Slashdot Mirror
AI Still Useless at Catching Hate Speech, Research Finds (theregister.co.uk)
New research has shown just how bad AI is at dealing with online trolls. From a report: Such systems struggle to automatically flag nudity and violence, don't understand text well enough to shoot down fake news and aren't effective at detecting abusive comments from trolls hiding behind their keyboards. A group of researchers from Aalto University and the University of Padua found this out when they tested seven state-of-the-art models used to detect hate speech. All of them failed to recognize foul language when subtle changes were made, according to a paper [PDF] on arXiv. Adversarial examples can be created automatically by using algorithms to misspell certain words, swap characters for numbers or add random spaces between words or attach innocuous words such as 'love' in sentences.
... they're fucking *brilliant* at CREATING it!
https://www.theverge.com/2016/...
How do you catch something that doesn't exist?
Trained snipe and jackalope(s) sniff it out.
Amused to find that within two minutes of posting, my contribution was modded "50% Insightful, 50% Redundant".
I have to admit, it's a fair enough grade. Even as I posted I was thinking to myself-- what I am saying is so blindingly obvious, does it really need to be said at all?
But I think the very existence of this "research" shows that it does need to be said, loudly, and often, and by as many people as possible.