Did you know They Look Real, Sound Smart, and Don’t Exist — AI Experts Are Fooling the News

 

AI generated material has become a problem even for giant media companies. They have been found to be treating AI generated characters as experts on different subjects. Media companies, like the BBC, the Sun, the Guardian, Newsweek, Medium, Fortune etc., have promoted fake characters by quoting them as reliable sources in their newspapers. But these experts do not even exist and are just the product of AI.

An investigation by the journalist Rob Waugh of the Press Gazette has revealed the infiltration of AI in mainstream media. He says that some famous newspapers published articles quoting AI generated characters as experts on various subjects. And they have been quoted again and again, indicating that the media outlets were totally unaware of their actual reality. This infiltration of AI has decreased the level of authenticity of even reliable media outlets and exposed the potential dangers of AI.  

Providing details on LinkedIn, Rob Waugh writes that one of such fake experts was an alleged psychologist Barbara Santini. Claimed to have a degree from Oxford, Santini was quoted hundreds of times in the British media as an expert psychologist, giving her opinion on different psychological problems. Rob found her to be fake when he tried to talk to her on phone and to confirm her qualifications, but she refused. She wanted to communicate only through WhatsApp: messages on WhatsApp would have allowed her to remain hidden as an AI generated character.

He gives another example of Rebecca Leigh of Academized. She also presented herself as an expert commentator on various subjects and had been quoted in the famous newspapers like Fortune and Business. Her profile on the website described her as a biochemist and science educator with a 12-year experience. But when Waugh asked the expert to prove herself to be a human, she stopped communicating. It was revealed later that she was using a fake name, picture and profile. The company also accepted that the profile had fake details. But to save itself from blame, it argued that fake name and profile details are used to keep writers anonymous.

Surprisingly, the similar photo was being used by another tech writer on another media service, LeadDev, under the profile name Sara Sparrow. Either AI is copying data of authentic writers, or it is using the same data at more than one place.

AI might be infiltrating media outlets, but the role played by humans in it should not be ignored. Qwoted and ResponseSource, two networking services connecting journalists with expert sources, have been found using fake AI experts. Similarly, humans could be behind other such usage of AI in the media, including giant media outlets, to speed up the process.

But it is not hard to detect AI in journalism. Qwoted does warn its users of a possibly AI generated content if the response of a query on the website is too fast to be from a human. Moreover, it also provides the option of Check for AI which can detect text generated by any AI system. So it is far better to establish the truth of a source than to start believing it just because data is coming from famous media outlets. This is no longer true because the authenticity of media has been compromised by AI.