Submitted via IRC for TheMightyBuzzard
Twitter has launched a new way to punish users for bad behavior, temporarily "limiting" their account.
Some users are receiving notices their accounts are limited for 12 hours, meaning only people who follow them can see their tweets or receive notifications. When they are retweeted, people outside their network can't see those retweets.
Some speculate these limitations are automatic based on keywords, but there is no hard evidence.
This would be fine if this was used uniformly to clamp down on harassment, but it appears to be used on people, simply for using politically incorrect language.
Source: http://heatst.com/tech/twitters-new-tool-to-crack-down-on-politically-incorrect-language/
(Score: 2) by gidds on Wednesday February 22 2017, @02:34PM
That sounds quite promising, but I can't help wondering if it's ultimately suffering from the same problem.
I generally don't mind individual words; what I want to avoid is things like hatred, bigotry, prejudice, and suchlike — and you can't filter those with a regexp. Nor can you filter deliberate disruption and trolling. Nor deliberate lies, astroturfing, advertising, scamming, or a variety of other things that humans can recognise (unless done verysubtly) but machines can't. (At least, not currently with sufficient accuracy.)
A site which could filter that (at user choice) would be a serious improvement.
[sig redacted]