Ancient Wisdom for Navigating Modern AI Fakes

You don't need sophisticated technology to fool people online. In a study, researchers showed over 3,000 high schoolers a grainy video of what appeared to be poll workers rigging an election. A simple, aggressive caption was enough to convince most students of U.S. voter fraud, even though the footage was actually from Russia. Only three students discovered the truth.

For a long time, experts have warned that crude, easily made "cheap fakes" were a greater immediate threat than complex deepfakes. Even with AI's recent advancements, the last election cycle saw debates about President Biden's fitness fueled by deceptively edited old-school videos.

The New Era of AI-Powered Deception

The age of cheap fakes is quickly coming to a close. We are now entering a more treacherous information landscape filled with viral deepfakes created by powerful new video generation tools. With products like Google’s Veo 3, OpenAI’s Sora 2, and Meta’s Vibes, a flood of AI-generated "slop" is metastasizing across our screens. This problem is compounded by social media platforms largely abandoning their fact-checking responsibilities. To navigate this new reality, we may need to turn to ancient wisdom: the emphasis on reputation that is central to Muslim, Jewish, Buddhist, and other faiths.

An Ancient Solution for a Modern Problem

For centuries, various traditions have prioritized the source of information. Devout Muslims trace the sayings of Muhammad through a “chain of narration,” or “isnad.” Religious Jews consider the rabbi behind Talmudic teachings. Tibetan Buddhists rely on an oral lineage tracing back to the Buddha. These traditions all encourage reasoning, but only after first tracing the information to its origin and assessing the reputation of those who shared it.

This principle of reputation is not just religious; it's a fundamental part of secular life. We rely on it when choosing a doctor, hiring a plumber, or picking a restaurant. We consult trusted friends and read reviews because we know that people and businesses are unlikely to advertise their own flaws. If reputation is so critical in these areas, why do we so often ignore it on the internet?

Why We Fail to Vet Sources Online

Our research group has consistently found that young people evaluate online content while completely disregarding its source. In the voter fraud video study, one student believed the footage because they felt their own eyes could detect fraud, while another simply stated the video clearly showed wrongdoing.

This pattern is even more pronounced with AI. A teacher, sharing their story with 404 Media, described a student who, when asked how they knew information from ChatGPT was true, simply held up their phone and said, "Look, it says it right here!" Our own pilot studies show a similar trend: students place immense trust in AI chatbots, even when the bots don't cite their sources.

Beyond Fool's Gold Recognizing True Credibility

Too many internet users either fail to consider reputation or mistakenly treat platforms like Google and ChatGPT as vetted sources instead of the flawed information aggregators they are. When people do try to evaluate a source, they often rely on easily faked signals: a .org domain, official-sounding language on an "About Us" page, or simply a professional look and feel.

These signals are like fool's gold. Anyone can acquire a .org domain, including hate groups. Holocaust denial websites often claim to provide "factual information." And with AI now able to generate everything from voice clones of loved ones to hyper-realistic videos of landmarks on fire, we can no longer trust our own senses.

How to Fight Back A Practical Approach

This landscape forces us into a no-win choice between blind belief and total cynicism. Both paths erode the informed citizenship we desperately need. There is a better way: before you analyze the content, first ask who is behind it.

The very tools that create this confusion can also help us find clarity. This doesn't mean outsourcing our thinking to technology, but using it to establish a source's reputation and sharpen our own judgment.

Remember the three students who weren't fooled by the voter fraud video? They didn't use any special software. They simply opened a new tab, searched a few keywords, and found articles from credible sources like the BBC and Snopes that debunked the clip. Similarly, with some basic knowledge of how LLMs work, AI itself can help us verify social media posts by providing missing context.

The Power of Education

AI tools often include boilerplate warnings like, "ChatGPT can make mistakes. Check important info." But these warnings are useless if people don't know how to verify information—a skill that most adults lack.

The good news is that we can all learn. Studies conducted everywhere from Nebraska and California to Georgia and Texas show that even a few hours of instruction on how to gauge a source's reputation can make a significant difference. Similar positive results have been found in Canada, Germany, India, and elsewhere.

When we can no longer tell what's real, it’s easy to feel like giving up. But by embracing the ancient lesson of focusing on reputation, we can learn to navigate today’s complex world with confidence.