Hacker News new | past | comments | ask | show | jobs | submit login

The paper (and I only read the abstract) is interesting, but as others have noted, fairly obvious.

One thing that seems to be an assumption is that the "company" needs to provide the explanation. I think it is even better if the user provides the explanation. The assumption a user can't provide it is probably because we've all seen Terms of Service agreements that are totally opaque.

Back in the day when I was doing a bit of admin work, I decided to simplify our TOS, and then when I had to block someone, I just kicked the ball back into their court: "If you would like your account restored, please point out in the TOS what rule you violated." It worked better than expected. People that cared enough about their access to the system usually figured it out pretty quickly, and we got the knowledge that they actually read the TOS to some degree. They got their account restored and that was the end of it. Repeat offenders at that point were willfully causing problems, so we just left them blocked.

Obviously this only works if a human can understand your TOS. Another interesting line of questioning might be "at what complexity level is your TOS useful in shaping behavior and where does it just become a legal shield."

I'm going through a related situation right now, but I'm the one who is the recipient of the negative action.

I posted a video of me playing the piano on YouTube. I got a copyright notification, that I was playing the melody to a song that someone else held the copyright for.

What's the problem? Well, the melody was published in 1886 (133 years ago) under the exact name identified in the copyright claim. The composer died in 1901 (118 years ago). It is not under copyright protection in any jurisdiction! Now, I'm having to appeal the copyright claim... not to YouTube, but to ASCAP (the company who is claiming the copyright in the first place)!!!! In fact, because it was MY arrangement and MY performance and MY production, I own the copyright to that video in every way legally recognized! In my mind, this is THEFT... from ME!

My point is, if YouTube had not at least told me what I'm being accused of, there is no way that I would have figured this out! I haven't done anything wrong! Someone else (ASCAP, ICE_CS) is fraudulently claiming copyright!

Under your system, I would have to "invent" things to confess to.

Of course, now the problem is that I have no power in this situation. ASCAP must agree that they don't want to monitize my video, and they have no incentive to do that. I have no protection or recourse. :(

And, for anyone who's interested, this still isn't resolved.

>I have no ... recourse.

In theory, you could sue ASCAP for damages and/or injunctive relief. Perhaps for libel, for grossly negligently communicating a false and disparaging claim about you to YouTube. Perhaps for tortious interference with business relations. But unfortunately, winning such a lawsuit likely requires an expensive lawyer, and sometimes you can only get as much justice as you can afford to buy.

Exactly why I described it as having no recourse. :/

I argue this is different for a //simple// "human readable" ToS.

In the lack of any clear infringement you can go through the terms of service and disprove every single one, and then come to a list of those that you're less clear about on a second pass and ask if you are mistaken about one of those portions and/or some other area that your conceptions are blind to.

At that level of genuine effort at the very least you've proven an attempt to understand (and maybe that the ToS got a little over-simplified in one or more areas, or that moderation was wrong).

Similarity detection software are unable to distinguish copyright infringement, plagiarism and accidental rediscovery.

I understand that problem, it's just that it is impossible for them to own the copyright that it matched against.

I assume most TOS are just a blob of legalese, and not a "this is what it means to behave on this server".

It feels like the trend has been in the other direction. In most services I've interacted with recently, the service often doesn't even tell you that the content has been removed, let alone why. One example is Discord - if a mod removes a comment you made, it just disappears. You might try to repost it, thinking it failed to send. Similarly if you are removed from a server, it simply disappears from your list. This kind of thing is very frustrating/confusing - not that I expect a mod to necessarily provide explanation (though, that's nice) but at least the service saying "hey, that got removed" prevents me from trying to repost it.

The usual explanation involves spammers/scammers. Something like, if they know they got removed for violating some rule or the other, they can use that information to work around it in their next attempt. But for actual human users it can be very frustrating.

It also only works if they agree that they actually did violate your TOS.

I assume the response "sorry, I looked at your TOS and don't see the problem" would not result in getting their account reinstated.

This assumes we blocked people for no reason just to engage them, or that there was some automated blocking or something. At the time, all this stuff was manual, so once someone was blocked, it was because there was a real problem.

Now, every once in awhile we would get the "johnny logged into my computer and did it, not me." For technical and privacy reasons, we can't really block people, just computers. So, ya, they stayed blocked.

Is it perfectly fair? Nope. But simplifying a TOS definitely cut down on arguing with people, which was a win for us.

That's probably a good thing. If people can't figure out which rule they broke, then they don't understand the rules, and they'll probably break them again soon.

(of course, if everyone complains that they don't see the problem, then maybe it's the mod that doesn't understand the rules, or the rules are ambiguous)

Or, its possible they should be able to appeal it because the company got something wrong, and the violation isn't there. Besides this feels excessively paternalistic.

Sounds like the old cop tactic to get users to confess violations you didn't know about...

I was about to comment to say the same, haha. It's pretty nasty. You have a broken blinker, and now you're going to have more trouble for speeding too, because you guessed wrong.

Originator of the action should be explaining it (even if decision making is automated and the decision making system is not open), seems pretty obvious as noone else can realy know, only guess.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact