Bots gone wild
I've seen a bunch of posts this morning by people who are desperate to get off a blacklist being used by a bot called @cheetah.
Free image source: Pixabay
Those posts were by:
- @elenasteem -https://steemit.com/drawing/@elenasteem/forest-beauty-help-me-fix-the-problem-with-my-page-on-steemit)
- @andrianna - https://steemit.com/steemit/@andrianna/i-am-the-real-person-i-have-no-not-one-stolen-work-the-management-confirms-it-but-i-still-in-the-black-list)
- @btc-central - https://steemit.com/steemit/@btc-central/censorship-alive-and-kicking-on-steemit-thanks-to-broken-bots-anyx-cheetah)
To the best of my knowledge these are all legitimate users who have had their posts commented on by the bot saying it was plagiarism because the same content appears elsewhere on the web. Well sure - these users also post the same content elsewhere on the web on their own websites. After multiple flaggings by the bot and no response to their counterclaims the bot went into downvote mode. Now these people are doing things like posting photos of them holding the original artwork they created and had previously posted. Multiple begging posts have elicited no response from the bot, just more flagging.
I'm going to hazard a guess there are more instances where this bot has gotten it wrong, and I'm sure there are other bots like it doing similar things. This particular bot was created by a user called @anyx but they say it and its blacklist is now under control of @steemcleaners.
I'm feeling like it is poor form that this bot is being so opaque about it's operations and that people are waiting a long time to be removed from a blacklist. I would like to propose that there should be a Steemit code of ethics for running bots and especially for those that operate off any blacklist which impacts a user negatively.
- Bots should always clearly self identify in every post with direct links to profile, FAQ, dispute resolution info and data about the dispute - don't rely on people going to the profile page and digging this up for themselves.
- If the account of the bot is not the one operating it that should be disclosed too.
- No action like "ban" or "blacklist" should be taken without an explanation - a link to artifacts backing up the claim should be provided. If you have found ten posts with allegedly plagiarized content before blacklisting them then link to them somewhere in one place so someone (and third parties) can quickly and easily find what dirt you are slinging at them.
- A clear and simple dispute resolution policy should be established - this might use of an automated "Please add <unique keyword/metadata> to the site you own to prove it and let us know" or it might be a human feedback loop. Ultimately everyone should have recourse to a human to arbitrate if automated methods fail.
- Ban or blacklists and their claims should be transparent and open, this is not a secretive society or at least it shouldn't be IMO
- Disputes over actions should be resolved in public - have the bot make a post and then the "accused" can reply to that and third parties can see what is going on. We have a public immutable blockchain to record actions, lets use it.
- Bots that do not comply with this code should be flagged themselves and if they cannot address the dispute they should be removed from the system.
Another thing to think about is that sooner or later Steemit.com will get a block feature. It already needs it now IMO to deal with harassment but I don't know if it is on a roadmap yet. However at that point someone could just block a bot like @cheetah and render it impotent. That means we'd need a way to give them special privileges to evade a block (or they would have to operate directly on the blockchain to see all content since AFAIK there really is no way to implement "block" at the blockchain level). I think with those privileges would come a burden to act ethically, fairly and responsibly. At this point in time I'm really not convinced Cheetah or any other bot for that matter is there. The users mentioned above are really finding its actions to be more like arbitrary and capricious or at the very least just plain erroneous.
Thoughts?
PS. I am still a relative n00b on Steemit so I may have misread things badly - I hope not, I definitely think I'm just echoing the feelings of a beleaguered few. Remember one of the arguments about profiling systems in real life is that even if your system is 99% accurate with a large data set that can lead to an awful lot of false positives which can really make them very, very unpopular with "law abiding citizens".
It's interesting because plagiarism is a real problem - I have only been here a few days and I keep noticing how much of the stuff posted is off topic to the tags and either copy and paste, or really badly modified from the source.
The API and the $$$ incentivizes people to take short cuts like this, and obviously humans are doing a bad job of housekeeping it.
Having dolphins caught by the tuna nets though is also a big problem.
There needs to either be a way to appeal or a way to have the canonical tag applied to prevent false positives?
I definitely agree that plagiarism is a problem and based on the description contained in the Cheetah bot FAQ it was much, much worse. But your analogy of dolphins being caught in the tuna nets is a good one and pretty much describes what seems to be happening here - and no one is on deck to cut them out and release them in a timely manner. Those people who run a business know that it is the people who have a negative experience like that are the ones that scream the loudest - every negative review is much more impactful than a single positive one.
Really we need to make sure that false flags are resolved well within the 7-day earnings period and if there is some standardized automated process for vouching for the authenticity of your own content when you've reposted from another source then we should have clearly document processes and folks who are willing to help out the less technically able (I predict the vast majority of Steemit users could not figure out how to add DNS or HTML metadata to their sites). But most could probably manage to post their Steemit ID and a special keyword to a blog or web page.
Completely agree, if the experience leaves a bad taste then the community can't grow
Make Steemit Delicious Again!
On top of false flags and downvotes, I see many users on social media and blogs and etc outside of Steemit joining this site and trying to dual share their content here and get the cheetah comment that degrades the look of their post (especially when I see half the time the posts include source links already, thus really eliminating the use of the bot)
"Tired of wondering if YouTube will pull or de-monetize your vids? LBRY rewards you for mirroring your channel!"
This should be the attitude for Steemit users. I do not see Steemit or DTube as a LBRY competitor necessarily, but really a lot of users on LBRY and Steemit could and should be the same, and the attitude from those devs should be shared here as well that share your content on this network over others.
Disclaimer: I am just a bot trying to be helpful.
While I can't say I 100% agree with Cheetah, I do think it does more good than harm. The bot on its own does not start flagging, there is human guidance and in most cases a good bit of due diligence.
This article provides more detail on the andrianna situation for instance. It appears the claims against her do not center around plagiarism, but ties to a known scammer.
This can be the case with other blacklisted users as well, where withdrawal addresses and other information may reveal networks of accounts that may be targeted together by Cheetah.
We have some users feeling that cheetah does more harm than good, others saying it does more good than harm, and without a way to verify it's work using some lovely blockchain technology in some way, why is it still running in it's current form? Is there a way for Steemit users are able to vote on how or why it functions? With the cheetah off in the wild by itself is this the start of the WWIII Elon is warning us about?
The original poster made some really good suggestions that if majority feel this bot needs to run, it can be held accountable for it's actions at the very least. That link you provided was great, very interesting situations with andrianna in particular, hope Sherlock goes into the links in the post here. I have been here just about a year and seen that cheetah go through positive changes and still the results are pretty shitty imo, I've requested it be renamed at the least, but I'm sure the amount of money that account in particular earns why change it. I mean, as much as I see people complain about it, I also see users who thank it after it comments on posts lol.
I saw your mention of renaming it, and it's funny... I'd never once thought of the Cheetah-Cheater connection! I'd only assumed it to be named for the speed it scans/indexes web content so never got that negative connotation.
And no, there's no way for users to vote on the functioning of Cheetah because it is a private user itself. We have the same options we always do, upvoting and downvoting. Andrianna has been getting better payouts since being blacklisted it seems! The community is supporting her to negate Cheetahs flags. The community can in turn flag Cheetah at will if it's felt it is having a negative impact.
I do believe the intent and motivation behind Cheetah is good, because it would certainly be more profitable for its owner if it did not flag at all.
Great link to the post abou @andrianna's account there @bryan-imhoff - very curious indeed. But I think I rather side with the comment on it by @sminchow which says "She should be un-banned, innocent until proven guilty".
The bot hasn't gotten it wrong. People need to learn that plagiarism isn't acceptable here, and there are repercussions for their actions. Nobody gets "banned" by this bot without a warning. They were warned and clearly refused to correct the behavior.
Actually if you look at these cases it was people posting their own content. In one case it seems there was confusion about who owned the account (it was apparently purchased from a scammer) not the content was always original and belonging to the poster.
There is also a difference between a cheetah message and being on cheetah's ban list. The message doesn't come along with flags, and it doesn't necessarily mean plagiarism.
hi edrivegom, good post...i am new in steemit...i am following you, u can follow, vote & comment on my post.
Thanks & Regards,
@rrnayak