During an episode of The Joe Rogan Experience with guest Andrew Doyle, the conversation turned to artificial intelligence and censorship when Doyle shared a troubling personal experience with AI content filtering.
Doyle explained he was using AI as a search engine to find an old article he’d referenced in his book. The article concerned a case in the UK where a man who had assaulted a 13-year-old girl received no jail time after the judge cited his religious upbringing as making him “s*xually naive.”
“It came up on ChatGPT. And then it deleted,” Doyle said. When he questioned why, the AI responded that it might violate terms of service. “I said, ‘Well, how could it? This is an article that’s in the public domain.’ So, it gave me the information again, deleted it again.”
The pattern repeated multiple times. “I said, ‘You keep deleting this. Stop it.’ And said, ‘I definitely won’t delete it.’ Then it did the same again,” Doyle recounted.
He explained the AI’s reasoning: “What it’s doing is it’s saying because this is a news story that could be deemed anti-immigrant or this is a news story that is politically sensitive, I’m not going to let you see it.”
Rogan suggested trying the search on Perplexity instead, noting his doubt that the alternative platform would censor the same way. With this, he accidentally put his AI sponsor on spot.
“Oh, I wonder if you could do it in America. Let’s find out. Well, let’s try Perplexity,” Rogan said.
When Jamie admitted he wasn’t sure which specific article Doyle was referring to, Rogan suggested a simple workaround.
“Well, why don’t you just ask the question he asked 10 years ago?” he said, encouraging them to recreate the original prompt rather than getting stuck on the missing reference.
As they revisited the incident, Rogan seemed genuinely baffled by the AI’s behavior, particularly the way it briefly produced an answer only to retract it seconds later.
He asked: “But it showed it to you and then it pulled it back, which is crazy,” he said. “Like, how does it not know?”
However, the conversation quickly drifted away before Rogan ever followed through on testing Perplexity. Doyle began contrasting the experience with Grok, Elon Musk’s AI platform, which he claimed was more willing to provide information without the same kind of filtering.
Rogan didn’t press the issue much further, and the moment passed almost as quickly as it surfaced.
In another episode, Rogan turned to Perplexity AI on-air to fact-check musician John Mellencamp after he made revisionist claims about the Civil War, including falsely suggesting Abraham Lincoln owned slaves. The AI quickly returned the historical consensus, directly contradicting Mellencamp’s assertions.
