In beta stage and available to a few users, it aims to be “like TikTok but for the news”
Providing personalized, reliable, and truthful information. This is what Instagram founders Kevin Systrom and Mike Krieger aim to achieve with their latest app, Artifact. Artifact, which plays with the words’ artificial’, ‘article’ and ‘fact’, is currently available in beta version for the exclusive use of a select few users as experimenters. It is a kind of ‘TikTok for news’, a social media to scroll through like the Chinese app to offer the user a constantly updated feed based on their previous activity on the app.
Social media will also allow users to comment on news, but only by sharing references to the relevant site or newspaper on their profile. In short, creating original posts with a previously published starting point will only be possible. A way of eliminating the risk of fake news from the outset. But Artifact does not present itself as a simple return to the text in the golden age of audiovisual content because, at the basis of the whole algorithm that will take care of the composition of the customized feed, there will be the Transformer system, created by Google in 2017, to process texts in a manner indistinguishable from that of human beings.
Artifact represents an innovation in the field of traditional text-based social media. If Facebook and Twitter propose new content to their users based on the pages they follow, the new app, like the way TikTok works, will display content independently of the pages users already follow, based at most on the topics that are most attractive to the user. But this work on the ‘freedom’ granted to the algorithm to propose new news to the user has nothing to do with the unreliability of the sources from which they will be drawn and disseminated, which will be scrupulously selected and checked in real-time by artificial intelligence. Another absolute novelty is the adoption of the criterion of the time taken to read a post as a benchmark for the algorithm in place of the less indicative number of clicks.
