts show real risks, but they do not show that trust would fall apart across society.
• Tools that check where media came from, and spot fake media, can help people keep trust.

How to read this

The number of points on each side does not show which side is right; strong proof matters more than a long list.

The bottom line

Deepfakes clearly threaten trust in recordings, public words, and online proof. They may also help scams, lies, and unfair attacks spread.

But the evidence does not show that deepfakes will cause trust to collapse across society. Checks that show where media came from, plus tools that find fakes, may help protect trust.

Deepfake technology clearly threatens some forms of trust, especially confidence in recordings, public statements and online evidence. But the available evidence does not show that it will cause an irreversible, society-wide collapse of trust.

The case for

The strongest argument is that deepfakes can weaken confidence in visual and audio evidence. As realistic fabrications become easier to create and share, people may no longer accept an apparently direct recording as proof. A peer-reviewed review links deepfakes to misinformation, media manipulation and uncertainty about what is authentic. A United Nations synthesis places the issue alongside wider concerns about information integrity, elections, privacy and human rights. (see Figure 1) 1

This problem can shift the burden of proof. Audiences, journalists, courts and institutions may need outside verification before trusting material that once appeared self-evident. That could make it harder to establish what really happened, even when genuine evidence exists.

Deepfakes also make impersonation and targeted deception easier to scale. Synthetic media can attach false words or actions to recognizable politicians, officials, companies or private individuals. The cited material describes links to political misinformation, fraud, reputational attacks and abuse. These harms can damage trust in particular people and institutions, even if their overall effect on society has not been measured. 2

A further danger is the so-called “liar’s dividend.” People accused of misconduct may dismiss authentic recordings as fabricated, taking advantage of the public’s uncertainty. Trust depends not only on true information existing, but on whether others can prove that it is genuine. False denials could therefore make accountability more difficult. 3

The case against

The central weakness in the broader claim is that an existential threat has not been demonstrated. Existing research shows serious risks and falling confidence in digital information more clearly than it shows a permanent, society-wide breakdown of trust. Reviews and institutional reports identify possible mechanisms of harm, but they do not establish how common the effects are or whether they will continue over the long term.

The evidence also does not directly show that deepfakes cause irreversible declines in social or institutional trust. Some user-focused research measures people’s perceptions rather than broader changes in society. Case examples, including an industry-produced compilation, illustrate what could happen but cannot reliably show how widespread the problem is.

There are also ways to respond. Detection tools, provenance records, labels and stronger rules could help preserve trust by giving people ways to check whether media is authentic. Research suggests that people’s willingness to rely on detection systems depends on how accurate and reliable they believe those systems are. These safeguards are not guaranteed to work everywhere, but their existence shows that trust need not depend entirely on unaided human judgment. 5

Trust is shaped by institutions and political experiences as well as technology. One cited voting study found that participation can increase political trust, suggesting that public confidence responds to many factors. Deepfakes may intensify existing problems, but they cannot be treated as the sole cause of changes in trust.

The bottom line

The evidence moderately supports a narrower claim: deceptive or poorly governed deepfakes can harm evidentiary, political and interpersonal trust, and can raise the cost of finding out what is real. It does not support the stronger claim that deepfakes pose an existential threat to trust as a whole.

The key uncertainty is whether these concentrated, context-dependent harms will add up to an irreversible collapse across society. So far, there is no strong population-level, long-term evidence showing that they will. Adaptive institutions and authentication systems may limit the damage, although their accessibility, accuracy and effectiveness remain uncertain.