ChatGPT is generating inaccurate URLs for news articles from its partner organisations, according to an investigation by the Nieman Journalism Lab.

When asked about specific investigations or reports by Nieman researchers, ChatGPT sometimes provided detailed summaries along with fabricated URLs that appeared legitimate but did not actually exist despite licensing deals promising proper attribution and linking.

Affected publications include the Associated Press, Wall Street Journal, Financial Times, The Times (UK), Le Monde, El País, The Atlantic, The Verge, Vox and Politico.

The finding highlighted concerns about ChatGPT's accuracy and reliability, notably its tendency to 'hallucinate' information and data, including citations and content links, thereby contributing to a poor user experience and the potential generation of misinformation and disinformation.

OpenAI acknowledged the issue pointed out by Nieman and stated they are working on addressing it.