Episode description:

Introduction

It’s pretty much a cliché to say that Americans live in two separate political realities. We shout at each other from our separate bunkers, unable to agree even on basic facts. One coping strategy—and maybe the default, given human nature—is to hunker down with our own tribe and demonize the other. But there are more constructive ways, and today’s guest makes a living examining those alternatives. He’s Jonathan Stray, senior scientist at University of California Berkeley’s Center for Human-Compatible AI and the creator of the excellent newsletter Better Conflict Bulletin.

Jonathan studies the growing field of peace-building—that is, helping people from different factions work together. He’s involved in research about AI primed to unite people rather than divide them. Some things we learn from this conversation: About a journalistic standard called multi-partiality, a more attainable goal than impartiality. About how you might construct an algorithm that prioritizes reliable news over popular news. And why, despite the state of discord right now, there are reasons to be optimistic.

Faithful listeners might recognize that we recorded this podcast about a year ago, but its relevance has only increased. We’re re-posting this the day after the most divisive state of the union address Eric has ever heard. This administration will not last forever, and we will as a country need to find our way back to working together. And we can really use some of Jonathan’s optimism about our ability to do that.

 

Summary

In this conversation, Jonathan Stray, Senior Scientist at the UC Berkeley Center for Human-Compatible AI, explains to Eric Schurenberg the intersection of AI, media, and conflict, emphasizing the challenges of objectivity in journalism and the need for a new approach to reporting that embraces complexity and ‘multipartiality’. He explores the role of AI in shaping social media narratives and the potential for algorithms to foster better understanding in political discourse. Stray also highlights reasons for hope in addressing political polarization and the importance of bridging divides through constructive dialogue.

 

Transcript

Eric Schurenberg (00:01.273)
Jonathan Stray, welcome to In Reality.

Jonathan Stray (00:04.162)
Hey, nice to be here.

Eric Schurenberg (00:05.765)
It’s great to have you. You’re the head of the Center for Human Compatible AI at Berkeley, University of California at Berkeley. You’re creator of the excellent Better Conflict, a bulletin newsletter and the podcast of the same name. I don’t see a direct line between human compatible AI and a newsletter that’s focused on, and an organization’s focused on making disagreement in America better. Walk us through that.

Jonathan Stray (00:35.49)
Well, first of all, I’m not the head of the Center for Human and Battle AI. I’m a senior scientist there. The head is Stuart Russell, who is an old time star in the AI field. Do you want to redo that intro? Yeah. All right.

Eric Schurenberg (00:46.231)
okay. Hey, Tom, let me redo that intro. I might as well not start out with a stumble. All right, all right. And once again with feeling. Jonathan Stray, welcome to In Reality.

Jonathan Stray (00:53.495)
Yeah, yeah.

Jonathan Stray (01:00.632)
Alright.

Jonathan Stray (01:04.664)
Good to be here.

Eric Schurenberg (01:05.765)
It is great to have you. You’re a senior scientist at the Center for Human-Compatible AI at University of California at Berkeley and the creator of the Better Conflict Bulletin newsletter and the podcast of the same name, both of which are excellent. It doesn’t seem to me though that there’s necessarily a direct line between human-compatible AI and an organization and content that’s dedicated to the idea that disagreement in the US can be better, better managed, better handled. Can you walk us through that line?

Jonathan Stray (01:39.79)
Sure, so I study three things. I study AI, media, and conflict, and how those things collide. And so in my day job, I do things like run experiments on different AI-driven social media algorithms to see if we can reduce polarization. I’m currently working on a test to understand what would a politically neutral chat bot or AI even mean.

But of course, conflict is a big element of this. And when I say conflict, mean, intergroup conflict. mean, political conflict, people fighting over something, everything from online flame wars to civil wars to interstate wars. And, you know, in my journey as a conflict researcher, I began to realize that if I didn’t pay attention to the specifics of a real conflict, at least one real conflict, I was going to sort of drift off into theory.

And so I started getting seriously into the American culture war. And that was some years ago and, you know, things have only gotten, let’s say more interesting since then.

Eric Schurenberg (02:46.559)
You also have a background in media, having taught at Columbia Journalism School, worked at the Associated Press. How did that shape your particular journey?

Jonathan Stray (02:56.3)
Well, media is super important, right? We experience very little about the world firsthand, which means mostly it comes from media. So it’s gotta have an effect in shaping our perceptions and attitudes and understanding of one another. And so at this point, media is becoming increasingly technological. Our media diets are mediated by machines in a big way. So it’s starting to… to intersect with the tech world in a serious way. And I think neither the media world nor the technology world are particularly happy about that. It’s a very awkward marriage.

Eric Schurenberg (03:35.461)
Well, let’s dive into that marriage and let’s start with, if you don’t mind, the big O, objectivity. It is often held up among professional journalists and also among consumers of the work of professional journalists as this brass ring, holy grail. Is it actually attainable as a concept? Is it kind of an ideal that is actually unrealistic?

Jonathan Stray (04:16.556)
Right, well, you know… news consumers and journalists, should say, want to want journalism to be something like the truth, right? There’s some sort of a, we would say in philosophy, versilimitude that they’re looking for. And objectivity was a word that started to be used in the early 20th century for basically two reasons. One was they wanted journalism to be more scientific, right? They wanted to borrow some of the

Authority and methods of science and say, know, this is the truth. This is the real world that we’re telling you What the word really means is it it exists whether no matter what humans think right? As Philip K Dick said Reality is that which when you ignore it doesn’t go away So that was part of the idea The other part of the idea was that in the 19th century American journalism was very partisan it was mostly funded by political parties and news organizations began to realize that that meant losing half their audience. So if they could somehow appeal to everybody, they could have a broader audience and then, you know, sell more papers, sell more ads. So it sort of took off at the time. And I think there’s some some core ideas that still make sense about it, but some ideas that don’t make sense. The idea that there is always some truth that is independent of who’s doing the reporting…

…doesn’t really hold up as well as it used to because audiences want something other than just the facts. They want to know, well, what does this mean? You know, what does this mean for me? What does this mean for my country? And that’s an interpretation. And you can’t have an interpretation without coming from somewhere. And of course, the all of the journalists’ background influences that somewhere. So as an idea, doesn’t really work.

The problem that we have is we don’t really have a good explanation of what should replace it because we still need some sort of idea of, this is not just your opinion,

Eric Schurenberg (06:21.861)
Hmm. I mean, part of the journalism training is to try to take yourself out of the picture, theoretically, to get a number of sources, including if you’re careful about your job, to get conflicting sources so that you have a rounded idea. And also, to be fair, journalism isn’t the only profession that tries to practice this. You could say that this is something that you are bound to do if you work in the intelligence community or in science or medicine. Is there something particular about journalism that makes it kind of unattainable on that platform?

Jonathan Stray (07:06.414)
Well, I mean, think what journalism studies, the subject matter on journalism is, of course, journalism covers a lot of things, but I’m sure today we’re going to be talking mostly about civic or political journalism. And so the subject matter is highly subjective. It’s not that there aren’t facts in the world. There are facts in the world, right? I would considered irresponsible not to say that actually to the best of the scientific community’s ability to discern it, climate change is manmade and that sort of thing. But on the other hand, what we should do about that fact, there’s no fact of the matter about what therefore the right climate policy is. That depends a lot on what you care about, whether the costs are going to hit you personally or someone else.

Who’s going to have to make what sacrifices, know, the classic subject matter of journalism, who wins and who loses. And so that’s not something about which there is some sort of reality independent of people. And that makes it different from the hard sciences, where you could talk about, you know, what’s happening on Mars without any reference to what humans think about

Eric Schurenberg (08:23.141)
Hmm, hmm. I guess also science, to take that example, defines what it aims to discover the truth of, to discover the reality of. With news, when you have somewhat limited range, even the story selection can be a kind of bias. You can’t really decide what the news is without values.

Jonathan Stray (08:48.738)
Yeah, exactly. So one of the things I tell my students is news stories do not occur in nature. They’re not like some sort of thing out there that you just sort of pluck from the world. So you could call it bias. I would think prefer to call it values. You have to have some sort of news values and say, this is a story worth covering. And you can see this in what sources choose to cover…

…know, leftish media didn’t pay a lot of attention to immigration over the last decade, let’s say. And, you know, rightish media has not really been covering the climate crisis in a way that people on the other side might want. So there’s no, you know, even just starting from what counts as news, there’s no sort of objectivity there.

Eric Schurenberg (09:37.925)
We’ve been talking up to this point, or at least I’ve been feeding you questions kind of with the assumption that journalism is a goal, I’m sorry, that objectivity is a goal that journalists aspire to, even if they fall short because of the way they’ve been brought up or what their values are and so forth. How do journalists actually feel about objectivity? Is there a consensus that parallels what I said or is it actually more complicated?

Jonathan Stray (10:11.662)
Well, so my research assistant and I did a study of three publications that journalists read, Columbia Journalism Review, Neiman Journalism Lab, and the Poynter Institute over three years. And we just looked at every time the word objectivity was mentioned, was it mentioned in a negative or a positive light? And we found that three to one, journalists were critical of it. So within journalism, objectivity is not really… an ideal anymore. However, when you survey audiences and you ask them things like, do you want your news to be impartial or to present both sides equally or something along this line, overwhelmingly and internationally, large majority, as we’re talking 70, 80%, say yes. So the problem we’re having now is objectivity as an idea has sort of fallen apart. Arguably, it never…

…really was a completely coherent idea for exactly the reasons we just discussed. Journalism is not science. It can’t be science. But we don’t have a sort of brand to replace it. And a lot of journalists are really struggling to understand what their position is, especially in the midst of a conflict. So when people are arguing over basic facts, what is the role of a journalist? And so I think the profession is still trying to figure that out.

I have some ideas, but it’s far from settled.

Eric Schurenberg (11:44.729)
I’m interested in hearing those ideas. What should replace objectivity?

Jonathan Stray (11:49.794)
Well, so I don’t think there’s going to be any one thing that replaces objectivity. I think it’s a collection of virtues that we should think about. So you want to think about things like accuracy. That’s a pretty basic one, right? Everybody should agree that if something is reported as true, then it should, in fact, be true. You want to think about comprehensiveness. You don’t want to lie or skew by omission. If there’s a basic piece of information or somebody you should have talked to, you’ve got to include that.

You want to think about things like intellectual honesty, right? The journalist shouldn’t omit or distort arguments to favor something that they hope is through. They themselves sort of have to have this practice of being able to challenge their own ideas as they’re working on it. And then you have a bunch of ideas around, let’s say, fairness. And fairness is kind of an interesting one.

And I think particularly relevant when you talk about conflicts and I view the United States as engaged in a pretty intense domestic conflict, which is escalating, you know, as we have seen occasionally has broken into political violence on both the left and the right in different ways. So I take a page from what people who work in conflict do. So we’re talking the peace building community, not the peacekeeping community, which is…

…soldiers with guns trying to prevent people from shooting each other. Not that that’s not important. It is important in certain contexts. But the peace building community are people who work often at the community level to try to help different factions of society figure out how to live with one another. And so they’re the people who go in both before and after armed conflict and try to build these bridges between factions.

And so they have, over many years, developed very specific ideas of what it means to be fair in a conflict situation. And the core of the idea, or one of the main ideas that they settled on, is not impartiality, but multipartiality. So that’s kind of a weird and academic word, but if impartiality means I’m not on anybody’s side, multipartiality means I’m on everybody’s side.

Jonathan Stray (14:11.116)
And I think that’s kind of the approach that journalists need to take. They’re there to serve the audience no matter what the audience’s political beliefs are.

Eric Schurenberg (14:19.791)
That is fascinating. I had not heard that term before, but that is a really…

let me say complex, think definition of how journalists should approach stories. It reminds me of a phrase that came up in earlier podcast episodes with Amanda Ripley, for example, of Better Conflict and Monica Guzman of complicating the narrative, which is

Jonathan Stray (14:44.024)
Mm-hmm.

Eric Schurenberg (14:52.503)
a technique that you’ve talked about a lot. Explain what that means. Is that sort of the same as sort of multi-partiality?

Jonathan Stray (15:00.994)
Yeah, it’s a similar idea, right? And I know Monica and Amanda well, I’m glad to hear that you’ve been talking to them as well. Yeah, so first credit where credit is due, the phrase complicating the narrative is from a brilliant essay by Amanda Ripley about what, five or six years ago now. So the idea there is that, you know, there’s never just one side to a story and it’s never, you know, one simple idea and it’s…

What conflict does is it flattens these ideas. They’re bad because they’re doing X. Well, there’s probably a reason they’re doing X and there’s some trade-offs involved in doing X and you could do option Y, which is your preferred option, but there’s always some story. So it’s sort of a reinvention of the idea. We used to say, get both sides of the story, right? And that’s not a bad principle.

Jonathan Stray (15:58.89)
But it’s normally saying that when people are fighting over something, there’s always more going on. It’s never the simplistic narrative of one side or the other. And so you should try to find that. And there’s two reasons for that. One is that it’s just more true to life. It’s more real. And the other is that complexity diffuses a lot of those conflict emotions, right? It’s very hard to say, they’re clearly the bad guys when you have a better understanding of all of the various motivations and factors that went into their decision.

Eric Schurenberg (16:36.809)
One example that it came across in the Better Conflict Bulletin is related to a relatively current event, which was the tragic airplane accident over the Potomac River. And the press that I tend to read because I am left-leaning was outraged that the president would bring up DEI as a potential contributor to that.

Eric Schurenberg (17:06.575)
But in reading your commentary on it, realized that there actually is a certain… reasonableness, or at least it wasn’t completely drawn out of left field, that accusation. Could you explain how you complicated the narrative about that accident and the relation between DEI and the Federal Aviation Administration?

Jonathan Stray (17:31.95)
Yeah, so this is a complicated story indeed. So I should start by saying that as of yet, nobody really knows why that accident occurred. first of all, this is entirely speculation about that particular tragedy. And because of that, to immediately say, well, you know, it was or wasn’t caused by this or that is politics, right? That’s political posturing. So let’s just start with that, right? People are already making this into a political football.

Jonathan Stray (18:01.272)
But there is this broader question which had been simmering for some time, which is, does DEI-style hiring, affirmative action, all of these programs to try to recruit underrepresented minorities, does it decrease the quality of employees? That’s just been an argument going on for 40 years, really. And more specifically, has it had this effect at the FAA? And there is a grain of truth behind this narrative, which…

…you know, the way these things work had started to appear in right-wing sources over the last five or six years, but not in left-wing sources, which is this. And the way that hiring used to work at the FAA is there were these colleges that were sort of FAA approved that you went to train at. There was a program to become an air traffic controller. And the FAA recruited from that using a standard aptitude test. And in 2014, they said, no, no, no, we’re not going to do that anymore. Everybody’s degrees are…

…you know, worthless, right? That’s, we’re not going to take that into account. And your aptitude test results are invalid. And we’re going to replace that pipeline. We’re going to keep the aptitude test, but we’re going to start with what they call the biographical questionnaire. And the idea was, this was a, sort of test designed to increase the fraction of minorities who made it into the FDA hiring pipeline, particularly black applicants.

So, okay, I think you can make a reasonable argument for that. The problem was this test was just really strange. It had questions like, what was your worst subject in high school? And you got the highest marks if you said science. And we know this because of a lawsuit where we’ve actually seen the grading results, right? And I don’t think anybody wants people who said they were bad at science directing airplanes or things like…

…you know, what was your your average grade, right? And D scored higher than C. It was very weird. And it seems to be designed specifically to, you know, basically weed out white people, or at least the type of candidates that the FAA normally had. And then there was a related scandal, which is that someone at the FAA who was also a member of an organization trying to increase diversity in hiring, which

Jonathan Stray (20:27.734)
I should say I have no problem with, I think that’s a perfectly reasonable goal and a perfectly reasonable thing for someone that the FAA to be working on. The problem was he then sent a message to the members of this, I forget that actual name of the organization, but it was basically an organization for people of color who wanted to be air traffic controllers. And he sent them the correct test answers. So it was this weird test, which nobody thought made sense. And this sort of a…

…people were given the answer key, but only some people. So this was a scandal. There were lawsuits. We only know all of this because of lawsuits. Congress passed a law in, I think it was 2016, which outlawed the biographical test and so on. So this was a strange incident, now nearly 10 years ago, that arguably damaged the quality of the FAA hiring pipeline. So it’s not.

Jonathan Stray (21:24.13)
completely insane to say that there might be a problem there. On the other hand, they still had the aptitude test. So it’s reasonable to imagine that they didn’t let anyone through who was truly unqualified. But they have been having staffing shortages because they severely damaged their hiring pipeline in this process. So it’s not a simple story. This has emerged over really the last couple of years and mostly outside of the mainstream media.

I have to give a shout out to a person who goes by the name of Tracing Woodgrains, who’s anonymous on the internet but claims to be a law student, who’s really put all of the documentation together on this. And then finally, last week, the story got picked up by the Atlantic. So it took several years to make it to mainstream liberal media.

Eric Schurenberg (22:16.857)
I’ve often had the experience of thinking that a declaration by President Trump or someone on the right side of the political spectrum is making some completely fabulous claim that has no basis in reality and seems just from the kind of conspiracy fever swamp. And yet then finding that there were elements of truth that…

…like the one you just outlined. And also another reaction is that these things have been bubbling on the right outside of my view for a long time and they’re drawing from a kind of set of received wisdom that is familiar to them but totally alien to someone like me.

Jonathan Stray (23:07.948)
Yeah, so I mean, this is the challenge, right? So to distinguish the truth from the bullshit, you have to get fairly far into this. And one of the sort of maxims that I try to convince people of is that, if you wanna keep track of what’s happening among your political opponents or people you disagree with, you have to read the sources that anger you because…

Jonathan Stray (23:34.67)
they’re going to know things that you’re not going to know. Now, often those things are going to be promoted by polemicists and you’re not going to like it, but you don’t have to agree with it. You have to be informed. yeah, your most strident enemies are the people you consider most unreasonable are also the people most likely to know true things that you don’t know precisely because you are unlikely to listen to them.

Eric Schurenberg (24:02.661)
That’s a hard thing to do. will, you know, you’re likely to get angry as you read it. Are there shortcuts? In my class at the University of Chicago, I recommend that people read all sides or ground news, things like that. Or do you think that someone needs to go, you know, jump in, you know, both feet to the, you know, to the…

Jonathan Stray (24:06.978)
Yeah.

Jonathan Stray (24:22.039)
Mm-hmm.

Eric Schurenberg (24:31.363)
…the far ends of the spectrum of their political disbelievers.

Jonathan Stray (24:37.55)
Well, I think there’s a market opportunity here, which some news organizations have started to jump in. So you’ve mentioned All Sides and Ground News. Those are both news aggregators, which try to tell you what the political orientation of the stories are and give you a balanced media diet. But journalism itself could be taking on this task. And I would have to single out Tangle, which is a newsletter run by Isaac Saul, and what his organization does…

…is they round up for whatever big story is happening. Here’s what people on the left are saying, here’s what people on the right are saying, here’s my take. And this is a format. You might call this multi-partial journalism, bit of a mouthful, but it’s journalism that aims to explicitly represent in the most charitable possible light what people with different perspectives think. And they’re growing very quickly…

…which I think suggests that people want something like this. My idea, which I’m starting to play with and working on an article about, is that you should have, really it’s a fairly obvious idea, you should have both liberals and conservatives in the newsroom working on the same story. And this is drawn from an idea that has emerged in science over about the last 15 years called adversarial collaboration. So,

Jonathan Stray (26:03.842)
The idea is you get two scientists or two groups of scientists who disagree about a hypothesis. So for example, going back to the DEI stuff, are women in academia and in academic science disadvantaged? Are they discriminated against? And there was recently a big review done by three authors who disagreed with each other. Some thought it was overblown, some thought it was absolutely yes. And they went and reviewed all the evidence…

…and tried to agree on a joint paper and wrote this joint paper and what they found is they examined, you know, six different criteria. It was like hiring rates and salaries and promotions and paper output and, all this different stuff. And long story short, they found that, you know, women were disadvantaged in some areas but advantaged in others, especially hiring where they had a big advantage. So I would like to see this type of thing happening in a room because I think if you can’t come to some sort of reasonable consensus within the newsroom, I don’t know how the audience is going to.

Eric Schurenberg (27:09.637)
That’s fascinating, Jonathan. By the way, Tangle is another news source that I highly recommend, and I’m a big fan of Tangle’s approach. I believe it’s been copied by other organizations like Semaphore, but not as successfully as Isaac Saul practices it at Tangle. In the example that you were talking about when you were talking about women being advantaged or disadvantaged in science, advantaged how? By… prejudice or lack of or by genetics or ability or sort of kind of social environment like expectations about motherhood or things like that. What did that exactly mean?

Jonathan Stray (27:52.386)
Yeah, I have been thinking I should put together a course on interpreting statistics about discrimination, because it’s complicated. So this particular study looked at outcomes. They looked at, women hired more? Are women paid equivalently promoted? Paper output, that kind of thing. And they found that, you know,

Jonathan Stray (28:17.112)
There’s different ways, right? So in a hiring process, you might imagine that the outcome mostly depends on who’s rating the resumes, who’s scheduling the interviews, who’s making the decisions, right? So that’s a very human process. And there you might say, well, what they found was that women were generally advantaged, probably because of decades of trying to get more women into science. And so universities really want to hire women into the sciences.

Jonathan Stray (28:46.326)
On the other hand, if you look at something like paper output, they saw that generally women had lower productivity than sort of equivalently positioned men. And the hypothesis there is that that’s probably partially biased on the part of colleagues and reviewers and that sort of thing.

I don’t think anybody would be bold enough to claim there is no sexism in academic science, but it’s also structural stuff like motherhood, right? We know domestic labor isn’t evenly distributed, and so when a couple has a child, the men do less of the work and the women do more, so they just simply have less time to be doing science. And then there’s things like pay gaps, right? What they found is that the pay rate gap was real…

…although not as big as people usually assume it is. The results they got were a few percent. And that just seems like straight unfairness. That seems like the sort of thing that an HR department should be able to solve. So whenever you ask about these disparities, it’s never, I think there’s a tendency on the left to say it’s always discrimination and a tendency on the right to say, it’s always ability without stepping into the whole is it nature or nurture argument.

Jonathan Stray (30:08.61)
But the Quotidian truth is that it’s always both, it actually takes a lot of work to untangle what exactly the forces are a complicated narrative. Let me switch back to the topic that you brought up as part of your responsibilities as a senior scientist at the Human Compatible AI Center, which is the creation of artificial intelligence and other kind of offshoots like social media algorithms that are pro-social and that foster human flourishing. So far, at least at the…

Jonathan Stray (30:22.188)
Exactly.

Eric Schurenberg (30:49.803)
…large at the scale of the largest social media platforms and as far as we can tell the biggest artificial intelligence models that is not happening. What progress are you making and do you remain optimistic about the ability to make that kind of change in the technology?

Jonathan Stray (31:09.996)
Yeah, well, you knew I was going to say it’s complicated, right? So the status quo is not great for a variety of reasons. And one of the things that many people have pointed to is the algorithms which select what we see, what is presented to us, are based largely on what we click on, what we share, what we like or put a heart on, that kind of thing…

Eric Schurenberg (31:12.729)
Yeah.

Jonathan Stray (31:38.318)
…in the business we call it engagement, meaning short-term reactions. And that’s not an insane thing to do because we really do spend more time with stuff that is genuinely valuable to us. And if you don’t do that, what you find is that the feeds are just uninteresting, right? By definition, it’s just stuff that we don’t want to read or we don’t want to see. So that’s fine. The problem is that it’s basically a popularity contest. And a popularity contest, that may be okay for

deciding what music we listen to, it’s not okay for deciding what news we read. So what’s the alternative? And there’s a variety of other things that people have explored. largely today I’ll talk about two types of things. One are quality signals. Is there some way that we can judge whether this is quality journalism? Judge by machine, which is a challenge. And another is an idea called bridging based ranking. So the idea is…

…when there’s a complicated issue, abortion, gun control, climate, guess no tariffs, right? Rather than showing you the thing which you’re most excited by, you a particular audience member, promote the thing that people who hold diverging views agree is pretty good, right? Because that’s likely to be the multi-partial article, right? That’s likely to, you want an article that…

…you know, is comprehensive, it’s fair, it’s accurate, you all of these things, right? That as much as possible, people who disagree about the substance or the values will say, okay, you know, that’s a pretty good write-up. Yeah, sure, it includes viewpoints that I disagree with, but it includes my viewpoint as well, and it’s presented in a way that I agree with. So that’s why we use the word bridging, right? It sort of crosses that divide.

Eric Schurenberg (33:25.263)
Mm-hmm. I see. Oh, well, that’s a very interesting approach. I didn’t realize that that was underway. It does seem difficult for machines to make decisions about all of that. It’s hard even for people to make a, to regard…

Eric Schurenberg (33:46.941)
…information that’s delivered, news that’s delivered from a different point of view to say that that is actually fair and not to be sidetracked already with the things that they are most likely to disagree with.

Jonathan Stray (34:00.524)
Well, you know, we just ran a huge experiment seeing if we could do it. So we ran this experiment called the Pro Social Ranking Challenge, where we asked people, teams from around the world, to submit prototype social media algorithms. And then we picked five of them based on peer review. And we ran them with 5,000 sort of random American social media users. And what we did was we used a…

…a custom browser extension, right? So you can install an extension in your web browser on your desktop to change what people see to try to add more of this bridging content as identified by different AI algorithms. And so we literally just finished collecting the data. So I don’t have the results of the experiment, but we are hoping that we will see people being less polarized, more informed, and so on.

Eric Schurenberg (34:42.041)
Hmm.

Jonathan Stray (34:59.944)
So it’s possible to test these things and we’re doing it.

Eric Schurenberg (35:03.205)
Jonathan, we’ll have to come back when the results are in from that experiment. It sounds fascinating. I’m gonna ask you on the way out the door here to review one of the newsletter posts that I thought was particularly interesting and a good way to sign off, which was reasons to be hopeful in 2025 about the state of the culture wars and political conflict, which…

Jonathan Stray (35:22.605)
Hmm.

Eric Schurenberg (35:31.365)
…from time to time seems utterly intractable. So reasons to be helpful.

Jonathan Stray (35:36.942)
I’m trying to remember what I said in that issue. Okay, well, but I do think there are some reasons to be hopeful. I think the main one is we have reached a sort of breaking point in how bad it’s become, right? A lot more people are talking about polarization and difference than they used to. And…

…attention and money are flooding into the field, right? It’s many, many people who care about our country and the state of our institutions and take our state of our democracy are thinking about this problem now and working on this problem. efforts that have started years ago are starting to come to fruition. there’s just now a field of bridging organizations. You you talked to Monica Guzman of Braver Angels and that’s very important work.

But even more important than providing spaces for people who want to talk about politics to talk about them might be changing civic education towards a more pluralist or constructive disagreements type of view. And so there’s programs in schools now teaching people at both the high school and university level how to do this and sort of modeling what it means to live in a country where you’re not going to agree with your neighbors.

So I think there is hope. And I think that that coexists. I don’t think that’s incompatible with political activism. Right. So people sometimes ask me, OK, but, I think Trump is very dangerous or I think the Democrats are very dangerous. You know, you know, I don’t I don’t want to get along better. I want to fight. And what I’m saying is that part of fighting effectively is fighting better. And that’s why it’s called the better conflict bulletin. I’m not telling people don’t fight.

I’m telling people fight in a way that is democratic, that you would support your political enemies fighting in, that preserves the integrity of the nation in the long run. And I think there’s an appetite for this kind of idea in a way that there wasn’t 10 years ago.

Eric Schurenberg (37:51.811)
I agree with you. And I think the growth of Tangle, for example, the very existence of the Alliance for Trusted Media, my nonprofit, shows that it filtered down even to me at the head of a commercial media organization a few years ago. So let’s leave it there. I think that it is, the problem can’t be solved until people recognize the problem and are concerned about it. And then once smart people put elbow grease into the solution, it has a much better chance of showing up.

Jonathan Stray (38:23.826)
Yeah, two of my heroes, Guy and Heidi Burgess, who were on a project called Beyond Intractability, as in intractable conflicts, the conflicts that no one thinks they can solve, they’ve said it’s going to take a generational project, much like a whole generation of people have gone into working and thinking and advocating for climate change. We’re starting to see a generation of people thinking and working and advocating for disagreeing better or bridging or however you want to call this sort of new idea.

Eric Schurenberg (38:54.949)
Great. Jonathan, thank you so much for this time and thank you for the work you do.

Jonathan Stray (38:59.777)
Thank you.

---

Created & produced by: Podcast Partners / Published: Feb 26 2026

---

Share this episode: