BuzzFeed’s quizzes will soon be written by artificial intelligence, which will also help the media to “personalize” other content for Internet users. The group is the first to assume a decision of such magnitude, which raises many questions about what Internet users consume online content.

It is the first major international media to publicly endorse the use of the ChatGPT conversational tool to create content on the internet. The infotainment group BuzzFeed will have some of its articles written with less journalistic value, Jonah Peretti, its CEO, announced in an internal memo. obtained by the Wall Street Journal on January 26, 2023.

This will involve both creating quizzes and customizing certain content for their audience, the WSJ pointed out. Jonah Peretti explained that he believed that artificial intelligence would help the creative process and improve content, while humans would remain responsible for ” cultural references » and could write « inspirational orders “.

ChatGPT (for “Generative Pre-trained Transformer”) actually works thanks to commands called “ prompt “, in English. In this case, these are orders that are written in order to direct the chatbot in the requested request. ChatGPT, which everyone discovered at the end of 2021, is an iteration of what can be done with GPT-3, the AI ​​developed by OpenAI, which makes its functionalities accessible to as many people as possible. Professionals and businesses can also use the OpenAI API to develop their custom tools.

The CEO of BuzzFeed believes that within 15 years, AI will “ccreate, personalize and animate content on your own “. Today, ChatGPT may be very functional, but it is only “fed” by what already existed on the web (until 2021). It is for this reason that some AI specialists want to put the prowess of ChatGPT into perspective.

However, the multinational should not impose these changes on journalists, who deal with serious or topical subjects. That did not prevent some from worrying publicly about this news.

This announcement is added to another: a partnership between the media group and Meta, which pays BuzzFeed to help attract more creators to Facebook and Instagram. The company’s stock (listed on the stock exchange since last year) has exploded by +119%.

Buzzfeed’s action on January 26, 2023 // Source: cnbc

BuzzFeed throws a stone into the pond

BuzzFeed became known in the 2000s for its many lively quizzes, before adding a journalistic brick to its activities, hiring editor Ben Smith in 2011. Nevertheless, the majority of its audience has always been dependent on lighter content, such as bulleted articles or lists.

By assuming to ask an artificial intelligence to create personalized quizzes, BuzzFeed is causing shock waves in the media world, in which the company has always navigated like a sort of UFO, capable of releasing quality journalistic scoops while flooding the web with the most casual quizzes.

The debate is vast. On the face of it, no one wants all the articles on the web to come from bots — CNET’s attempt recently caused an uproar because the tech site failed to clearly indicate that some articles were written using GPT- 3.

Nevertheless, many contents with very low added value already abound online, without it being indicated that they are vulgar translations or copy-pasted. We could thus see, in the public announcement of BuzzFeed, a certain transparency on their methods.

Similarly, some articles, which require little application or research work, are akin to assembly line work that is not very stimulating intellectually: automating some of them could save time for editors, who would focus on other missions with better added value.

Jonah Peretti, CEO of BuzzFeed // Source: Flickr/CC/Fortune Brainstorm TECH
Jonah Peretti, CEO of BuzzFeed // Source: Flickr/CC/Fortune Brainstorm TECH

In addition, the group has long been known to pay low, if any, some contributors. In 2019, Matthew Perpetua, Head of Quiz, had revealed after he was fired, that many of BuzzFeed’s survey writers were unpaid. The example of a student in particular had been much commented: Rachel McMahon, was the second author who reported the largest audiences of the site throughout the world, while she was not paid. His quizzes reportedly grossed over $3 million to the group.

Progress or regression?

Using a robot rather than exploiting a human could be seen as a sign of economic and social progress. Nevertheless, several nuances prevent giving in to such a Manichean comparison.

  • On the one hand, we are not talking here about assembly line work, but about production that calls for intellect and creativity. We are thus approaching the debates that are inflaming the art world, for example around the prowess of Midjourney, which worries artists a lot about the notions of copyright and original creation.
  • On the other hand, remember that GPT-3 does not technically create original content. It has been fed by billions of online and offline content, and only offers texts from already existing aggregates. To put it more prosaically, after exploiting Rachel McMahon to save money, BuzzFeed will be able to generate new quizzes (even more quickly and easily) from the hundreds of quizzes she invented for free, years ago.
  • If we project ourselves further into the future, and imagining the potential abuses of such use, we can consider that the content of BuzzFeed would end up spinning in a vacuum: robotic quizzes inspired by other quizzes robotized… and end up creating a mass of non-original content that folds in on itself.

Will you assume to enjoy a quiz made by a robot?

It would be possible to imagine a positive future, where humans and robots cohabit, and in which original creation would thus be all the more valued. However, this assumption only works if we consider that all the actors will be honest and transparent. However, knowing the many protagonists of the web, it is unlikely that this sincerity is omnipresent in the production chain.

BuzzFeed’s decision, however, has the merit of putting the foot in the dish regarding the low quality of much content consumed online. She cynically endorses the fact that a large part of Internet users absorbs useless, merely entertaining productions to occupy their attention.

From now on, Internet users will no longer be able to escape the question: do I agree to answer a humorous quiz that has not been written by a human? What would the opposite bring me? What am I really looking for when reading these productions? Why would I want content with such low added value to come from the intellect of one of my peers? The questions are just beginning.


Subscribe to Numerama on Google News to not miss any news!

Understand everything about experimenting with OpenAI, ChatGPT

California18

Welcome to California18, your number one source for Breaking News from the World. We’re dedicated to giving you the very best of News.

Leave a Reply