Most of us now get our information through AI chatbots and search engines. Even Google shows us an AI summary first before taking us to the sources from which it compiled the answers.
A new study from Yale suggests that while AI-generated answers are fast, convenient, and easy to read, they can also influence our opinions. Daniel Karell, an assistant professor of sociology at Yale University, and his team wanted to find out whether reading AI-written summaries of historical events helps people learn better than reading those written by humans.
To test this, participants were shown short summaries of historical events, some written by humans and others by AI tools like ChatGPT, and then asked what they remembered.
The result? People who read summaries written by AI consistently answered more questions correctly.
Can AI simply spread information better than humans?
Karell attributes this to the way AI presents information. “It’s like the model took Wikipedia and made it more readable,” he said. The AI summaries were smoother, clearer, and easier to remember, regardless of whether participants knew they were reading AI-generated content.
That is, even if people were told that the summary was written by AI, they still learned more from it than from the human-written version.
Should that worry you?
This is where things get interesting. In a follow-up paper published in PNAS Nexus, the same researchers found that AI summaries not only teach better, but also influence political opinions.
When the AI summary had a liberal bias, readers came away with more liberal opinions. A conservative attitude had the opposite effect. Researchers believe this is because AI not only presents facts, but also phrases them in a way that feels more logical and convincing.
AI tools are becoming the standard way people learn about history and current events. That’s not necessarily a bad thing. But knowing that the tool that shapes what you learn can also quietly shape what you think is something worth keeping in mind.
At the same time, AI hallucinations remain a major problem, and AI-generated summaries can be even more misleading to humans. A study conducted by researchers at USC’s Information Sciences Institute found that AI systems can carry out propaganda campaigns with minimal human effort.
Add to this the idea that AI can be more persuasive than humans, and it’s frightening to imagine how these tools can be used to manipulate human thought and thinking and lead us into a more fragmented world.
