Slashdot Mirror

← Back to Stories (view on slashdot.org)

'It Took 10 Seconds For Instagram To Push Me Into an Anti-Vaxx Rabbit Hole' (vice.com)

Posted by EditorDavid on from the guided-by-algorithms dept.

eatmorekix quotes Vice: It only took around ten seconds. On Wednesday, I created a fresh Instagram account, and followed 'Beware the Needle', a user with 34,000 followers which posts a steady stream of anti-vaccination content. I also followed the user's "backup" account mentioned in its bio, the creator clearly aware that Instagram may soon ban them. Instagram's "Suggested for You" feature then recommended I follow other accounts, including "Vaccines are Genocide" and "Vaccine Truth." I followed the latter, and checked which accounts Instagram now thought would be a good fit for me: another 24 accounts that were either explicitly against vaccinations in their profile description, or that posted anti-vaccine content.

They included pseudo-scientists claiming that vaccines cause autism; accounts with tens of thousands of followers promising the "truth" around vaccinations through memes and images of misleading statistics, as well as individual mothers spouting the perceived, but false, dangers of vaccinating children against measles, polio, and other diseases.
"Instagram told Motherboard it will be looking at different ways to minimize these sorts of recommendations," the article reports, but "did not give a more specific timeframe for this change...."

"For the moment, however, Instagram remains a hot bed of easy to discover misinformation on vaccinations."

4 of 143 comments (clear)

  1. Well, all the platforms are quick to do that. by ruddk · · Score: 4, Interesting

    I was watching a video about the dismantling of Ignalina Nuclear Power Plant and then one about Sellafield and then YouTube started recommending a bunch of related videos.
    And I know this is what happens, so I am leary about watching a conspiracy theory video or an anti-vax video just to laugh at them because I YouTube then will start recommending more of that bullshit instantly. :D

    --
    L'Idiot
    1. Re:Well, all the platforms are quick to do that. by noodler · · Score: 3, Interesting

      "Part of the problem is that these services have too much blind faith in their ability to make predictions based upon incomplete knowledge. "

      You're seriously overthinking this which makes you conclude they are in some way contemplating their actions.
      They're not.
      All they do is shove more of the same shit you watched towards you. That's all they do.
      They don't care for the consequences. There is no 'faith' or 'prediction' involved.
      The algorithm can be sumerized in five words: More of the same shit.

    2. Re: Well, all the platforms are quick to do that. by Evtim · · Score: 3, Interesting

      Clearing the viewing history really helps. With or without being subscribed. Freshly deleted YouTube non-subscribed only gives me viral and trending videos as well as the most popular videos in my region. So there's geo-location at work but that is all...

      BTW, I have yet to see one, even one loony video recommended to me....according to the crazies in the mainstream media I should have, since I follow "gateway to alt-right" personalities...

      This is non-story....unless you are peddling towards total authoritarian control.

  2. They should use PubMed by Krishnoid · · Score: 4, Interesting

    Too bad these sites don't provide preferential recommendations/promotions to accounts and posts that contain links to PubMed articles. If they're going to use these algorithms, they could certainly promote accounts that link to the original publication even if those accounts promote a conspiracy theory, that the research was real and is being suppressed, etc.

    That would at least get actual scientific research reports out front. If you can find anti-vax peer-reviewed research on PubMed, you go right ahead and link/share it. If you can't find it there because it's "being suppressed", find (and let the sites promote) any other sites that publish experimental research in any kind of refereed journal. It would hopefully get this information more in the open and at least cursorily examined for reputability and accuracy by people on the fence.