3 ms·

Couple of thoughts, unsorted brain dump: * Ranking/Karma systems can and will be abused. See the digg bury brigade, etc. It's really difficult to separate hone

by groby_b 12y ago

Couple of thoughts, unsorted brain dump: * Ranking/Karma systems can and will be abused. See the digg bury brigade, etc. It's really difficult to separate honest feedback from feedback deliberately tailored to vanish unpopular opinion, or hurt people. (It's not like eBay ratings aren't gamed, either. Or Amazon ratings/reviews) * Related to that: One person's troll is another person's idol. Post on any political topic if you don't believe that :) * Related: I believe comment-based rating works slightly better than user-based rating. Just because it takes a lot of effort to silence a user by downvoting every single comment. It still allows burying unpopular opinion * I believe rating systems are gamed because rating is "free". If you only had limited rating capital to expend, I'd assume most people would not want to waste it on burying things. * On the flip side, reputation doesn't matter to many people because it is assigned to online IDs, not actual people. I.e. reputation is free, as well. * Instead of a narrow AI, I'm wondering if a simple Bayesian filter might do the trick for the vast majority of Internet trollery. Yes, it's gameable, but as troll filters become more sophisticated over time, gaming becomes OK. [3] * An immutable ledger of past reputation transactions might be meaningful. (Yes, I'm on the Bitcoin bandwagon with that, sorry). It does not solve the issue of multiple accounts with different reputations, but it's in each participants interest to have each individual entity represent consistent reputation. * I believe any reputation system will have a social component. What your friends think about somebody you never met matters more than what random strangers think, because chances are your friends and you have at least similar value systems. * Society-level interventions might start with educating people in "direct communication" [1] and "nonviolent communication"[2]. And yes, education in general helps too. I'm not sure "biological interventions" are without ethical issues. * You do increase the risk of an information bubble - people might confuse civility with agreement. But I do think that over time, that might sort itself out. In my experience, people I consider of high integrity tend to rank other people's integrity based on behavior, not opinion or political leaning. I might be an optimist. * It seems I use politeness and trust interchangeable. I'll have to ponder that some more, but it seems fundamentally correct to me. I care about respectful politeness, not fake politeness with ensuing backstabbing. [1] https://csde.washington.edu/~mbw/direct-and-indirect-communication.pdf https://csde.washington.edu/~mbw/direct-and-indirect-communi... [2] http://en.wikipedia.org/wiki/Nonviolent_Communication http://en.wikipedia.org/wiki/Nonviolent_Communication [3] http://xkcd.com/810/ http://xkcd.com/810/

A_COMPUTER 12y ago
>I'm wondering if a simple Bayesian filter might do the trick for the vast majority of Internet trollery. Not actually possible because "trolling" so often isn't in the actual words as it is the context. For example, presenting an intentionally poorly-reasoned argument designed to infuriate people. Or for instance, it is possible to appropriate the rhetoric of "social justice" and construct a troll that is indistinguishable from a legitimate statement, the only difference is the context, e.g. used by an oppressor rather than the oppressed. Finally, in many communities it is acceptable as part of the in-group to flame the out-group, but not vice-versa. One person's "eat shit and die" is acceptable, the other person's is not. I suppose this could be mitigated by training your own bayesian filter based on your ingroup and your outgroup separately, based on who you designate your ingroup via social network or something. Sort of like "if I flame an outgroup, I won't be punished for doing so by people who consider me part of their ingroup."

by groby_b 12y ago

A_COMPUTER 12y ago

>I'm wondering if a simple Bayesian filter might do the trick for the vast majority of Internet trollery. Not actually possible because "trolling" so often isn't in the actual words as it is the context. For example, presenting an intentionally poorly-reasoned argument designed to infuriate people. Or for instance, it is possible to appropriate the rhetoric of "social justice" and construct a troll that is indistinguishable from a legitimate statement, the only difference is the context, e.g. used by an oppressor rather than the oppressed. Finally, in many communities it is acceptable as part of the in-group to flame the out-group, but not vice-versa. One person's "eat shit and die" is acceptable, the other person's is not. I suppose this could be mitigated by training your own bayesian filter based on your ingroup and your outgroup separately, based on who you designate your ingroup via social network or something. Sort of like "if I flame an outgroup, I won't be punished for doing so by people who consider me part of their ingroup."