search

MichaelKreil / openbots

slides, graphics, code and data of the talk "The Army that Never Existed: The Failure of Social Bots Research"
https://michaelkreil.github.io/openbots
6 stars 0 forks source link

Verified accounts can be bots #1

Open mitar opened 4 years ago

mitar commented 4 years ago

There is no reason why they could not be bots. Verified accounts do not verify that they are not bots. Please cite any source to the contrary. See FAQ here: https://help.twitter.com/en/managing-your-account/twitter-verified-accounts

In fact many companies and newsrooms with verified accounts are using bots.

MichaelKreil commented 4 years ago

Of course many media companies use automation. The question is whether this is enough to categorize Twitter accounts of newsrooms as social bots.

Two possible answers:

  1. YES, MANY NEWSROOMS ARE SOCIAL BOTS. In this case, it makes no sense to scandalize "social bots", especially if newsrooms scandalize … well … themselves.

  2. NO, NEWSROOMS SHOULD NOT BE CLASSIFIED AS SOCIAL BOTS. In that case the detection method of Comprop/OII does not work that well and I demand that they publish their list of social bots during the US election to check the extent of their misclassification.

Does this close the issue?

mitar commented 4 years ago

I am not sure if I follow your reasoning here or how does this relate to the issue I opened.

I am pointing to the issue that your line of thought to disapprove Comprop/Oxford research is bogus as well. I am not claiming that what they did is correct or not, it still might be invalid. But the way you disprove it does not work as well. So you can find another way to disprove it, great. But you cannot use comparison with verified accounts. Concretely;

One would expect that among those accounts, the amount of social bots is higher than in a pool that consists exclusively of verified users.

If verified users can be bots or not, then there is no basis for this expectation.

Also:

Why should there be more social bots among users that have been verified by Twitter than among accounts tweeting about elections? There is no other scientific explanation for these results than a flaw in the method itself.

Or maybe simply verified accounts are power users which use bots.

But yes, there work seems to maybe detect bots, and it is unclear what is difference between social bots and bots. So why would social bots tweet more than regular bots.

But I still do not get why comparison against verified accounts tells us anything. Why social bots would not be among verified accounts? Why bots would not be?