The US media company Buzzfeed plans to use ChatGPT to generate quizzes and personalized content. The Wall Street Journal reports this, citing an internal statement. In it, managing director Jonah Peretti explains that he believes that artificial intelligence “helps with the creative process” and improves content. It is the task of people to provide ideas, as well as to contribute “cultural capital” and “inspired specifications”. In the future, AI would help to “create, personalize and animate” the content itself.
No “terrible use”
Internally, some employees have also expressed concern, according to the report. Peretti was asked how the company intends to deal with the “unavoidable legal problems”. However, that referred to generated images, because in the past few weeks and months more and more legal questions had arisen in relation to AI image generators. This involves the often copyrighted material with which the algorithms were trained. In the beginning, Buzzfeed should only be about automatically produced texts, Peretti assured.
In a meeting, Peretti criticized the fact that digital media has so far mainly used AI to save costs and create poor-quality content. This is a “terrible use” of technology and not Buzzfeed’s approach. There are far more creative ways to use it. Buzzfeed has been known for years for content that is primarily designed to go viral on social media. It was only this week that the company’s contract with Meta became known, for Buzzfeed to produce content for Facebook & Co. for 10 million US dollars and teach others how to create content that will be widely distributed.
The portal’s AI plans are now coming to light just days after a similar attempt was canceled on news outlet CNET. Content had been published there for weeks that came from an AI and was only identified as such in a somewhat hidden way. After severe criticism and the discovery of some fundamental errors, the practice has since been discontinued. It had also been discovered that some passages contained plagiarism of other texts. CNET’s editor-in-chief has defended the move as an “experiment.” It is not known how Buzzfeed intends to identify the generated content.
(my)
