Intel gave a presentation at GDC a number of weeks again, however I’m guessing no one truly watched it as a result of it took until this week for anyone to notice this totally absurd pitch from the corporate, which desires to make use of AI to observe and censor “hate speech” in your on-line voice chat, and let customers toggle simply how a lot hate they wish to hear.
It’s a service launching later this yr known as Bleep, which “is a user-facing software that makes use of AI to detect and redact audio primarily based on consumer preferences”. Which principally means it’s going to monitor audio because it comes out of your system, and mute/beep your audio system or headphones when it detects unhealthy phrases.
It’ll detect these phrases utilizing AI, but it surely’s the consumer preferences a part of this that’s the most hilarious/horrifying. Right here’s a have a look at Bleep’s backend settings, and it’s a technological hellscape.
Be at liberty to advantageous tune! Would you want to listen to folks body-shaming you? Oh, you’ll? However solely a little? OK, positive!
So many selections! I deeply admire the actual fact I can solely hear somebody screaming white nationalist taunts down a microphone “most” of the time—typically you want a break, in spite of everything!—and will likely be considering very laborious about whether or not I wish to toggle that N-word swap to its “on” or “off” place.
It’s ghastly that one thing like this ever left a whiteboard, not to mention made all of it the best way into a significant presentation, however then we’re years previous the purpose the place we ought to be anticipating firms like Intel to consider something besides methods it will probably waste hundreds of thousands attempting to make use of its personal expertise to fight deeply human issues.
Perhaps there was an excellent intention right here in some unspecified time in the future. Letting folks get pleasure from a safer on-line expertise is, in spite of everything, an excellent factor! However this, this isn’t the best way to do it. Hateful speech is one thing that must be educated and fought, not toggled on a settings display screen.
You’ll be able to watch the presentation at across the 29:30 mark within the video beneath (although it ought to autoplay at that time in the event you click on on it anyway).