The new feature could help the spread of good information, but also encourage fake news.
It's been a long time that Facebook aims to revolutionize the news sector. This time, Mark Zuckerberg's social network may have the right weapon in the drawer: it is an enhanced tool through the use of artificial intelligence and, according to the first rumors, would be able to make a short summary of the news without any intervention of a person in the flesh.
The information comes directly from a report by BuzzFeed News and speak of a particularly powerful tool that can read, obviously in a virtual way, the news and produce a summary in order to avoid the user to spend more time reading. Obviously, this is not the first step taken by the platform in the field of information, of great interest to Zuckerberg and partners. Only last year, the company decided to launch a section dedicated to news on the platform: Facebook News, this is the name, is currently available exclusively for users in the United States of America but, according to internal rumors, the project would be extended to other countries such as Germany, India and Brazil.
Facebook and AI in the news, the characteristics of the project
The project that integrates artificial intelligence is called TL;DR, which is the acronym used in many forums and stands for "Too Long; Didn't Read". Usually, it is used by users to justify a missed reading of previous posts or to introduce a short summary of what the user wrote.
In practice, the tool will provide a short summary of the news, eliminating information considered redundant or not very useful to the understanding of the focal point of the article. The operation, shown by the company only a few hours ago, would follow that of tools already present on the network, such as the AutoTLDR bot on the Reddit platform.
According to the report, other features could be added to the main functionality of TL;DR. Among these could be the voice reading of the text and the possibility to exploit a virtual assistant, able to answer questions related to the content of the news formulated by the user.
Will AI boost fake news?
In an era when time seems to be as important as money, saving a few minutes to get to the gist of the information might seem like an advantageous choice. But what are the risks involved in choosing to entrust the reading of an article to AI?
In addition to an error linked directly to the artificial intelligence algorithm, which may not clearly understand the meaning of the news or extract incorrect or problematic information if taken out of context, there remains the age-old problem of fake news. The possibility that a fake news story could spread through the web like wildfire, conveyed by a summary that does not require the analysis of a complete article (and therefore a personal understanding of the topic), could quickly generate a wave of misinformation, more serious than the time lost in reading.
In addition, already in the past researchers had been able to overcome AI controls, managing to evade verification systems and publish potentially dangerous or strongly negative content. It therefore becomes clear that, if this were to happen with TL;DR news, the possibility of finding the network flooded with short fake and harmful content increases exponentially.