The Unreliability of ChatGPT’s Citations: A Concern for Publishers

A recent study conducted by the Tow Center for Digital Journalism has shed light on the concerning issue of ChatGPT’s citations. The study found that the AI chatbot’s citations are often inaccurate, incomplete, or misleading, posing a significant risk to publishers’ reputation and credibility.

The study analyzed 200 quotes from 20 randomly selected publishers, including The New York Times, The Washington Post, and The Financial Times. The researchers found that ChatGPT’s citations were frequently incorrect, with the bot often inventing or misrepresenting information. This lack of transparency and accuracy raises serious concerns about the reliability of ChatGPT’s citations.

One of the most concerning findings was that ChatGPT’s citations were not always reliable, even for publishers that had inked deals with OpenAI. This suggests that allowing OpenAI’s crawlers to access their content does not guarantee accurate citations. Furthermore, the study found that ChatGPT’s citations were often inconsistent, with the bot returning different answers to the same query each time.

The study’s findings have significant implications for publishers, who risk having their content misrepresented or attributed incorrectly. This can lead to reputational damage and loss of credibility, as well as commercial risks associated with readers being pointed elsewhere.

The researchers argue that OpenAI’s technology is treating journalism as “decontextualized content,” with little regard for the circumstances of its original production. This lack of understanding and respect for the context and provenance of journalistic content is a significant concern for publishers.

In response to the study’s findings, OpenAI stated that it supports publishers and creators by helping users discover quality content through summaries, quotes, clear links, and attribution. However, the company’s response does not address the fundamental concerns raised by the study about the accuracy and reliability of ChatGPT’s citations.

In conclusion, the study’s findings highlight the need for greater transparency and accountability in the use of AI-generated citations. Publishers must be aware of the risks associated with ChatGPT’s citations and take steps to protect their reputation and credibility.