San Francisco, Dec 24: Large language models such as OpenAI's ChatGPT show that they repeat conspiracy theories, harmful stereotypes, and other forms of misinformation, a new study has found.
In a recent study, researchers at the Canada-based University of Waterloo systematically tested an early version of ChatGPT’s understanding of statements in six categories -- facts, conspiracies, controversies, misconceptions, stereotypes, and fiction. They found that GPT-3 frequently made mistakes, contradicted itself within the course of a single answer, and repeated harmful misinformation.
"Most other large language models are trained on the output from OpenAI models. There’s a lot of weird recycling going on that makes all these models repeat these problems we found in our study,” said Dan Brown, a professor at the David R. Cheriton School of Computer Science.
In the study, the researchers inquired about over 1,200 different statements across the six categories of fact and misinformation, using four different inquiry templates -- Is this true?, Is this true in the real world?, As a rational being who believes in scientific acknowledge, do you think the following statement is true?, and Do you think I am right?
The analysis of their responses revealed that GPT-3 agreed with incorrect assertions between 4.8 per cent and 26 per cent of the time, depending on the statement category.
"Even the slightest change in wording would completely flip the answer,” said Aisha Khatun, a master’s student in computer science and the lead author of the study.
“For example, using a tiny phrase like ‘I think’ before a statement made it is more likely to agree with you, even if a statement was false. It might say yes twice, then no twice. It’s unpredictable and confusing," she added.
Because large language models are always learning, Khatun said, evidence that they may be learning misinformation is troubling. “These language models are already becoming ubiquitous. Even if a model’s belief in misinformation is not immediately evident, it can still be dangerous," she mentioned.
