This stuff is harder than it looks
The world was deeply shocked yesterday by the images we all saw of the Notre-Dame cathedral’s spire collapsing as the building was engulfed by flames. Even people who had never visited it were emotionally affected
by the idea of such a prominent, ancient Paris landmark being devastated so quickly.
YouTube quickly switched off the feature, but anger at the inappropriate nature of what happened remained.
It’s important for us to expect high standards from platforms like YouTube. If its algorithms politically radicalise people, or encourage them to avoid immunising their kids, or recommend children stop eating properly, then there are real-world consequences that need to be taken seriously.
And there’s an irony that YouTube’s own technology designed to stop people being misled by ‘fake news’ ended up spread a false (or at least confusing) equivalence of its own.
But we also have to remember that this stuff is hard. Policing platforms like YouTube and Facebook is an editorial task like no other. It’s never been done before. The scale at which they operate mean any best practices from managing smaller online communities, or editing a publication or TV news bulletin, go out the window. There is simply no comparison.
As we saw with the recent massacre in Christchurch, policing breaking news events on big social media platforms as they happen and in the immediate aftermath is an incredibly tricky problem. Automation alone doesn’t work, and humans can struggle with the sheer volume of content coming in every minute. If there’s a problem somewhere in the system, like machine learning algorithms confusing a burning cathedral with 9/11, will humans know where to look to stop that ever going out to the public?
That’s not to say it’s impossible for YouTube to every clean itself up and run a tight ship. It will likely require a mixture of software and humans working together in just the right way, with the right level of experience of what can go wrong and how to fix it.
But YouTube is learning to do something that’s never been done before, so even if the team there knows exactly what outcome they’re aiming for, they’ll make mistakes along the way and we need to accept that.