Why relying on algorithms is bad
About two years ago I started playing chess online and since then I also watch chess videos, usually over dinner or lunch. A funny thing that happened last year in the online chess community was that a live interview between the (then) most popular chess YouTuber Agadmator and chess grandmaster Hikaru Nakamura was banned. for hate speech. Apparently, the algorithm interpreted phrases like “white is better here”, “black defends”, “white attack” and the like as incitement to violence, and completely failed to recognize that the conversation was about a board game.
At the same time, outspoken racists and transphobes were spouting and often continuing to spout their bile on YouTube unfettered under the guise of “Humour” or “Just Asking Kwestchions.”
Today, the algorithm has struck again wonderfully.
I don’t remember specifically when I saw something called fractal wood burning on YouTube, but I think it was last year. I thought it looked cool so I researched how it was done. And I immediately went to the conclusion that it might be cool, but I’m definitely not doing it, not even for a big clock. And the “How To Cook That” YouTube channel posted a great video a few weeks ago explaining why fractal wood burning isn’t a good DIY for woodworkers:
And of course, a great YouTube video can’t go unpunished – the algorithm took it down for allegedly encouraging harmful and dangerous acts. And although it was banned, that same algorithm actually recommended a video to me showing the hack in action. Wonderful work – a warning about an unsafe practice is prohibited because promotion of said practice and actual promotion of it are promoted. Straight logic like a corkscrew.
The video was reinstated after YouTube was pushed back, but I wonder how many really good and possibly important videos get taken down and never come back because the channels that created them were small and didn’t have millions of likes. subscribers to cry foul on their behalf. Because let’s be real – YouTube asks a real human to do the review only when there’s an outcry, otherwise they don’t care.
I think the overreliance on algorithms has great potential for real harm. Human social interactions are so complex that there are humans (like me) who are barely able to navigate them. I don’t think AI is there yet.