Oumou Ly: Welcome to the Berkman Klein Center’s new inter­view series The Breakdown. I’m Oumou. I’m a staff fel­low on the Berkman Klein Center’s Assembly Disinformation pro­gram. Today we’re inter­view­ing Renée DiResta. She is a tech­ni­cal research man­ag­er at the Stanford Internet Observatory. She stud­ies the spread of false nar­ra­tives across social net­works and helps pol­i­cy­mak­ers in devis­ing respons­es to the dis­in­for­ma­tion problem.

Thanks Renée. So today we are gonna talk about COVID-19 and the way dis- and mis­in­for­ma­tion relat­ed to COVID has per­co­lat­ed across the Internet and in so many ways cre­at­ed new prob­lems in dis­in­for­ma­tion that I think pol­i­cy­mak­ers and those of us who were focused on the issue weren’t nec­es­sar­i­ly track­ing before and gives us a whole new lens through which to study issues that kind of are not unique to this prob­lem but tell us a lot about dis­in­for­ma­tion as a whole.

So one of the first ques­tions I had for you, Renée, is just whether there’s any­thing new about what we’re learn­ing about dis­in­for­ma­tion from COVID?

Renée DiResta: Yeah. It’s been real­ly inter­est­ing to see the entire world pay atten­tion to one top­ic, right. This is some­thing some­what unprece­dent­ed. We have had out­breaks in the era of social media mis­in­for­ma­tion before. Zika in 2015, Ebola 2018, right. So there have been a range of moments in which dis­eases have cap­ti­vat­ed pub­lic atten­tion. But usu­al­ly they tend to stay at least some­what geo­graph­i­cal­ly con­fined in terms of attention.

What’s inter­est­ing with the coro­n­avirus pan­dem­ic is of course the entire world has been affect­ed by the virus. And so the oth­er inter­est­ing thing about it has been that very lit­tle in the way—you know, even the insti­tu­tions don’t real­ly have a very strong sense of what is hap­pen­ing, unfor­tu­nate­ly. There’s a lot of these unknowns, as the dis­ease man­i­fes­tion, treat­ment, a lot of the mechan­ics of the out­break of the pan­dem­ic itself are poor­ly under­stood so the infor­ma­tion around them is sim­i­lar­ly fuzzy. And so one of the chal­lenges that we real­ly see here is the chal­lenge of how do we even know what is author­i­ta­tive infor­ma­tion? How do you help the pub­lic make sense of some­thing when the author­i­ties are still try­ing to make sense of it them­selves, and researchers are still try­ing to make sense of it themselves.

But what we see with COVID is a lot of real, sus­tained atten­tion. And I think that what that’s shown us is there’s this demand for infor­ma­tion and it’s revealed gaps where plat­form cura­tion, they don’t have enough to sur­face things. They’re strug­gling with what an author­i­ty is, what author­i­ta­tive infor­ma­tion looks like. I think that that’s been one of the real inter­est­ing dynam­ics that’s come out of this.

Ly: Thanks for that. So one ques­tion I have about the bit on author­i­ta­tive sources is, what makes it so dif­fi­cult for so many of the plat­forms to pri­or­i­tize author­i­ta­tive sources of infor­ma­tion and depri­or­i­tize false con­tent and oth­er sources? Do you think that the polit­i­cal attacks or par­ti­san attacks on tra­di­tion­al­ly author­i­ta­tive sources of infor­ma­tion like the CDC and WHO com­pli­cate the task of plat­forms to pri­or­i­tize that we call ​“good” information?

DiResta: So, for plat­forms to sur­face infor­ma­tion when peo­ple are search­ing for a par­tic­u­lar key­word or top­ic, they have rec­og­nized that sur­fac­ing the thing that is most pop­u­lar is not the right answer either, that pop­u­lar­i­ty can be quite eas­i­ly gamed on these sys­tems. But the ques­tion becomes what do you give to peo­ple? Is an author­i­ta­tive source only an insti­tu­tion­al­ly author­i­ta­tive source? I think the answer is quite clear­ly no. But how do we decide what an author­i­ta­tive is?

So you saw Twitter begin­ning to try to ver­i­fy and give blue checks to doc­tors and virol­o­gists and epi­demi­ol­o­gists and oth­ers who were out there doing the work of real-time sci­ence com­mu­ni­ca­tion, who were rep­utable. And so the ques­tion became, for the plat­forms, how do you find these sources that are accu­rate and that are author­i­ta­tive that are not nec­es­sar­i­ly just the two insti­tu­tions that have been deemed kind of pur­vey­ors of good infor­ma­tion in the past. And per your point, unfor­tu­nate­ly attacks on cred­i­bil­i­ty do have the effect of erod­ing trust and con­fi­dence in the long term.

The plat­forms did begin to take steps to deal with health mis­in­for­ma­tion last year, actu­al­ly, and so a lot of the poli­cies that’re in place now, why health is treat­ed dif­fer­ent­ly than polit­i­cal con­tent, is that there has been a sense that there are right answers in health. There are things that are quite clear­ly true or not true. And those truths can have quite a mate­r­i­al impact on your life.

So Google’s name for that pol­i­cy was Your Money Or Your Life. It was the idea that Google search results should­n’t show you the most pop­u­lar result, because again pop­u­lar­i­ty can be gamed, but it should in fact show you some­thing author­i­ta­tive for ques­tions relat­ed to health or finance because those could have a mate­r­i­al impact on your life. And that was a frame­work that Google used for search begin­ning back I think in 2013, def­i­nite­ly in 2015. But it inter­est­ing­ly was­n’t rolled out to things like YouTube and oth­er places that were seen more as enter­tain­ment plat­forms. So, the oth­er social net­work com­pa­nies began to incor­po­rate that in 2019, in large part actu­al­ly in response to the measles outbreaks.

Ly: Do you think that there are any new insights that this has offered us into maybe the def­i­n­i­tion, the nature, or just gen­er­al char­ac­ter of disinformation?

DiResta: Um, one of the things that we’ve been look­ing at at Stanford Internet Observatory is actu­al­ly the reach of broad­cast media. This is some­thing that—the idea of net­worked pro­pa­gan­da, right. Of course the title came out of some Harvard pro­fes­sors, right, Rob Faris and Yochai Benkler. 

So, the idea of broad­cast media and the inter­est­ing inter­sec­tion between…you know, broad­cast is no longer dis­tinct from the Internet, right, and they all have Facebook pages so there’s…for some rea­son I think peo­ple still have this men­tal mod­el where the media is this thing over here and the Internet is this oth­er thing but I don’t see it that way.

So when you look at some­thing like state media prop­er­ties on Facebook, you do see this real­ly inter­est­ing dynam­ic where overt, attrib­ut­able actors (mean­ing this is quite clear­ly Chinese state media, Iranian state media, Russian state media), they’re not con­ceal­ing who they are—this is not like a troll fac­to­ry or a troll farm ampli­fy­ing some­thing subversively—they’re quite overt­ly putting out things that are…uh… (nice way to say it) con­spir­a­to­r­i­al at best? And so the chal­lenge there is this is no longer just being done sur­rep­ti­tious­ly, this is actu­al­ly being done on chan­nels with phe­nom­e­nal reach. And so, again it’s an inter­est­ing ques­tion of that inter­sec­tion between qual­i­ty of sources, dis­sem­i­na­tion social plat­forms, dis­sem­i­na­tion if you go direct­ly to the source—meaning to their web site or their pro­gram, and just real­ly think­ing about the infor­ma­tion envi­ron­ment as a sys­tem not as this dis­tinct silo in which what is hap­pen­ing on broad­cast and what is hap­pen­ing on the Internet are two dif­fer­ent things.

Ly: Yeah. Sort of relat­ed to that, one of the things we’ve talked about I know…even in our con­ver­sa­tions amongst our groups at Harvard is how dif­fi­cult it is to come up with answers to ques­tions of impact. How do we know, for exam­ple, that after expo­sure to a piece of false con­tent some­one went out and changed their behav­ior in any sub­stan­tial way? And that’s of course dif­fi­cult, giv­en the fact that we don’t know how peo­ple were going to behave to begin with. So, do you think that this has offered us any new insights into how we might study ques­tions of impact? Do you think maybe for instance, push­es of cures and treat­ments for COVID might be illus­tra­tive of the poten­tial for answers to those ques­tions here?

DiResta: Yeah. I think peo­ple are doing a lot of look­ing at search query results. You know, the very real—you know—

Ly: Yeah.

DiResta: When what we’ll call like ​“blue check dis­in­for­ma­tion” or ​“blue check mis­in­for­ma­tion,” maybe char­i­ta­bly, comes out, does that change peo­ple’s search behav­iors? Do they go look for infor­ma­tion in response to that prompt? One of the things that plat­forms have some vis­i­bil­i­ty into that unfor­tu­nate­ly those of us on the out­side still don’t is actu­al­ly the con­nec­tion path­ways from join­ing one group to join­ing the next group, right. And that is the thing that—you know, I would love to have vis­i­bil­i­ty into that. That is like the ques­tion for me, which is, when you join a group relat­ed to reopen, and a lot of the peo­ple in the reopen groups are anti-vaxxers, are you then more like­ly to go join…you know, how does that influ­ence path­way play out? Do you then kind of find your­self join­ing groups relat­ed to con­spir­a­cies that’ve been incor­po­rat­ed by oth­er mem­bers of the group?

I think there’s a lot of inter­est­ing dynam­ics there that we just don’t have vis­i­bil­i­ty into. But per your point, one of the things we can see, unfor­tu­nate­ly, is stuff like sto­ries of peo­ple tak­ing hydrox­y­chloro­quine and oth­er drugs that are dan­ger­ous for healthy peo­ple to take. Again, one of the chal­lenges to under­stand­ing that is you don’t want the media to report on like, the one guy who did it as if that’s part of a nation­al trend, because then that is also—

Ly: Right.

DiResta: —harm­ful.

Ly: Right.

DiResta: So it’s real­ly appro­pri­ate­ly con­tex­tu­al­iz­ing what peo­ple do in response, I think is a big part of our gaps in understanding.

Ly: Yeah. Definitely for sure. Okay. If you could change one thing about how the plat­forms are respond­ing to COVID-19 dis­in­for­ma­tion, what would it be and why?

DiResta: I real­ly wish that we could expand our ideas of author­i­ta­tive sources and have a broad­er base of trust­ed insti­tu­tions like local pedi­atric hos­pi­tals and oth­er enti­ties that still occu­py a high­er degree of trust, ver­sus major, behe­moth, politi­cized orga­ni­za­tions. That’s my kind of per­son­al wishlist.

I think the oth­er thing is that I real­ly want to see us not screw up is every­body who works on man­u­fac­tur­ing treat­ments and vac­cines for this dis­ease as we move for­ward is going to become a tar­get. And there is absolute­ly no doubt that that is going to hap­pen. Happens every sin­gle time. Somebody like Bill Gates could become the focus of con­spir­a­cy the­o­ries and peo­ple show­ing up at his house and all these oth­er things, you know. He’s a pub­lic fig­ure with secu­ri­ty and resources. That is not going to be true for a lot of the peo­ple who are doing some of the front­line devel­op­ment work who’re going to become inad­ver­tent­ly ​“famous” or inad­ver­tent­ly pub­lic fig­ures, unfor­tu­nate­ly, just by virtue of try­ing to do life-saving work. We see doc­tors get­ting tar­get­ed already.

Ly: Yeah.

DiResta: And I think that the plat­forms real­ly have to do a bet­ter job of under­stand­ing that there will be per­son­al smears put out about these peo­ple. There will be disinform—videos made, web sites made, Facebook pages made, designed to erode con­fi­dence of the pub­lic in the work that they’re doing by attack­ing them per­son­al­ly. And I think we absolute­ly have to do a bet­ter job of know­ing that is com­ing, and hav­ing the appro­pri­ate pro­vi­sions in place to pre­vent it.

Ly: What do you think are those appro­pri­ate provisions?

DiResta: If you believe that good infor­ma­tion is the best counter to bad infor­ma­tion, or that more voices—you know, Zuckerberg has said repeatedly—is the anti­dote to—you know, good speech is the anti­dote to bad speech, and authen­tic com­mu­ni­ca­tion coun­ters con­spir­a­cies and these oth­er things, then you have to under­stand that harass­ment is a tool by which those voic­es are pushed out of the con­ver­sa­tion. And so that is where the dynam­ic comes into play where you want to ensure that the cost of par­tic­i­pat­ing in vac­cine research or health com­mu­ni­ca­tion to the pub­lic is not that peo­ple stalk your kids, right. I mean that’s an unrea­son­able cost to ask some­one to bear. And so I think that that is of course the real chal­lenge here. If you want to have that coun­ter­speech, then there has to be recog­ni­tion of the dynam­ics at play to ensure that peo­ple still feel com­fort­able tak­ing on that role and doing that work.