How To Not Get Punked By Automation
"If humanity is going to retake social media and push back against the tide of automated attention-manipulation, journalists need to get smarter themselves so as not to fall prey to these electric demons—and start covering bots, and their changing strategies, as a beat", Nicholas Diakopoulos claims.
Diakopoulos recently wrote an article for Colombia Journalism Review titled The bots beat: How to not get punched by automation. In this piece, he tackles how we all should tackle bots. Because we have to--they are everywhere.
Diakopoulos points to a report by the New York Times report that dives into the social media black market, where popularity easily can be bought. Millions of fake followers and bots are out there; an astonishing 3.5 million of them.
Diakopolous writes that bots aren't all bad, "[b]ut they can easily be set to bully, intimidate, harass, pollute, and push political agendas online".
Four main bot issues
Diakopolous identifies at least four ways bots can co-opt public media and attention:
1. Bots can manipulate the credibility of people or issues
2. They can amplify and spread propaganda and junk news
3. They can dampen or suppress opposition and debate
4. They can intimidate or deny access to authentic people who want to participate
Journalists may be part of solution
Discovering bots can be a challenge. Because of this, good bots like e @probabot exists. This is a bot that hunts and shames other bots like this:
Diakopoulos recognizes that it's difficult for journalists to handle many of the issues concerning automation. He can picture a solution.
"Platforms should be building these various forensics lens into their tools (e.g., Tweetdeck in Twitter’s case) so journalists can flag suspicious accounts—and funding fellowships so that more reporters can cover the bots beat".
This would, according to the researcher, " facilitate future automatic bot detection and removal efforts."
And that isn't the end. Diakopoulos believes journalists even can help develop techniques to help identify bots at scale, and identify patterns to translate into data mining techniques that make it possible to thwart this type of attack in the future.