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.
I was going to comment the same thing. It's someones opinion. The fact that you don't agree with it, does not make it "hate speech".
In the Islamic world, saying "kill all the gays", is perfectly acceptable. On a far-right political conference, saying "deport all the muslims" is perfectly acceptable as well.
The concept of what is acceptable and what not is defined by the circumstances and subjective interpretation by its recipients.
Simply having a small minority of right- or left leaning individuals "determining" that something is hate speech, does not make it such.
You may believe hate speech is an acceptable form of expression, but that doesn't mean it doesn't exist. It's literally defined in the dictionary now
https://www.merriam-webster.co...
You may believe hate speech is an acceptable form of expression, but that doesn't mean it doesn't exist. It's literally defined in the dictionary now
https://www.merriam-webster.co...
I think you're just feeding a flamboyant troll. My personal theory is that some of the low-digit accounts have been hacked and hijacked by professional trolls. Either that, or more Libertarian insanity. I've yet to meet a Libertarian who actually understood his worship words...
However, these years I prefer to think in terms of solutions. Imagine that the troublemakers, hate speakers in this case, were assisted in rendering themselves invisible? Not absolutely invisible. After all, they still have their rights under the First Amendment (in the American flavor), but if we see a nut screaming conspiracy theories into a bullhorn all of us should also have the right to walk far around him.
My solution approach is currently tagged MEPR for Multidimensional Earned Public Reputation. In Slashdot terms, you might think of it as a kind of symmetric karma on steroids. In mathematical terms, it would be defined as a convergence between what you do in public and how people react to what you did (while also considering the MEPRs of the people with the reactions). It is NOT an attempt to reduce people to a single number (in the Chinese and Facebook fashions), but rather a kind of lens that would help you know where to look and not look.
In theory, this is the kind of thing AI could be good at. In practice, you know the gamesters will look for new wrinkles, so you still have to continue evolving...
Time's up, but ADSAuPR, atAJG.
Freedom = (Meaningful - Coerced) Choice != (Speech | Beer^2), and sad sock puppets' bad mods avail them naught.
"Hate crimes" are just crimes
When did it become fashionable to pretend that there sub categories don't exist.
I propose we should no longer distinguish between digital computers and analog computers because they're just computers.
Unless there are "happy murders" and "love frauds" ?
Congratulations, that's about the stupidest thing I've read on the internet today. That's a pretty high bar.
Since when does something have to be done with love for it to not be done with hate? Someone murdering you to nick your wallet neither loves nor hates you.
SJW n. One who posts facts.