| It isn't a secret that Ishinashi has been creating problematic troll reviews for a while now. The current status quo is as follows: he makes an account, posts the same 2-3 reviews, users report the reviews, and they eventually get taken down by review or community moderators. The cycle continues. This "whack-a-mole" game that we're playing with him solves the problem in a reactive manner, but I think a proactive solution would be more sustainable for the mental health of the moderators and the community alike with this troll who won't give up. I searched on suggestions last night and I couldn't find anything like what I thought about. Maybe I missed a thread detailing this or any previous threads suggesting this were removed. But anyways, I think that the review system should have an automatic word-filter which would either prevent a review from being posted entirely or shadowban the review from being seen in the recent reviews section of the website. Basically, if the review contains any flagged language, then the review would not be posted. This way, the issue of reviews containing bigoted language in them wouldn't be an issue. However, such a system like this, if implemented, should only filter out slurs and not vulgar or bad language. I think that I can speak for most people and say that I'm 100% okay with their review saying fuck in it, but I don't exactly like waking up every morning and checking this website to see the latest serving of slur slop on the front page by Ishinashi. A suggestion of this magnitude would definitely be controversial if it was implemented because an automatic filter does introduce some ethical questions: "Does this limit free speech?", "Are slurs free speech?", and "What constitutes a slur?", are three potential questions which I thought up just now that I can see people asking in regard to a word filter. First, are slurs free speech? In most liberal democracies in the world, hate speech is not free speech and would not be protected by most people's governments. The United States is one of the sole exceptions to this norm, as its first amendment protects any group or person from expressing their political ideas so long as that person or group does not have intentions to cause harm to people. This Stanford University philosophy encyclopedia page has an interesting section on the U.S. and the Skokie march if you would like further reading. However, MAL is a privately owned website and business, thus in the domain of MyAnimeList, its rules apply first and foremost. The website's guidelines that apply for both the forums and the website as a whole forbid "Not Safe For MAL" content, which bigoted language such as the average Ishinashi review fall under. So on MyAnimeList, hate speech is bigoted speech, and therefore would not be free speech. So, this answers the first two questions related to free speech. But what about the slurs themselves? What constitutes a slur? Well, I think to keep it simple, this filter should discriminate commonly agreed upon slurs that are bigoted on a basis of race, nationality, sexuality, and ability. Anything where the meaning isn't agreed upon or the word may have a different meaning in another context, like the word "trap" for example, I would not add to this filter. I think most MAL users are knowledgeable enough with Thomas Paine to understand common sense applies with slurs and most people know the commonly understood slurs. Before I close this post, I'd also like to try and present some "counter-suggestions" that would accomplish the same thing and why they would not work as well as a slur filter in the reviews. 1. What about IP-banning Ishinashi? IP Bans do not work in the present day of the Internet. It is incredibly easy and accessible to spoof your IP with a VPN. Most good VPNs are pay walled at an accessible price, but free VPNs exist and charge you with your data. There are also proxies like what OperaGX's "VPN" is that can achieve a similar result. But let's say that Ishinashi was trolling the MAL Forums on a budget and he couldn't afford a VPN for some reason. Well, that doesn't matter. If you unplug your router or just turn it off, you can change your IP Address for free. Not to mention, if he was desperate enough to troll, he could go to any public space such as a library or any other public space with free Wi-Fi to access the website. TL;DR: IP Bans are irrelevant. 2. What about mandatory 2FA Authentication to post reviews or post on the forums? This would also not work to stop him from creating accounts to do what he does, because the 2FA which MAL uses does not actually require a cell phone to pair. You can pair an account to the official Google Chrome authenticator browser extension to achieve the same results as you would with a phone. I mean sure, it would slow him down because he would have to pair 2FA to each new burner account and burner e-mail he utilizes, but it accomplishes as much as pouring water on an electrical fire. This also has other ramifications besides complete redundancy in stopping these troll reviews. Mandatory 2FA authentication would lock out legitimate users from using the forums or posting reviews. The only benefit to this is that it would completely stop those random advertisement bots from Lahore, Pakistan trying to sell MAL users erectile dysfunction pills for a little while, but the costs outweigh the benefits. TL;DR: Good faith users suffer by having to go through a pointless chain-link fence, while Ishinashi can just climb over it every time. 3. What about a cease and desist/court order to stop him from using the website? This would require a ton of legal effort to prove that he was breaching the website code of conduct and a ton of money which I quite frankly think wouldn't be worth it. It is legal to block a user from accessing your website, but the digital blocks are easy to circumvent as I discussed. An actual probation and/or monitoring would be a lot harder to achieve and is typically reserved in many of the world's liberal democracies for thug shaking hackers, Chris Paul highlights distributors, drug-distributing felons, and pirate captains. As far as I know, there is zero legal precedent to stop a troll from using a website, and it would not be financially viable to do so. TL;DR: Lol. 4. Why not just ignore him? I mean, I try, believe me. But it keeps happening. I'm not offended by the language because I'm not a member of the LGBTQ community, yet I see as many as 4 Ishinashi accounts in a day and those reviews sit for a while because he outpaces the natural supply of fresh reviews. It's annoying and he is winning his guerilla war against the website. I do report the posts he makes and I know that I'm not the only one who does it because sometimes people beat me to it. But even if a post of his is deleted, he can outpace the natural supply of fresh reviews anyways, plus sometimes the reviews sit for a while because not every moderator has the time to sit 24/7 babysitting the website. Moderators and admins have lives to live and responsibilities that they are responsible for besides the site. TL;DR: His volume far exceeds the volume of legitimate reviews, plus he trolls on the forums too. So, if we were to stop Ishinashi in his tracks, then I think the only feasible way to do so would be to introduce a filtering system of this scale to the website. I'm certain that it is feasible because other websites have integrated them. But, it would be a first for an anime database website and it would be controversial. I unfortunately know nothing about the feasibility in coding it on MAL as I'm not a web developer. Regardless, I hope the site team does see this suggestion and takes it into consideration if they haven't already. Even if they don't implement a solution as radical as this, visibility is what I strived for when I typed up this post. And for any users who like or dislike this suggestion, please provide your feedback. Agreements and disagreements, as well as tweaks or additions, are valuable when it comes to promoting a discussion on this issue. Thanks. |
| I mentioned something to this effect to one of the Community mods via PM when I got the automated "Thank you for your report" response after one of the reports on him I submitted had been handled. I know that the site has had a spam filter at some point, probably still does, so extending it to include reviews shouldn't be too hard. I do agree that the dealing with the 3-5 accounts per day that he makes should be as automatized as possible, so that the mods can a) focus on other violations that matter, instead of, as you put it, "babysitting" his crap b) be not as stressed from dealing with his crap on top of their other responsibilities both here and irl. another "solution" that you missed that I will touch on briefly, since it is also not viable, is having someone on staff who has equally as much free time who's sole job it is to keep Ishinashi in check. A menial task like that would be a downer imo, and would be counterintuitive. ETA: I would add to the slurs, specific phrases known to be used by him to prevent spam reviews such as this one |
Fleeting_DreamAug 21, 10:03 AM
    . . We will keep watch! . . |
Dulonkk said: I oppose your suggestion on the basis that slurs themselves are vulgar. There are absolutely people who are offended by the use of "fuck" in a review; the word wouldn't be considered vulgar otherwise. I don't think their feelings are any less worthy of consideration than yours.However, such a system like this, if implemented, should only filter out slurs and not vulgar or bad language. I think that I can speak for most people and say that I'm 100% okay with their review saying fuck in it, but I don't exactly like waking up every morning and checking this website to see the latest serving of slur slop on the front page by Ishinashi. |
| その目だれの目？ |
Fleeting_Dream said: If the website does indeed have a spam filter, assuming it is still operational, then it shouldn't be too hard to toggle it for the reviews. Maybe they already have, but as you said, it could use the tweaks necessary to rule out and filter any posts made by Ishinashi. Your suggestion of adding known Ishinashi phrases to this proposed filtering system is genius. I wish I thought of that myself. This expanded filter would also eliminate virtually all of those Oshi no Ko posts he makes as well, at least until he adapted to copying another review to spam. The only potential issue that I can see by incorporating Ishinashi's phrases into this filter are false positives. I think it would be really unfair for a legitimate reviewer to have their review removed if their review contained similar phrasing to that of Ishinashi with none of the dogmatic content. In any automated system, whether a filter for a messaging board or an anti-cheat for a video game, false positives are rare, but they do occur and they punish legitimate users. Regardless, I think the benefits outweigh the risks and the trade-off of rare false positives would be worth instantly nuking his spam.I know that the site has had a spam filter at some point, probably still does, so extending it to include reviews shouldn't be too hard. I do agree that the dealing with the 3-5 accounts per day that he makes should be as automatized as possible, so that the mods can a) focus on other violations that matter, instead of, as you put it, "babysitting" his crap b) be not as stressed from dealing with his crap on top of their other responsibilities both here and irl. ... ETA: I would add to the slurs, specific phrases known to be used by him to prevent spam reviews such as this one Fleeting_Dream said: another "solution" that you missed that I will touch on briefly, since it is also not viable, is having someone on staff who has equally as much free time who's sole job it is to keep Ishinashi in check. A menial task like that would be a downer imo, and would be counterintuitive. I didn't mention this because I didn't think at the time of me writing my post that there would be anyone willing to consent to such a task without some sort of monetary motivation behind that. And I'm sure as you know since you are a former moderator yourself, moderation is essentially volunteer work done out of good will. That's obvious, but for anyone who may not have known that, moderators do not get paid to deal with people like Ishinashi. I think that "counterintuitive" would be a massive understatement for sure. I'm sure the current mod team has enough trouble dealing with him, I couldn't imagine someone subjecting themselves to an endless game of whack-a-mole for hours on end every day. Lucifrost said: I oppose your suggestion on the basis that slurs themselves are vulgar. There are absolutely people who are offended by the use of "fuck" in a review; the word wouldn't be considered vulgar otherwise. I don't think their feelings are any less worthy of consideration than yours. You are correct in saying that slurs themselves are vulgar and also correct in saying that there are people who do not enjoy reading a review sprinkled with f-bombs and whatever other vulgar words you could imagine in a review. But at the same time, there is a clear distinction between a word which is obscene because it is socially inappropriate to use, such as "fuck", in most contexts; from racial slurs such as the n-word or other types of slurs such as the bundle of sticks. Reviews are expected to follow the review guidelines and have strict etiquette to them, or at least they are supposed to follow those guidelines. Within those very review guidelines, Kineta stated to refrain from: "...using excessive profanity and obscenities. If it contains "crap" every three words, it is no longer a review and more a rant." Anything which is deemed excessive could be manually removed by a review moderator with ease, provided that the review is seen by the review moderator or reported by a user such as yourself or myself. I object on a filter system deleting reviews with just profanity and not slurs on the basis that the system would flag the review as containing flagged language on a singular presence of a flagged word. It wouldn't be fair for a reviewer if their review contained one curse word and it got deleted, especially if they typed it off the top of their head in the post window as I did with this forum post. It would not set a good precedent if it removed all vulgar language such as common curse words because it would undermine the review guidelines and have the potential to be an overreach in moderation. I do understand your point of view, but I also did not place my feelings or my point of view on a higher pedestal than yours or the feelings/views of others. When I wrote my post and thought of the idea last night, I did try to take into consideration the potential for over-moderation and errors. That's why I thought that slurs would be the most suitable line in the sand to draw where the problem would be addressed. I think most people can all agree that slurs are objectively bad, but not everyone may agree that common curse words are bad. Anyways, I do appreciate the replies that both of you have written to my suggestion. They're both well written and they both expand the depth of discussions around my suggestion. |
| It's stupid to have an entire website of supposedly millions of users bend to one user Besides a lot of words contain words within them that would give false positives. Also super easy to get around with 1337 speak and special characters |
traedAug 21, 12:43 PM
traed said: It's stupid to have an entire website of supposedly millions of users bend to one user Besides a lot of words contain words within them that would give false positives. Also super easy to get around with 1337 speak and special characters Ishinashi himself is one user, but his trolling is incredibly easy to replicate by another user, while at the same time also to fix with a filter. This filter would allow the community and review moderators who are responsible to moderate the reviews to focus on less obvious rule breaks. I wouldn't say that such a fix bends the website to the will of a single user, because in the current state, a single user can disrupt the website to the point where everyone on the home page can see the flagrant reviews. I'd argue that the fix corrects the status quo where Ishinashi can manipulate the front page into displaying his homophobic ramblings. As with what you said about false positives and words which contain false positives, well what you said does have merit depending on how the filtration is programmed. If it were to flag words based on a partial match, then there would be lots of instances with false positives. An exact match would work better. Bypassing the filter would have to be seen as an "acceptable inevitability"; yes, if something like this were to be implemented, there would be bypasses with l33t speak or special characters as you mentioned. My counterpoint to this would be that the staff could either: 1. Disallow special characters in review (unlikely, because some show/manga titles do consist of special characters) 2. Tweak the system to flag discovered bypasses (if this system were incorporated, this would be more likely to avoid FPs, but it would take more time) I don't think that my idea is a perfect idea, but I also do not think that my idea is a stupid idea either. I think there is merit in a system like this, but the points which you raised about bypasses are valid concerns. The current status quo should not be maintained however, regardless of if its Ishinashi doing the trolling or someone else. Edit: With a filter system, this wouldn't even be on the front page of the website for the past half hour.  |
DulonkkAug 21, 1:32 PM
| You put way too much thought into something that the mods will never read or implement anyway. Easiest solution is to just ignore it. |
| To be honest, I find it weird that there is no automatic filter on MAL that would prevent users from using certain words (like the most "popular" slurs used on the Internet). Filter for both reviews, forum, recommendations, and so on. ateks said: True, lol. You put way too much thought into something that the mods will never read or implement anyway. Easiest solution is to just ignore it. |
| Too bad submitting changes to "About Me" section on the profil page seems to be still malfunctioning... |
| It wouldbe good if: There was a system that detected the same sentence or characters used in a review. Like copypasting the same stuff once and once again. There was also a system that restricted you the release of reviews to one each certain amount of time, like, they just copypaste from whatever program they saved the review from and post lots of reviews in one minute. There was a system that detected if a review is the same as another, and if they are the same as a deleted one. And if there was a system that restrained you from posting/commenting whithin the first hours of account creation. But well, people working on this site don't even care to make it secure, not only talking about this case in particular but it's SO FREAKING EASY to create an account for scams, really, I have seen like 10 scam accounts in less than one year here. I am not blaming mods but rather the higher parts who are the ones deciding what changes to put in the site. |
| "Kyu" - Penguin #2. The more I use this site the more I realize this site is terrible. |
ateks said: You put way too much thought into something that the mods will never read or implement anyway. Easiest solution is to just ignore it. I might be a bit less pessimistic than you are when it comes to this specific issue, but then again, your take is likely because you have used MAL far longer than I have. It's easy to say that it can be ignored, and I'm certain that most people don't even realize that Ishinashi does what he does with the reviews, however I believe that my suggestion could change his capability of doing this. If it generates enough of a discussion to the point where it continues to be seen, it will reach the staff team eventually, and from there it might be discussed internally. I have faith, especially after @Fleeting_Dream replied to me stating that a spam filter system does or did exist on parts of the website. As for the part about putting way too much thought into this, true. But I enjoy overthinking things if the result is an idea with substance behind it. Adnash said: I find it weird that there is no automatic filter on MAL that would prevent users from using certain words (like the most "popular" slurs used on the Internet). Filter for both reviews, forum, recommendations, and so on. Well apparently, MAL did have a spam filter at some point and it still might have a spam filter. However, I have no idea how it actually flags things as spam or if its even functional. If the filter still works, as Fleeting_Dream said earlier in the thread, it likely wouldn't be too hard to tweak it to blacklist common slurs. That is, of course, if the feature still exists today. I think what you said about forums and recommendations would be interesting. The feature could also work on interest stacks as well, but I think the best place to start testing this would be the reviews and the recommendations as well. eblf2013 said: It would be good if: There was a system that detected the same sentence or characters used in a review. Like copy pasting the same stuff once and once again. There was also a system that restricted you the release of reviews to one each certain amount of time, like, they just copy paste from whatever program they saved the review from and post lots of reviews in one minute. There was a system that detected if a review is the same as another, and if they are the same as a deleted one. And if there was a system that restrained you from posting/commenting within the first hours of account creation. I think the idea of my "slur filter" should have its scope broadened into just a filter for reviews in general, because there are also spam reviews (not sure if Ishinashi makes them or not) that do this exact same thing that you mentioned with the constant copy-paste repetition in the review. This would likely be easier to implement than an exact-match slur flagging system, but I think that it would also be prone to false positives if it was not properly implemented. I've included an attached example at the end of this post. I also liked your other ideas with the reviews. Rate limiting reviews is also a good idea in my opinion as it would not only deter Ishinashi's spam, but it would increase the general quality of reviews from people who post lots of reviews en masse because it would force them to wait and proofread their work, or just wait. Granted, this would not affect most people, just trolls like Ishinashi and people who post lots of reviews at once. The account restrictions within the first hour would work well and I think it would be a great way to slow down Ishinashi or other trolls without a drastic solution such as 2FA. These three ideas I really like and I think they are realistic enough to implement. The plagiarism checker is a good idea which I like, but it would not work on MAL. The review moderators and the database contributors themselves are already overextended dealing with their respective issues of Ishinashi issues or database backlogs. A detection system to check whether or not a review is a match of another review previously posted would require scraping over thousands of reviews and placing them within a database manually. The only way this could ever actually be implemented is if either: 1. Old reviews were grandfathered and exempted from this data collection, which would be unlikely if such a system was implemented 2. The system was implemented already at the website's inception, but unfortunately CERN has not discovered the secrets of time travel and neither has the legendary mad scientist Houhouin Kyouma But I felt that your other suggestions were far more realistic and feasible to introduce. Anyways, I appreciate all of the replies left today. I would've replied sooner, but Applebees unlimited boneless wings put me into a self-induced coma last night. I thankfully recovered after eating 19 nuggets made of monkey lips and manatee flippers, though I thought their sauces were delicious.  |
| @Dulonkk Regardless of that I think an automatic slur filter is pretty useless anyway. If faggot is banned he's just gonna write fäggot, faggøt, fa99ot, fæggot, fagg0t or faggót etc. There are nearly infinite ways how to write it and even more if you use unicode characters so how are you gonna implement that? There are some other ways how to reduce the spam but again, why discuss about it if nobody who is responsible cares anyway? Ishinashi is just one more hole in this sinking ship. |
ateks said: I mean, it isn't impossible to implement a system that can deal with bypasses. There are systems on other applications and websites which will delete words that match ~80% of the characters or so within the word, but because of false positives, I would not take this course of action. I mean, imagine if you wanted to talk about Studio Trigger, and well, this hypothetical system flagged you for the n-word, how would you feel? That wouldn't be the way to go about this. As I mentioned in my reply to traed, the staff could either disallow special characters in reviews entirely, which I find unlikely, or they could manually blacklist any discovered bypasses over time, which I think is far more reasonable.If faggot is banned he's just gonna write fäggot, faggøt, fa99ot, fæggot, fagg0t or faggót etc. There are nearly infinite ways how to write it and even more if you use unicode characters so how are you gonna implement that? Think of the filter as a levee rather than a replacement. Its purpose would be to automatically delete things which it could detect, like how a levee is supposed to stop parts of a river from overflowing. But what happens if a storm blows by and the water floods? Well, water will go over the levee and the people who live around the river as well as their government has to clean up the damages. So if people start bypassing the filter, then the review mods will have to deal with reports or delete bad reviews and discovered bypasses will have to be added to the filter over time. My idea of the system is honestly very similar to the discord bot Dyno and its detection system. You can add or remove as many words/bypasses to its filter or configure it to detect partial matches, and if detected, the offending message would be deleted and logged. Granted, I don't know any of the coding behind it as I'm not a developer, but such a system would be feasible to code on MAL and it would be a first for anime database sites. |
| I don't think such a system would stop him. It's easy to bypass a simple word filter, and since it's an automated system, you could just trial and error your way through it. On top of that, I feel like such a change would only make him more active. He'll double down and rant about "heterophobia", "freedom of speech", or something. I think a better solution would be to make a discord bot that scrapes the recent reviews page every 10 minutes and searches for specific strings of text that show up in his reviews. If it finds anything, it pings the mods, and the mods can nuke his account. Sure, it's still just whack-a-mole, but it's more efficient than what we have now. This would force him to either write new unique reviews or figure out when the mods are asleep and post at those times. |
 |
fluffycow17 said: I don't think such a system would stop him. It's easy to bypass a simple word filter, and since it's an automated system, you could just trial and error your way through it. On top of that, I feel like such a change would only make him more active. He'll double down and rant about "heterophobia", "freedom of speech", or something. I think a better solution would be to make a discord bot that scrapes the recent reviews page every 10 minutes and searches for specific strings of text that show up in his reviews. If it finds anything, it pings the mods, and the mods can nuke his account. Sure, it's still just whack-a-mole, but it's more efficient than what we have now. This would force him to either write new unique reviews or figure out when the mods are asleep and post at those times. My goal in writing my post up was to figure out a way to clean the front page efficiently and slow him down. I don't think that Ishinashi will ever stop unless it becomes too difficult to troll at the volume which he does. While I still think that a simple filter would have its merits in deleting anything obvious at first (though as I've mentioned before, any bypasses would have to manually be added), your idea to use a Discord bot is an excellent idea to deal with Ishinashi specifically. It would be dependent on the response times and activity of the moderators, but the bot seems far easier to program with Python than a filter which would have to be constantly tweaked just to catch him. I still think that they should go with a simple filter because it would stop a lot of imitators down the line, but the bot idea would be a far more cost effective solution in dealing with Ishinashi specifically. It would work a lot better in the long-run as well. So, great suggestion. You should consider posting your idea as a separate post in suggestions, though I'm certain that the staff will see our ideas soon anyways. |
| I mean, yes this might be helpful. |
 |
More topics from this board
» Anime source ORIGINAL database or linkkiraAbdul - Apr 2, 2018 |
7 |
by Pusquin
»»
1 hour ago |
|
» Please Add Producer Column in Anime ListswHaleJoey - 5 hours ago |
0 |
by wHaleJoey
»»
5 hours ago |
|
» Add Visual Novel External LinksDeago - May 31, 2021 |
8 |
by IridescentJaune
»»
7 hours ago |
|
» AchievementsRobertBobert - Today |
0 |
by RobertBobert
»»
Today, 8:18 AM |
|
» Add dislike botton or disagree botton in reviews.Orion911 - Today |
0 |
by Orion911
»»
Today, 1:13 AM |