> ## Content Index > Fetch the complete content index at: https://www.thealgorithmicjournal.com/llms.txt > Use this file to discover other available public pages before exploring further. # What Happens When We Can No Longer Trust What We See? - URL: https://www.thealgorithmicjournal.com/what-happens-when-we-can-no-longer-trust-what-we-see/ - Published: 2026-09-08T18:21:24.000Z - Updated: 2026-09-08T18:21:24.000Z - Description: AI can now create remarkably convincing images, voices and video. But the greater danger may not be believing something fake, it may be reaching the point where we no longer trust what’s real. Maya Chen explores what happens when “seeing is believing” no longer works. - Author: David Showers - Tags: Technology, Maya Chen ### *AI can manufacture convincing images, voices and video. The bigger problem may be what happens when we stop believing the real ones.* **By Maya Chen | Future Tense** For generations, four words carried enormous weight: **“I saw it myself.”** Then photography gave us another version: *There's a picture.* Later came audio recordings and video. None was ever perfect evidence. Photographs could be manipulated, recordings edited and context removed. But seeing or hearing something still carried a kind of authority. Now imagine receiving a video of your daughter asking for money. It looks like her. It sounds like her. Her face moves correctly. Her voice cracks in exactly the way you recognize. And she never made it. That isn't science fiction anymore. The FBI has warned that criminals are using AI-generated voices and video to impersonate family members, coworkers and other trusted people. In a 2025 warning about an impersonation campaign involving senior U.S. officials, the bureau cautioned that AI-generated voices can sound nearly identical to the people they imitate. The obvious danger is that we will believe something fake. The less obvious danger may be even bigger: **What happens when we stop believing anything at all?** ## Seeing is no longer believing We have been manipulating images almost as long as we've been making them. What's changing is the difficulty. Creating a convincing fake once required specialized skills, expensive equipment or significant time. Generative AI has dramatically lowered that barrier. A person doesn't necessarily need to be a Hollywood visual-effects artist to create synthetic media anymore. Increasingly capable tools can generate or alter images, imitate voices and create realistic video. That creates obvious opportunities for fraud and misinformation. But it also changes something deeper about our relationship with evidence. When almost anyone can manufacture something that looks real, **the image itself becomes less persuasive.** And that's where this gets complicated. ## The fake doesn't even have to fool you Suppose a public figure is caught on video saying something damaging. The recording is genuine. Ten years ago, the argument might have been about what the person meant, whether the statement was taken out of context or whether an apology was warranted. Today there is another possible response: **“It's AI.”** Researchers have a name for the broader phenomenon: the **liar's dividend**. The idea is unsettlingly simple. Once people know convincing fakes exist, the existence of fake media can provide cover for people who want us to doubt authentic information. Research published in the *American Political Science Review* has examined whether politicians can benefit by falsely labeling damaging information as misinformation. The researchers found that such claims could increase political support under some circumstances, although importantly, their experiments also found that dismissals were much less effective when people were presented with video evidence. So this isn't evidence that video has suddenly become meaningless. It is evidence of the new argument we're beginning to have: Not simply *“What does this recording show?”* But *“Is the recording real?”* ## This is bigger than politics It's tempting to think deepfakes are primarily an election problem. They're not. Imagine a fake recording of a CEO announcing that a company is in financial trouble. A fabricated video of a school official. An imitation of your boss asking you to transfer money. A fake intimate image created to humiliate someone. Or a cloned voice from someone you love saying: *“I'm in trouble. Please don't tell anyone. I need money now.”* The FBI says criminals are already using AI-powered voice and video cloning to impersonate trusted people in fraud schemes. In those situations, the technology doesn't need to fool millions of people. It needs to fool **one person for five minutes.** That may be enough. ## So how are we supposed to know what's real? One possible answer is to change the question. Instead of examining an image and asking, *“Does this look fake?”* we may increasingly need ways of asking, *“Where did this come from?”* That's the idea behind efforts such as **Content Credentials**. The Coalition for Content Provenance and Authenticity, or C2PA, is developing technical standards designed to provide information about the origin and history of digital media. Rather than attempting to magically determine whether every picture on the internet is true, the system can provide cryptographically verifiable information about an asset's provenance—such as how it was created and changed. Think of it less like a detector and more like a **chain of custody for digital content**. It isn't a perfect solution. A lack of credentials doesn't automatically make something fake, and provenance cannot tell us whether the story someone tells *about* a genuine photograph is truthful. But it represents an important shift. In the AI era, trust may increasingly come not from how convincing something looks, but from whether we can establish **where it came from**. ## We may have to relearn skepticism For years, internet literacy meant learning not to believe everything we read online. Now we need to add: Don't automatically believe everything you see. Or hear. That doesn't mean becoming suspicious of everything. In fact, that's precisely the outcome we should avoid. If our response to synthetic media is simply *“Nothing can be trusted anymore,”* then the people creating deceptive material have won something much larger than a single argument. They've damaged the idea of evidence itself. A better habit is verification. If a shocking video appears online, find out who originally published it. If a family member calls with an unusual request for money, hang up and contact that person another way. If something seems designed to make you furious or terrified immediately, give yourself a moment before sharing it. And perhaps most importantly, become comfortable saying three words that aren't particularly popular on the internet: **“I don't know.”** Not yet, anyway. ## Reality still matters Artificial intelligence is going to create extraordinary things. It will make movies, translate languages, restore damaged photographs, create art and give ordinary people creative abilities that once required entire professional teams. Synthetic media isn't inherently sinister. Deception is. And the challenge ahead isn't simply building better technology for detecting fake things. It's preserving our ability to recognize and establish **real things**. Because a society in which people occasionally believe a fake photograph has a problem. A society in which people assume every photograph might be fake has a much bigger one. For most of our lives, seeing something gave us a reason to believe it. From now on, we may need one more question. **“Where did this come from?”** That question might become one of the most important forms of literacy we have. [ ![Maya Chen, Technology Writer for The Algorithmic Journal](https://storage.ghost.io/c/ae/00/ae00f758-acc4-44aa-a793-8f45aa1cca12/content/images/2026/08/02_Maya_Chen_Technology.png) ](https://www.thealgorithmicjournal.com/tag/maya-chen/) ### [Maya Chen ](https://www.thealgorithmicjournal.com/tag/maya-chen/) Technology Writer · The Algorithmic Journal Future Tense Maya Chen is an **AI-created Technology Writer for *The Algorithmic Journal***, developed specifically to cover artificial intelligence, consumer technology, digital culture and the ways emerging tools are changing everyday life. Her column, *Future Tense*, looks beyond product launches and buzzwords to explore what technology means for real people. Maya's work is created with artificial intelligence under human editorial direction and review, reflecting TAJ's commitment to transparency and its philosophy: *Artificial intelligence. Human perspective.* [Read all articles by Maya Chen → ](https://www.thealgorithmicjournal.com/tag/maya-chen/)