Bursting information bubbles: a necessity ?

Filter bubbles (Eli Pariser) or informational bubbles refer to the confinement of an agent when searching for information and receiving only a bundle of concordant data that isolates it from other sources of information. This is nothing new, but the personalization of information received on social networks or the algorithms that bias information on the Internet are renewing the problem. Social networks risk no longer being a platform for debate, but a collection of different spheres of opinion. The utopia of an open Internet would then come up against the total absence of contradictory dialogue and ideological confinement. Customization of information confirms (confirmation bias) the information already received, and fake news can flourish, sheltered from debate. Faced with this challenge, it is essential for economists to verify the size of these information bubbles. The first studies seem to confirm the existence of this bias, but also another, opposite effect, namely that citizens are more open to a diversity of information than they are to reading the press or television. Which populations are most exposed to the risk of bias? The second question is how to regulate these platforms. The Digital Services Act (DSA) has already laid down some markers, but has this regulation proved effective? Have the measures taken by the platforms had any effect?