Debatten om påverkan på amerikanska valet 2016 fortsätter
Diskussionen om hur i hur hög grad ryska annonser, hackning av kampanjdata/strategiska läckor, trollarméer ( “discourse saboteurs”) m.m. faktiskt påverkade det amerikanska presidentvalet 2016 har tagit fart igen. Den respekterade forskaren Kathleen Hall Jamieson driver i sin nya bok, ”Cyberwar; How Russian Hackers and Trolls Helped Elect a President – What We Don’t, Can’t, and Do Know”, tesen att de var avgörande (se artikel av Jane Mayer i New Yorker, citerad från nedan). Bl.a. har hon studerat effekten på valdebatterna och mediebevakningen. En del andra forskare ställer sig tvivlande, det är inte så lätt att få folk att ändra politisk åsikt, och en annan ny bok tar motsatt ställning i frågan. Några bekanta namn kommenterar, kontentan är väl att frågan är komplex. Längre ner några ytterligare synpunker:
’Other academics may also be skeptical of “Cyberwar.” A forthcoming book on the 2016 campaign, “Identity Crisis,” by the political scientists John Sides, Michael Tesler, and Lynn Vavreck, argues that Russian interference was not a major factor in the Presidential election, and that the hacked e-mails “did not clearly affect” perceptions of Clinton. Instead, they write, Trump’s exploitation of divisive race, gender, religious, and ethnicity issues accounted for his win. But the two books are not necessarily incompatible: Jamieson shows that Russian saboteurs inflamed polarizing identity issues, including resentment among whites that minority groups were benefitting at the expense of “real” Americans—which is exactly what “Identity Crisis” says swung the election.
Recently, Brendan Nyhan, a professor of public policy at the University of Michigan, suggested, in the Times, that most fears about the impact of Russian information warfare in the 2016 campaign are exaggerated. He wrote that “a growing number of studies conclude” that “most forms of political persuasion seem to have little effect at all,” and cited studies suggesting that television ads, direct mail, and door-to-door canvassing rarely sway voters. Moreover, he argued, though the number of Russian-sponsored messages during the 2016 campaign might sound alarmingly large, the universe of information that most voters are exposed to is so vast that the impact of fake news, and other malicious online misinformation, is diluted. He noted, “Twitter, for instance, reported that Russian bots tweeted 2.1 million times before the election—certainly a worrisome number. But these represented only 1 percent of all election-related tweets and 0.5 percent of views of election-related tweets.” He concluded, “It’s hard to change people’s minds!”
Conservative news outlets like the Daily Caller have embraced the view that Russia’s social-media operations were negligible. But Jonathan Albright, of Columbia University’s Tow Center for Digital Journalism, who has been at the forefront of documenting the spread of online disinformation from Russia and other sources, describes the “drop in the bucket” argument as a case of “false framing.” He told me, “A better way to think of it is not as a drop of solution but as pollution. Every piece of messaging has some effect.” It’s misleading, he says, to focus only on the relatively small number of obvious propaganda messages, such as paid Facebook ads. Far more important, in the 2016 campaign, was “organic content”: the countless messages, created by masked Russian social-media accounts, that were spread by algorithms, bots, and unwitting American users. The reach of such content, he told me, “turned out to be huge.” Of the four hundred and seventy Facebook accounts known to have been created by Russian saboteurs during the campaign, a mere six of them generated content that was shared at least three hundred and forty million times. The Facebook page for a fake group, Blacktivist, which stoked racial tensions by posting militant slogans and stomach-churning videos of police violence against African-Americans, garnered more hits than the Facebook page for Black Lives Matter.
It is now understood that Russia’s influence was far larger than social-media companies originally acknowledged. Facebook initially claimed that Russian disinformation posted during the campaign had likely reached only ten million Facebook users; it subsequently amended the figure to a hundred and twenty-six million. Twitter recently acknowledged that it, too, was deeply infiltrated, hosting more than fifty thousand impostor accounts.
James Clapper told me, “It’s hard to convey to people how massive an assault this was,” and added, “I think the Russians have more to do with making Clinton lose than Trump did.” Yet he remains cautious about saying that this is provable. So does Albright, of the Tow Center. He has accumulated a huge quantity of data documenting Russian meddling, but he believes that it remains “difficult to quantify what the impact was on the outcome.” He told me that Russian interference “provoked outrage, created discontent with social systems such as police and safety, pushed certain urban and disadvantaged communities to feel marginalized, and amplified wedge issues beyond authentic reach through social media, which then magnified media coverage of certain issues.” He went on, “That’s an impact. But to translate that into voting patterns is very difficult.” Michael Hayden, the former director of the C.I.A. and the N.S.A., is also agnostic. He has called the Russian attacks “the most successful covert influence operation in history,” but concludes that, although Russian hackers may have, as he put it, “put their thumb on the American electoral scale,” there’s simply “no telling the impact. . . . It’s not just unknown, it’s unknowable.”
Jamieson, ever the social scientist, emphasizes in her book that there is much that Americans still don’t know about the campaign, including the detailed targeting information that would clarify exactly whom the Russian disinformation was aimed at, and when it was sent. She told me, “We need to know the extent to which the Russians targeted the three key states, and which citizens’ voting patterns differed substantially from the ones you would have predicted in the past.”
Philip Howard, the Oxford professor, believes that Facebook possesses this data, down to the location of a user’s computer, and that such information could conceivably reveal whether an undecided voter was swayed after viewing certain content. He also thinks that, if there was any collusion between the St. Petersburg trolls and the Trump campaign, Facebook’s internal data could document it, by revealing coördination on political posts. But, he says, Facebook has so far resisted divulging such data to researchers, claiming that doing so would be a breach of its user agreement.
Even if this targeting information were released, though, questions would remain. Jamieson notes that postelection interviews are often unhelpful, since few voters are able to accurately recount what influenced their decision. Scholars know even less about nonvoters. As a result, she writes in “Cyberwar,” efforts to make an “ironclad” case will be “thwarted by unknowns.” Nevertheless, her book concludes that “Russian trolls and hackers helped elect a US president.”
Jamieson told me that one of her greatest concerns is that voters were unaware of the foreign effort to manipulate them on social media. Had the public known, she believes, there likely would have been a significant backlash. “We want voters to be aware of who is trying to influence them,” she said. “That’s the reason we have disclosure requirements on our campaign ads. We’ve known, at least since Aristotle in Western culture, that the source is judged as part of the message.”’
Jfr Henry Farrell som har en tråd i ämnet, han tar upp en rad frågor och en tredje ny bok i ämnet:
1 https://t.co/dpZppMTxxR I think @JaneMayerNYer has this wrong. NB – this is not the Kavanaugh story. It is her piece summarizing a new book by Kathleen Hall Jamieson which claims that Russian influence operations etc probably won the election for Trump.
— Henry Farrell (@henryfarrell) September 24, 2018
”6 Mayer does not discuss the other important new forthcoming book on this, which takes a data/information science rather than political science approach – Benkler, Faris and Roberts (discussed at https://www.newyorker.com/news/daily-comment/a-new-book-details-the-damage-done-by-the-right-wing-media-in-2016 … by @JeffreyToobin )
7. Chapter Eight provides an extensive account of the operations, based on new data, which is markedly skeptical of claims that the Russian efforts had influence. Russian efforts to split Sanders supporters from Clinton ”seem to have failed.”
8. Russian efforts to push the ”election is rigged” theme were swamped by similar claims that Trump had already made. Benkler, Faris and Roberts repeatedly emphasize that the work is being done by the American media ecosystem – not by Russian hackers.
9. This doesn’t mean that Jamieson is necessarily wrong. Again, perhaps there is better material in the book than in the precis. There may also be other effects that the research literature doesn’t catch. As Benkler, Faris and Roberts note, it would be easy to tell whether
10. efforts to suppress turnout had worked, if we had proper access to Facebook’s internal data. But it is fair to say that Jamieson’s research, as Mayer describes it, does not reflect a consensus in political science, and goes directly against other work grounded in the data.
11. This emphasis has political consequences. First – if it’s the US media ecosystem rather than Russian intervention that is at fault, then the remedies involve confronting what e.g. Fox News has become rather than blaming Putin for everything.
12. (this does not rule out retaliation against Putin – but implies that even if Russian interference disappeared miraculously overnight, US democracy would still be in the same trouble it is). In other words, Mayer on Koch may be more helpful than Mayer on Russian influence
13. Second, as Benkler, Faris and Roberts mention (and as I argue at greater length here – https://foreignpolicy.com/2018/01/17/american-democracy-was-asking-for-it/ … ), if you do take influence operations seriously, overly grandiose claims about their consequences worsen the problem rather than contributing to the solution.
14. The fantastic work of @mollyeroberts and others on ”flooding” in authoritarian regimes as well as the journalism of @AdrianChen and @peterpomeranzev highlights how general sense of chaos, ’nothing is true’ etc can reinforce power of authoritarian leaders. See also @emptywheel
on how sketchy data is leading to a lot of confusion about what Russian influence operations are, or are not doing – https://www.emptywheel.net/2018/01/24/the-gizmo-correlation-doesnt-equal-adversary-nation/ …
15. @schneierblog and I have a new paper on common knowledge attacks which highlights how this can be weaponized against democracy. One of the implications is that panic attacks about influence operations amplifies them, so that weak attacks can have major second order effects.
16. This implies two problems. First – that overly exaggerated stories about Russian influence distract from the real problems of American democracy. Second, that they actually amplify that influence, by creating strong second order consequences out of weak first order attacks.
17. Perhaps the Jamieson book is really strong and will put paid to these worries. Perhaps alternatively, Mayer should have written a piece that better reflected the disagreements of social scientists around this question, rather than highlighting one person’s alarming sounding
18. claims and treating other, less sexy and more skeptical accounts in a relatively perfunctory and dismissive way. I suppose we’ll see when the book becomes available. Finis.”