ChatGPT is still among the best-regarded chatbots in the industry despite several competitors, but it's still far from perfect. With that said, you might not want to put too much trust in the AI service when it comes to dispensing important information. 

ChatGPT is Still 'Hallucinating' Links

Nieman Lab ran an experiment to see how OpenAI's chatbot would fare in providing the right links as sources, and it turns out that ChatGPT is still not up to the task. Instead of citing the correct links, the AI tool instead makes up URLs.

This has been a common error in the chatbot, wherein it would sometimes make up information mixed with facts, making it easily believable. This is what the AI industry refers to as hallucinating, as mentioned in Engadget. 

In the experiment, Nieman Labs asked the chatbot to provide links to exclusive stories from the publishers OpenAI had made deals with such as Vox, The Atlantic, The Wall Street Journal, Associated Press, The Financial Times, and more.

Upon clicking on the links, it simply led to 404 error pages since they did not exist and the chatbot simply made them up. It means that ChatGPT cannot provide proper URLs even from the news outlets that it already has access to.

A company spokesperson said that OpenAI was still building "an experience that blends conversational capabilities with their latest news content, ensuring proper attribution and linking to source material," although it's still not available for ChatGPT.

The ChatGPT maker is already paying millions to the mentioned news publishers for access to their archives, so it's starting to seem like a waste since the company has yet to take advantage of this perk. However, if the allegations are to be believed, the new deals weren't even needed in the first place.

Read Also: OpenAI Introduces ChatGPT Edu, Nonprofits for Cheaper Price

OpenAI Has Been Accused of Using Online Content Anyway

OpenAI has seen a lot of copyright complaints and lawsuits in the past few months, claiming that the AI company is using copyrighted works to train its AI models without permission. Just a month ago, eight US news publishers took OpenAI to a New York federal court.

Taking issue with both ChatGPT and Copilot, Microsoft was also included in the lawsuit. All of the news outlets such as the New York Daily News, the Chicago Tribune, the Orlando Sentinel, the Sun Sentinel in Florida, and more operate under the hedge fund Alden Global Capital.

An OpenAI spokesperson responded saying that the company takes great care in its products and designs to support news organizations, as mentioned in CNBC. Microsoft, on the other hand, had no comments on the matter.

With that said, the news organizations might as well get paid for having their data used to train AI models. The latest to strike a partnership with OpenAI is Time Magazine, which is a multi-year content deal to bring Time's journalism to OPenAI's products. 

"Through this collaboration, OpenAI will gain access to current and historic content from TIME's extensive archives from the last 101 years to enhance its products and display in response to user inquiries," TIME stated. 

Related: How to Limit OpenAI from Collecting Personal Data on ChatGPT?