It’s not always easy to do the right thing in tech, and it can lead to some embarrassing (and funny) mistakes. Just ask Twitter.
As you all probably know, Twitter has been trying to stop misinformation about covid-19, the disease caused by the novel coronavirus, from spreading wildly on its platform. One way it does this is by sticking misinformation labels on tweets with false information about covid-19, such as posts talking about a cure for the disease (there is unfortunately no cure yet although scientists all over the world are working on a vaccine).
In recent weeks, Twitter has begun adding these labels to tweets that link the rollout of 5G to the spread of covid-19, a topic that has prompted multiple conspiracy theories. Some conspiracy theorists, for example, claim that radiation from 5G towers weakens immune systems and makes people more susceptible to getting covid-19. There is no scientific evidence to support this. However, misinformation about 5G towers has led people to burn cellphone towers in the UK.
Flagging tweets that spit out 5G coronavirus conspiracy theories sounds like a good thing, right? No doubt about that, except when the filtering system you use gets confused.
Many Twitter users noticed on Friday that Twitter was apparently sticking labels on any tweet that mentioned “5G,” “oxygen,” and “frequency” for some strange reason, even if those tweets were not about harmful 5G conspiracy theories. The label links to a Twitter Moment titled “No, 5G isn’t causing coronavirus,” and includes information from reputable sources debunking the claims.
The end result, of course, was Twitter users posting endless combinations with the words to see if their tweets would get flagged. These types of tweets were obviously not meant to be harmful and were just an attempt to have fun, but Twitter’s filter didn’t have a way of knowing that.
When it comes to why Twitter singled out “oxygen” and “frequency,” the folks at The Week theorize that it’s probably because of a conspiracy theory that claims that 5G is a dangerous frequency that sucks the oxygen out of the atmosphere, thereby disrupting our bodies’ normal functions. This is, of course, false. Twitter hasn’t confirmed that this is the reason why its system started flagging the tweets.
Nonetheless, it’s important to note that this type of “moderation by algorithm” can lead to other issues.
“One of the flaws of attempting moderation at scale by algorithm, a problem that has no bearing on 5G, is that it lets tech companies suck the oxygen out of efforts at reform & regulation, as they shrug & turn ‘we tried one thing with code and it did not work’ into ‘can’t be done,’” defense technology writer Kelsey D. Atherton pointed out in a tweet, which also got slapped with a covid-19 misinformation label.
As the day went on, Twitter eventually stopped labeling the tweets with the keywords we mentioned. A Twitter spokesperson told Motherboard this weekend that the company’s algorithm is “imperfect and constantly changing” based on what’s happening on Twitter. The spokeswoman added that Twitter prioritized over-labeling to err on the side of caution and reduce harm while providing the necessary context.
“In the last few weeks, you may have seen tweets with labels linking to additional info about covid-19. Not all of those tweets had potentially misleading content associating covid-19 and 5G. We apologize for any confusion and we’re working to improve our labeling process,” Twitter Support tweeted on Saturday. “As we improve this process to be more precise, our goal is to show fewer labels on unrelated tweets.”