Originally published at: https://boingboing.net/2019/04/29/uber-apologizes-for-tweeting-r.html
…
1 Like
Need a human in the loop.
4 Likes
I’m definitely inclined to blame Twitter rather than Uber for this.
9 Likes
Seems the simplest thing to do is include only the @username in the reply or just drop the personalization altogether.
4 Likes
Nadir? Navigator? Naysayer? Necromancer? Negotiator? Nether? Newscaster? Nibbler? Nipper? Niter? Nitpicker? Noisemaker? Nonbeliever? Nor’easter? Numerator? Nutcracker? The suspense is killing me.
7 Likes
Why not both. Two huge tech companies and nobody in a position of power in either of them had the foresight to think of something as simple as a filter for a list of banned words, that at least flags a post for automatic human review?
13 Likes
If someone says that they are a -adjective- and that they identify as -adjective- as their name; then who am I to disagree with them? Wouldn’t it be disrespectful to call them something else?
This is agiprop nonsense. They played stupid games and won stupid prizes. It should take everyone two seconds to figure out what happened, and yet some people so want to be offended that they will be willfully offended by agiprop nonsense.
2 Likes
IS anyone offended? I know it’s an offensive word but this just seems like a silly gag.
1 Like
Do you want ■■■■■ censorship? Because that’s how you get ■■■■■ censorship!
4 Likes
I want ■■■■■ censorship
2 Likes
■■■■■ should be free.
3 Likes
I submit that if this strikes you as a cause for even mild outrage, you must have been paying remarkably little attention to what is actually bad about Uber, Twitter and racism.
As to whether it’s a good joke… meh… I did chuckle, but if the prankster goes on tour I will probably skip their date in my town
5 Likes
Clearly it’s a six-letter word. Use a little context.
Obviously the word is “Nagger.”
2 Likes
I’m torn between the common-sense wisdom that any automated system that makes user-input content public should be at least minimally guarded against users inputting publicly inappropriate things and Falsehood Programmers Believe About Names #31, “I can safely assume that this dictionary of bad words contains no people’s names in it.”
6 Likes
You nailed it, @hawkeward. Trying to blacklist “bad words” results in both the scunthorpe problem, which you alluded to, and also, such blacklists are trivially evaded. If you blacklist a word like “shingles”, then the twitter trolls simply set their names to “shnigles” or “shïngles” or “sh1ngles” instead.
Better to just not include user input in your bot responses in the first place, rather than trying to identify the “bad” words.
A smart programmer realizes users might provide naughty words as input, and tries to detect naughtiness. A wise programmer realizes “simple” attempts at classifying naughtiness are ineffective and sometimes cause more harm than good – and thus simply doesn’t reproduce potentially naughty input unless absolutely necessary – or accepts that naughtiness will happen, and is a problem that can be solved in other ways (moderation), rather than at the software level.
3 Likes
Their twitter bot might be easily tricked, but their self driving cars will be ready for any situation on the road.
9 Likes
This guy spent a lot of time and effort figuring out how to keep out racist trolls from making his banana say bad words. It was quite a back and forth of fix and work around.
ETA - Oops - forgot the link!
1 Like
I thought that, by now, people would have eliminated “we’d like to” from their apologies.
1 Like
■■■■■ Von Lipwig
2 Likes
It’s easy to blame a faceless corporation for lacking sensible policies. But we ought to also spread some blame onto the degenerate and anti-social use base that thrives on social networks.
1 Like
