15 February 2020, Bavaria, Munich: Mark Zuckerberg, Chairman of Facebook, speaks on the second day … [+]
dpa/picture alliance via Getty Images
The machines are here, and it’s time to start worrying.
I’m mostly joking about that, by the way. While there are legitimate concerns about ceding control over to machine learning and automation during the coronavirus crisis, it pales in comparison to the pandemic sweeping the globe right now.
For those of us who are hunkered down in our homes, trying to practice social distancing when we are out in limited spurts, there is a concern that some of the artificial intelligence technology doing work for us is lagging.
Recently, Facebook announced that there was a glitch in how posts were moderated. They are working on it, they said. I’ve heard rumors that the humans who moderate posts for accuracy were sent home and that the bots doing the moderation were glitching, but I’m still not sure who or what to believe. Most of the spam detection and false removals have been related to coronavirus news.
Cue the dark lord music and a Terminator with red eyes.
Honestly, I’m worried enough already.
I keep hearing reports about deleted posts from everyone under the sun (and maybe a few living underground). Random colleagues have told me they have posted a link to a legitimate news source — even my own column! — and then received a troubling spam detection notice. It’s gotten to the point where I’d be surprised if a post about the virus actually did stay put.
We’re talking about The Atlantic and Politica. And CNN.
These are not fake news purveyors, so it’s extremely odd to see the spam warning. (I have not actually seen this myself.) You blink and your post is gone. It’s a sign that robotic detection is not quite perfectly accurate yet.
The reason it’s a bit worrying, at least from the standpoint of social media and spreading accurate information, is that we didn’t really know those human moderators were so critical. I’ve always assumed they were a back-up. I can imagine a small team sitting in cubicle farms glancing at reports once in a while and guiding the bots who do the hard work. It is likely the other way around. The human moderators are possibly only assisted by normal automation and not much AI.
And it makes me wonder how the bots are going to “help us” in other ways. Making sure bridges don’t collapse? Flying airplanes?
(Facebook did not respond immediately to a request for more information.)
What will actually help is if companies like Facebook and Google ramp up their efforts at machine learning to do this type of laborious work. We’re worried enough. You might view social media as superfluous in the current crisis, but in reality, it is now more important than ever. Accuracy is critical.
Without help from machine learning and bots, without artificial intelligence augmenting the mundane tasks humans must still perform, it will be an even more earth-shattering experience. If there is ever a time we need the bots, it’s now.