Anatomy of an online misinformation network
journals.plos.org
Anatomy of an online misinformation network
1–10 of 111 posts
Re: Anatomy of an online misinformation network
#2The truth is one, but lies are many
Re: Anatomy of an online misinformation network
#3The article specifically mentions snopes and politifact as sources of truth. They can be, but have also been shown to be less than perfect, and do tend to lean left.
Yesterday, twitter was full of pictures of the immigrant kids that Trump put in cages. This morning, we find out that the pictures were taken in 2014. It's a difficult problem because the retractions are never broadcast/retweeted as prevalently as the original fake news.
This is a difficult area, and the stakes are so high that the arms race will doubtless continue. Because the stakes are so high, you are competing with organizations from allied countries as well as non-allied countries, and from multinational corporations and industry groups, all throwing tens of millions of dollars each on think-tanks and lobbying groups, whose main activities are to publish source material that can be used as supporting "facts" for various news stories.
Re: Anatomy of an online misinformation network
#4Interesting content, but I get really worried when studies and research show a tendency being influenced by trends; it's often a sign of poorly aligned incentives and sometimes even produces hasty or bad results in favour of career maximization.
Re: Anatomy of an online misinformation network
#5Not everything is a technical problem to solve and acting like it is is a dangerous path.
This will never work in practice, truth is not a popularity contest.
Re: Anatomy of an online misinformation network
#6Not everything is a technical problem to solve and acting like it is is a dangerous path. This will never work in practice, truth is not a popularity contest.
What aspect of the study makes you believe they are relying on popularity? They are mostly relying on a very limited set of approved sources (which could be criticized), but their approach seems pretty straightforward. You've gotta have some baseline to compare to.
I'm glad someone is making a public/academic attempt to measure this, because it's hard to imagine just how easy it is to spread disinformation.
What would you have done differently?
Re: Anatomy of an online misinformation network
#7In addition to identifying the network, the authors also calculate that taking down as few as ten influential accounts can halve the retweets of misinformation. I'm not convinced that it's quite that easy, because the network would respond and adapt, but it's still an eye-opening demonstration of the power of social media companies.
The paper also finds that many misinformation accounts are not bots. Currently, misinformation is not a bannable offense for most social networks. Twitter, Facebook, and YouTube leaders have said, verbatim, that they don't want to be "arbiters of truth". Of course, that leaves open the question of what responsibility they have and what actions they should take.
Re: Anatomy of an online misinformation network
#8The truth is one, but lies are many
"... If falsehood had, like truth, but one face only, we should be upon better terms; for we should then take for certain the contrary to what the liar says: but the reverse of truth has a hundred thousand forms, and a field indefinite, without bound or limit. The Pythagoreans make good to be certain and finite, and evil, infinite and uncertain. There are a thousand ways to miss the white, there is only one to hit it. ..." -- Montaigne
https://ebooks.adelaide.edu.au/m/montaigne/michel/essays/boo...
Re: Anatomy of an online misinformation network
#9Why not to use detected bots for automatic false-news alerter, like ad-block or spam filter, but for false news?
I currently use few accounts, which are spreading false-news, as indicators with great efficiency. A bit of automation and a browser plugin from trusted source (like Bellingcat) can help to fight that problem.
Re: Anatomy of an online misinformation network
#10The article specifically mentions snopes and politifact as sources of truth. They can be, but have also been shown to be less than perfect, and do tend to lean left. Yesterday, twitter was full of pictures of the immigrant kids that Trump put in cages. This morning, we find out that the pictures were taken in 2014. It's a difficult problem because the retractions are never broadcast/retweeted as prevalently as the or…
> have also been shown to be less than perfect
Perfection shouldn't be the objective. If it's better than what we have then it's worthwhile.
> do tend to lean left
Citation needed.
And if you know of a fact-checking resource that has less bias, please share.