NewsBeat

20 NewsBeat October 2018 (Continued from previous page) interest in countering harassment and blocking some kinds of disinformation. What's more disturbing is the bad behaviors that may simply be part and parcel of the new media environment. Even when all participants in a system are acting in good faith, ugly behaviors still emerge. Whether it's due to algorithms (as Eli Pariser believes) or simply due to homophily and the cognitive dangers from birds of a feather flocking together (as Cass Sunstein and I, separately, have argued), echo chambers and filter bubbles seem to form online, surrounding us with content that tends to polarize us. Some of these isolated conversations seem to gravitate towards the paranoid, leading to the spread of conspiracy theories and other unreliable content. Maintaining these communities requires some degree of control, which leads to complaints about the biases of algorithmic filters and control asserted by moderators. And even with control, mobs form and punish what they perceive to be bad behavior. What's toughest about this fourth bucket of problems, the known bugs, is that they seem to appear even when everyone's behaving well within a media ecosystem. Much as I dislike what Breitbart has to say, they're producing content for audiences that want to hear it. That center-right audiences gravitate — or are algorithmically nudged — to the far right doesn't appear to be the result of anyone's sinister plan. Worse, it's simply an emergent property of a complex system, which makes it harder to determine who's to blame and who's responsible for fixing it. Taken together, even this minimal model of the problems raised for media and democracy seems pretty daunting. But understanding the problems as buckets of interrelated challenges that can be solved without unraveling the entire Gordian knot makes the situation more tractable. Financing the News The problem of financing the news is hardly new, and it's unlikely that my friends on the Knight Commission are going to come up with something radically better than what the collective wisdom of online and offline publishers, journalism and business schools have in the past decade. I think it's unlikely that there's a solution to this set of problems without insisting on the necessity of news for participation in a democracy, which opens options too seldom considered in our country, where market solutions are always preferred. News may be too important to leave to the whims of the market. Whether that means financing public media providers that are positioned to give us the essential facts of the world we live in, or building a robust set of membership models that allow a small set of subscribers to make critical information available to wider audiences, there are solutions to the problem of financing the news. Unfortunately, these solutions are rarely popular because they're expensive and hard to sell in the U.S., where they run counter to conventional thinking about markets and speech. Harder is the problem of getting people to pay attention to the high quality news that's already produced. This is the sort of problem that will require news publishers to work closely with discovery engines, and might open intriguing possibilities, like social media systems supported in part by public media funds, designed to help people mix opinions and perspectives from friends with a backbone of information from trusted sources. Unlike the problem of financing media, many of the concerns about the addictive nature of social media are new. There's a need for psychological research to figure out how deep public concerns should be, given a a long history of public overreaction to the dangers of new media technologies. It's likely that the field also needs strong consumer protection groups that can demand that technology platforms take our well-being seriously. And if there's evidence that companies are building these tools with an addiction model in mind, regulating how these companies operate seems like a very reasonable step. While bad actors, the third set of problems outlined here, have provoked the most dialog since the 2016 election, they are the easiest problem to tackle, as platforms and consumers have aligned interests. Bots aren't good for Twitter — they inflate user numbers in the short term, but lead investors to worry about whether the site's users are real — nor are they good for users of the platform. Companies like Facebook, Google and Twitter will need to work closely with scientists in academia, opening up data sets and collaborating closely so these problems can be tackled — Twitter's move towards collaboration on healthy media ecosystems is a step in the right direction, and if done well, the expertise of academics and activists who are already fighting harassment online can be incorporated into key systems. Platforms should be able to tackle bots with machine learning systems, as their signatures are highly recognizable. And emerging movements towards transparency in advertising are steps in the right direction in combating dark ads and other forms of paid influence. What's most needed here is a way for platforms and their critics to share data, tools and methods in a way that's less threatening to both sides… a problem that has become much harder with the fallout of the Cambridge Analytica scandal. Right now, academics and activists can't do their best work on these platforms without data they need, and platforms are understandably afraid of being excoriated in

• Cover