Across the world—from the United States to the United Kingdom, from Europe to South Asia and Latin America—politics and media are stuck in a spiral that incentivizes divisive rhetoric, hyper-partisanship, and disinformation. The main beneficiaries of this spiral are a generation of politicians, often labeled “populist.” What connects them is not their policies, but their ability to capture attention. They use intentionally inflammatory language and controversial ideas in order to focus attention on themselves and to divide electorates into crude wars of “us” versus “them.” Sometimes they are supported by online squadrons of social media militias, as well as intensely biased publications with low editorial standards, that help push their messages. Traditional, quality media that aspire to accuracy and balance find themselves caught in a catch-22: A failure to report on these politicians will result in accusations of censorship, but challenging them risks accusations of “fake news” by the politicians themselves.

Invariably, “quality” media end up reporting on these politicians, reinforcing their agendas and language. But the way they report is also shaped by the peculiar financial incentives of the current moment. As traditional media, especially print publications, struggle to survive, they find themselves operating in an online advertising market that favors “clickable” stories—namely, the scandalous, personality-driven, polarizing content that this new breed of politicians provide. The ad-tech market in which they operate is deeply opaque. It does not distinguish between publications that have a public service-spirited mission and editorial standards on the one hand, and partisan or deliberately misleading websites and social media pages on the other. The ad-tech market is in turn powered by the algorithmic architecture of the internet and social media, which is skewed toward highly emotive, hyper-partisan material that appeals to existing confirmation biases and feeds more “shares” and “likes.” The very architecture of the internet fosters an environment where it is profitable for news organizations and individual users to take ever more extreme and polarizing positions¹—an algorithmic logic that in turn encourages the populist politicians. They, in their turn, create content that mainstream media feels they are obliged to describe…and so the spiral spins on.

This spiral can be clearly seen in action in Italy, where politicians like the Minister of Interior, Matteo Salvini, have learned to dominate the information space. Salvini’s central agenda is his attack on refugees and immigration, an agenda he follows regardless of how these issues are playing out in real life. Our research—conducted by the LSE Arena Project, Venice University, and the Italian newspaper of record, Corriere della Sera—clearly shows that even as the numbers of refugees fell by over 80 percent in 2018, the number of articles on the subject rose sharply, spiking when Salvini became minister in June. Salvini himself was responsible for this rise, both because of his comments and because he engineered events that seemed to sharpen the crisis. In August 2018, for example, Salvini blocked a ship full of refugees, the Dicciotti, from docking in Italy. When the Italian courts investigated Salvini, he took to Facebook to insist his cause was right. The Facebook post garnered millions of views. When Corriere della Sera wrote about Salvini’s Facebook post, it was the most engaged with article on the newspaper’s social media pages. At the same time, our research in the Italian Twitter-sphere shows that Salvini-supporting internet publications nevertheless accused “mainstream” media of bias, splitting the media environment into two polarized blocks—not Left and Right, but “establishment” media and “alternative” sources.

In this research project we also asked whether it was possible to report on migration in ways that that didn’t play into this spiral of scandal and polarization. Over the course of a year, Corriere produced different types of content about migration, using different media (video and multimedia, text, and infographics) and different techniques (fact-checking, human interest, and constructive news, among others). Corriere created more than a hundred pieces of content, while the LSE and Venice analyzed engagement with them on Corriere’s Facebook page. We wanted to go beyond the usual commercial metrics of “likes” and “shares,” and looked for what we called “public service spirited” metrics of engagement that could help us understand what sort of content fostered a more civil debate, smoothed polarization, and enhanced trust toward accurate content. We found, for example, that data-driven content and fact-checking actually escalated criticism of Corriere della Sera. At the same time, old-fashioned, objective reporting enhanced trust. We saw that constructive news (stories that try to introduce practical policy solutions) inspired a lot of debate, and a lot of criticism, but also fostered a more civil conversation than more strident editorial pieces. We noticed that video and multimedia content is more trusted than infographics and plain text, which was perhaps to be expected as visual evidence is more compelling than dry stats. However we were more surprised to see that human interest stories, a genre many editors turn to when they want to humanize a complex story, received vociferous criticism and some of the lowest trust—perhaps (and we can only speculate) because audiences feel they are being manipulated by highly emotional tales.

Our research is not meant to be final, and our sample size was limited. Our aim was to encourage thinking and conversation about an editorial framework that avoids the polarizing games of so-called “populist” politicians. Instead, we ask whether it is possible to create content that is at once popular,  accurate, and encouraging of constructive, thoughtful engagement. This is hard enough at the best of times, but to do so in the context of the malign spiral we have described is nearly impossible. There is little incentive to produce public-service-spirited content for a publication that has to compete in the current ad-tech market and internet architecture, especially, as in Italy, where subscription models have not caught on. In order to avoid playing into the media strategies of populist politicians, and for a new public sphere to emerge, the whole negative spiral has to be dismantled.

How can this be done? Breaking the polarization spiral will require, first and foremost, greater public oversight of the algorithms and social media models that currently encourage extremism. Such regulation is already well on its way in Europe, and public pressure is growing in the United States. It is important that any regulation is focused not on censorship and content “take-downs,” but on encouraging accurate content, high editorial standards, and providing people with a balanced diet of content instead of encasing them in “echo chambers.”

Breaking the polarization spiral will also mean reforming the ad-tech system. As a new white paper by the Global Disinformation Index elaborates, this will require both automated analysis that looks at the metadata of news domains to see whether they show telltale signs of being created in nontransparent ways, as well as a qualitative review of the content and editorial practices of news sites to determine which ones follow journalistic standards around accuracy, transparency, corrections, and reliability. Defining “quality” news sites is fraught with complications, so such decision-making needs to be done in the open, by non-profit organizations. Non-profit organizations can then apply pressure on companies and advertising brokers to direct advertising away from “junk” news sites, and to put financial incentives in place to create content that fosters a public sphere that encourages a more deliberative democracy.

But whose job is it to create such content? Most journalists justifiably argue that their mission is merely to hold politicians to account. Public service broadcasters have a clearer mission to be a crucible for a more balanced discourse, but, apart from the unique and perhaps non-replicable case of the BBC, public service broadcasters tend to be marginal players in many markets, and in countries such as Hungary and Poland they have been captured by the regime. Perhaps it is time for new actors to emerge: organizations that specialize in creating content but also constantly analyze and engage with audiences to foster a 21^st-century public information sphere, which will in turn encourage the creation of a politics that revolves less around populist personalities, disinformation, and polarization.