frank news is dedicated to storytelling across all mediums. A space for debate, discussion, and connection between experts and a curious readership. Topics are presented monthly with content delivered daily.


Tatti Ribeiro
Clare McLaughlin
Want to share your story?
Become a contributor
Contact Us
November: TBD
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
No articles
© Frank


Your Brain on the Algorithm

by Phil Napoli
June 18, 2021

This interview with Phil Napoli, Shepley Distinguished Professor of Public Policy at Duke University, was conducted and condensed by franknews. 

frank | I want to start with how you see the collapse of local news and the internet becoming more ubiquitous in everybody's lives working in tandem? 

Phil| Obviously, one piece of this puzzle is the way digital platforms and their online ad model evolved in ways that made it really impossible for local news outlets to compete. Once a global company like Facebook or Google could offer advertising through all of the different geographic and demographic segments, there was no way for local news sites to really even attract advertisers. 

There are platforms like Facebook or Google that have accelerated this problem, but, to avoid sounding bitter, the internet was always going to democratize certain spaces and allow people to advertise more widely, probably more freely. If that was always coming, is there another business model that makes sense? Is there a world in which you see local news doing well?

I think that we've been in about a decade and a half of searching for new business models. Certainly, there has been a lot of foundation work to try to support experimentation.

Personally, I think we are at a point where we have to embrace a true public support model. 

We have to recognize that the model we have to embrace, that we haven't in this country at all, especially compared to other developed countries, is a true public service, public support model. We are at the bottom compared to other developed nations in terms of what we spend per capita, in terms of public support for journalism, for media. We have NPR and, you know, that's primarily member-supported. Our public service media system is minuscule compared to other countries. I really think, whether it's at the state level or the federal level, we need to rethink how many tax dollars go to supporting journalism. This creates a tension that we are justifiably concerned about, but there are, I think, plenty of proposals out there that could make that work in a way that would create the necessary firewall between government and journalism.

Do you have any examples?

One proposal that has been around for about a decade, and is starting to resurface again, is for each individual on your tax returns to pick which outlets you wanted to support. Our primary concern would be if funds were allocated in a way that expressed the political interests of the government at that moment in time. But if you can imagine every individual news consumer being granted the autonomy to determine where their contribution goes, that alleviates some concern. That is one avenue that is possible. 

It seems like the is money there. There isn't really a competitor to PBS. It's not a dense marketplace. Is there resistance from donors?

Involvement from the philanthropic sector is growing. There are more foundations turning their attention to this space than I’ve ever seen before. The irony, I think, is that many in the news business are resistant to a more robust and wide-ranging public service or foundation support model out of concerns about infringements on their editorial freedom. 

Of course, advertisers have always had that capacity as well. Advertising dollars often come with strings attached or with efforts to influence the news coverage. I mean, if you look at the history of the traditional print newspaper over a few decades, news sections always respond to particular advertiser needs: we need a travel section, we need an auto section. Advertising demand drove a lot of this.

I mean, Ed Murrow was complaining about this in his prime. His questions about sustainability when you need growth constantly were valid then and obviously now. 

You talk about algorithmic news – can you explain what you mean?

It can mean a few different things. On the production side, algorithmic systems can play a role in actual content generation. Processes of actually automating news production by taking various databases and feeding them into a software package that's capable of spitting out text in the form of a news story exist. Some of the new local news networks sprouting up are primarily producing journalism that falls into that category. 

It's truly automated news. The same story is being generated across hundreds of outlets.

On the consumption side, it refers to how we engage with news and how we are allowing algorithmic recommendation systems to dictate to a certain extent the stories that we consume.

To what extent are you reliant on what shows up in your Facebook feed? To what extent are you reliant on the recommendation system that Youtube provides? I've seen studies that showed 70% of what people watch on YouTube is determined by the YouTube recommendation engine

Where does public policy come in?

In terms of the most regulated sector, you have broadcasting. Broadcast TV until the late eighties, had to operate under something called the fairness doctrine – in which the government-mandated balance in opposing viewpoints. That's gone away. Broadcasters were also required to provide minimum amounts of local news at one point in time. That has gone away. That's the most regulated space — some regulations remain today on children's programming and political advertising and that sort of thing.

Then we can think about the online space. Section 230 applies to platforms that engage in the hosting of third-party content. Section 230 basically says, you are entitled to make editorial decisions about the content that you host, but you are not liable for any of that content. You cannot sue Facebook if something false about you is shared on Facebook. You can certainly sue the originator of that falsity. So we have created this incredible incentive for hosting third-party content to be the default business model online; there's no civil liability that comes from hosting that content.

Is there any regulatory framework that exists there now?

There are some ideas being kicked around about modifying or eliminating Section 230. 

But what is it that we want to achieve? Do we want these platforms to be more aggressive in policing falsity, which some folks want, while others want the platforms to act like neutral hosts? 

If you were to create less immunity from liability for these platforms, you hope that they would in fact then be more aggressive about what they police. For some folks that sounds great. The only thing about that that bothers me is that what we're usually talking about is immunity from civil liability for libel and defamation and things of that nature.

The whole enforcement mechanism would come down to who has the resources and the wherewithal to sue. 

And if you think about all the kinds of falsity that can be disseminated, for example, say someone spreads the falsity that masking does not prevent the spread of the coronavirus. Okay, who have you libeled there? Nobody. We're kidding ourselves if we think tweaking Section 230 is a solution to the range of disinformation and misinformation problems that we have in that space. 

We hear Zuckerberg saying, ‘we don't want private companies to have so much control in the decision-making process about how to balance social equities and democracy’. Do you feel he is saying that in earnest? Is he saying to the government, do what you have to do?

That's a great question. I mean, I think they never wanted to be in this space of having to make decisions about what to host and what not to host. It reflects the fact that the user base for these platforms is so different and the uses of the platform are so different from when they originated. They evolved in ways that I don't think the creators ever imagined. They have always been so far behind in terms of thinking of themselves as genuine curators of content. They've thought of themselves as “providers of tools.” 

They certainly would have loved a regulatory environment where everyone is on their own to protect themselves, but they've come to realize that the political environment is not supportive of that. At this point, I think they would be interested in delegating that decision-making and all the political blowback that comes with it. 

The tying together of news as content is deeply disturbing because they are not the same thing. Is there a way for news to be separated from other kinds of content?

Once everything became part of an undifferentiated news feed, that was hugely damaging for our notions of informed citizens.

I talk a lot in my book about the good old-fashioned distinction between push and pull media, and how that relates to the distinction between misinformed people and uninformed people. Being informed used to require an active effort. I was going to consciously make a decision to watch the news or read a newspaper or visit a news site. When the news became something that just got pushed into your social media scroll, that was ironically less conducive, rather than more conducive to an informed citizenry because platforms were not doing anything meaningful to distinguish between news and journalism and other types of content. And when some, like Snapchat, have tried to separate news from other content categories, users didn’t embrace it.

To this day, any of us can start an account on Facebook and call ourselves a news organization. There's so little rigor and scrutiny about what constitutes news. We sort of over democratized the notion of what a journalist is, which has contributed to news consumers not being very discriminating about what constitutes a legitimate news organization or source. 

I think it encourages individuals to stand out – as personalities, as celebrities. To control their own audience in some ways. 

The platform creates these opportunities for individuals to create brands for themselves that absolutely did not exist before. We joke about how disconcerting it is just how many undergraduate students when asked what their career aspirations are, say that they want to be an influencer. That's a career path now, and that's starting to intersect with the realm of journalism, and I don't think that it is healthy. 

Do you think it’s possible to change our habits of consumption?

Well, we have to think about this as generational change: can we alter the habits of the next generation of news consumers? This goes to how we educate kids in elementary school about how you go about informing yourself. Other countries are way ahead of us on the basics of news literacy and digital literacy and training people to be critical news consumers, which you have to be in this media environment. I think that is a big part of what we need to figure out going forward.

Do you think local journalism needs to be non-profit in order to succeed? 

I've seen some models that I find interesting. In New Jersey, the state had some broadcast licenses that they auctioned off, and part of the proceeds from that auction went to a local news fund that would help support local news nonprofits across the state. 

I'm leaning that way.? If you think about who is making money, the New York Times, the Washington Post. One of the key things that is vital to success is economies of scale. Can we distribute our costs across as large an audience as possible? Can we sell and resell our content to more and more audiences? And there are just no economies of scale in local journalism. Every municipality needs its own news organization, and the bottom line is, if you're writing about the school board in your town, it is not interesting to anybody in any other town. That is part of the problem. When we analyze local news outlets, a big part of what they do is just regurgitate national news.

If you can’t take advantage of economies of scale, that makes a commercial model really difficult. 

The big caveat here is if we can make people start to pay more directly for local journalism. During the pandemic, local news outlets saw an uptick in subscribers. There are certain things that seem to be able to drive it. But again, maybe this is something we need to train young people to think differently about – the way generations were retrained to pay for music instead of downloading it off of a file-sharing platform.

Our site is both free and ad-free. In some ways, I think I’m doing a disservice to the mentality that this information is provided to you for free. On the other hand, there is this feeling that it is for the public and no one should be excluded. When I see information pay-walled behind a major newspaper and that information is helpful to the public, it feels wrong. 

So basically, am I problematic?

With a stronger consumer support model, you start having to worry about these gaps between who can and who cannot afford to be informed.

That's why I come back to the ideal model, being a public service model of journalism that really truly filters down to the local level, that is accessible to everybody, and, at the same time, is insulated to some extent from the negative influences the commercial model certainly has on content. 

Last semester I had a student who was working at a local news organization. There was a security camera outside their building, and it kept catching this squirrel. Every day they would put up on their website a video of this squirrel. The whole town loved to watch it. They were like, I can't believe that this is where all our revenue is coming from. All of our clicks are coming from this stupid squirrel video every day, but they could not bring themselves to stop showcasing the squirrel on the roof of their building. 

That feels like a great analogy for Trump. CNN could not stop showing us the squirrel. 

I've heard CNN's primetime ratings are down something like 70%. The worst part is it creates this incredible incentive system to try to figure out how to resuscitate the ratings. So people are saying, the only reason we're hearing as much as we're hearing about Marjorie Taylor Green is that she helps generate some of the outrage and viewership and clicks that were lost when Trump left office.

It can have a very perverted effect on our politics.

I tell people that if you are watching any of the cable news networks in prime time, you are not watching the news. These are preachers. Whether it's CNN, MSNBC, CNBC, or Fox News, they all do it. It goes to the cost issue. They realize that what's really cheap to do relative to doing actual reporting, is to put a talking head up on the screen and let them spout their opinions for hours on end. And, it turns out, that unfortunately, this is something many viewers prefer to actual reporting.

And they all speak in the same voice.

I was interviewed a while back by a typical YouTube guy in his basement discussing the news. When he interviewed me, he said, I've actually just been purchased by CNN. And they specifically don't want me to change anything. They want me to look like a YouTuber in my basement with a homemade set. This guy was a CNN employee. I was just fascinated by the fact that this is how they were having to start to think about how to reach younger news consumers. It really doesn't bear much of a resemblance to what we traditionally think of as bonafide news reporting.

Can we escape algorithm-driven news? 

The bottom line is those kinds of data-driven decision-making systems that Netflix uses, hell, that a lot of news outlets use to decide what to feature or what not to feature, inherently can't innovate because it's all historical data. Data tells us people like this, so let's give them more of what the data tells us. They've already demonstrated that they like it. 

How do you discover a new show that you would not have known you would have liked? Schitt's Creek is a perfect example. 

When it comes to news, that's when we get to these concerns about reinforcing existing biases and political beliefs, and polarized viewpoints, which is what the data show people want the most. We give them more of it, and it ends up having these reinforcing and polarizing effects. 

Not to call consumers children, but where are the adults in here being like, who cares what they want, this isn't about what people want, this is about what you need to know to vote properly. 

Again, that's where the divorcing of the traditional commercial model becomes particularly important.

The bottom line is that's easier information to get now than it ever was before. It's possible for a news organization to know so much more about people's demonstrated preferences. It's so fascinating to think about how in decades past, newspapers had no clue which stories were being read and which stories were not being read. Granular data on the performance of individual stories and the performance of individual writers was not available. You knew how many copies you sold, and that was it. There is a case to be made that a degree of uncertainty about consumer demand as it relates to the news has some benefits.