,
14 tweets,
18 min read
Read on Twitter
3/ I lodged a complaint with @TwitterSupport about @Botometer. I took a representative samplings of the users in our Community of Interest and many of them, who I know are HUMAN, are rated very high as "BOT" like. At the time of my assessment, I was in the mid 2's...
4/ Many users are starting to utilize Botometer tool & its assessments for their Twitter experiences. It's always good to be able to identify automated trolls on Twitter. After viewing results, I've determined the tool to be scary & unfair...
5/ Scary because their assessments could influence users to decided not to engage other users who score high. Scary also in the fact that the app accesses information about the Twitter habits of users. The APP does ask for permission logon using current user credentials...
6/ That permission should be LIMITED to that user's dataset alone. Instead, you can fetch Twitter habits of any user which is a violation of privacy. How is it that User X can give a random APP the authority to retrieve another user's datasets without that user's permission?...
7/ Twitter needs to investigate this and fix it immediately. Especially in light of the fact that the APP makes assessments about user's based on datasets it assessed without permission. I know the easy way out of this argument is to claim public profiles are public...
8/ However, users should be able to place a disclaimer on their profiles indicating that "bulk access" and use of their datasets is a violation of their privacy. And any APP encountering this disclaimer should comply!...
9/ What's unfair about this APP is that Bot scores are displayed on a 0-to-5 scale with zero being most human-like and five being the most bot-like. A score in the middle of the scale is a signal that the classifier is uncertain about the classification...
10/ So basically there are 3 outcomes:
Human;
Don't know;
BOT.
This just isn't a way to run a ship...
Human;
Don't know;
BOT.
This just isn't a way to run a ship...
11/ Most users more than likely will not look into the detailed breakdown of an assessment, and my fear is that they will make decisions about engaging with a user based on the summary assessment, without looking deeper into the Complete Automation Probability score...
12/ Although Twitter is a very good platform for establishing Communities of Interests, its most annoying characteristic is forcing users to PROVE that they're HUMAN, over & over again!
Maybe I don't completely understand Botometer, so I'll continue to read the documentation.
Maybe I don't completely understand Botometer, so I'll continue to read the documentation.
@Botometer @TwitterSupport @TwitterAPI @TaggartRehnn @BandieraNancy @AVestige1 @BJSquirrel @DinaSaS69 @SaneSophisticat @GomesChantel @LockeShiny @Carmenr1Carmen @DenbrotS @fenrir_71 @Mschade2020 @teenytinyflame @holachola @impinkestgirl @BeverlyBednarc1 FYI team; Botometer is responding to our concern.
@ginadem @crpswarrior1 @TheRealJTIII @SkiBuni @DemocracyJourno @iche_me @angelarlacy @sagemendo @joyrevels28 @media_auntie @DrTDogg @Chris_09765 @Nasty_Woman1 @MFrance923 @CrippledJerk @Chrisnotmypotus @ooshdesign @_WeStandUnited @stemeakabrunson FYI team; Botometer is responding to our concern.
