Tech.LGBT

Logo

Information Site for the tech.lgbt Fediverse Instance

13 October 2023

Retrospective On thebad.space Situation

by tech.lgbt Moderation Team

While our latest post has responded to the fediblock post from simcha.lgbt (to load, put the link in the search bar of your instance), more recently, mastodon.art has announced to fediblock tech.lgbt on the same grounds, including additional reasons.

In our post, we have talked more broadly about the issues that were brought up in rage.love’s statement, in this post we’ll be responding to the events that led up to weirder.earth’s fediblock post, our response and reassessment, and we’ll respond and do the same with the point noted by mastodon.art about the events surrounding thebad.space, now that we have reached consensus.

The latter, including fatigue and smidges of burnout over this situation, contributed to the delay of this post. We apologize for that.

Weirder.earth’s fediblock

The simcha.lgbt fediblock post mentions the fediblock post from weirder.earth, of which we have touched more upon the responses in our last post. In this section, we’ll talk about the events leading up to these responses, and explain the events surrounding each point made.

Due to the nature of the chronology, we’ll be responding to each point out-of-order, starting with the second.

The second point

In another incident, one of our community members was suspended by tech.lgbt for calling out the racism of one of their users.

Back in January of this year, a tech.lgbt user engaged in a discussion about AAVE words and cultural appropriation. The user made a statement apathetic to the trend. Hours later, a user from weirder.earth (who has since closed their account) engaged in hostility to the statement from the tech.lgbt user. They sent a report and, within several minutes, posted a fediblock to publicly call out the user for racism with screenshots. This led to multiple reports being sent to tech.lgbt about this user. The moderation team saw these reports and determined that the statement did not intend to cause harm despite it potentially being unconsciously racist, and resolved them. This decision included our POC members. This was not the right call. We have since adjusted our review process to include the criteria of apathy towards racism as enabling racism, and check for this at multiple points in the assessment.

The first point

In one incident, a member of their mod team made a “revenge report” claiming they were getting spammed with pro-capitalism reports. This was 3 minutes after tech.lgbt received a single forwarded report for antisemitism. We felt that this was an extremely bad knee-jerk reaction, and it is disgustingly antisemitic to decide anyone reporting antisemitism means they are pro-capitalism. We did reach out to their admin to privately let them know that perhaps they need to review things with their mod team etc and to clarify the situation, but that did not lead to anything.

A post from from one of our moderators was taken out of context, and was reported as being anti-Semitic by a weirder.earth user. Specifically, the post read “gender neutral language suggestion: instead of ‘landlord’ or ‘landlady’, consider ‘parasite’.” The context being that they were referencing the capitalist hellscape we live in, where landlords/landladies own homes, rent them out at exorbitant pricing, while contributing none of that profit back to society. (It is a practice in business that has accelerated housing issues in many places in the modern age and is a sad trend.) Our users are generally aligned on such topics and understood the context of the post as simply a joke. Additionally, the mod added another post reading “similarly, don’t use ‘policeman’ or ‘policewoman’. consider a more inclusive term, such as ‘pig’ or ‘bastard’” only a few minutes later (referencing ACAB sentiment). It is clear to understand these as anti-authoritarian jokes and one needs to willfully misinterpret the intent at this point to claim that this was anti-Semitic in nature, and our Jewish mod at the time also found no fault with the post.

The report was issued by a weirder.earth user who did such willful misinterpretation after engaging in hostility in the thread, espousing their own strict misrepresentation that it could only be anti-Semitic in nature by conflating the term “parasite” with the harmful stereotype “Jewish parasite”. Our team at the time saw this report and dismissed it as it contained no details of any kind, and the only rule stated to be broken was listed as “Other”. Furthermore, the report came from an anonymous server admin account and not from the server itself as normally occurs with forwarded reports. The mod who noticed this concluded the report was made in bad faith as a means to assert a pro-capitalist stance and that they abused the anonymized nature of the server account, and the mod reported this server account alleging this claim. This assumption and subsequent action were not correct, and our reaction prompted a discussion between both server admin teams. We failed to respond in good faith with their communications at the time, taking several days between responses.

There was also a dispute between one of our staff members and the same user on weirder.earth from a previous fediblock callout post (discussed in the response to the second point), who had also previously harassed this staff member, which exacerbated the issue further. This is not how moderator staff should handle moderation conflicts.

In response to this issue, we have since created official instance communications via the @mods account and the email, mods@tech.lgbt, to prevent future miscommunications with other instance teams, and our staff are not permitted to directly engage in active controversies and persons that were reported. Each staff ember is entitled to their own views and can air them, but this restriction aims to stall controversies from spreading to individual moderators, as their position of authority could be viewed as the perspective of the entire moderation team.

The Third Point

The final separate incident was a tech.lgbt mod (who has since left) “revenge reporting” a person for calling out the coddling of Nazis.

The context of this issue was the appearance of an Ukrainian Mastodon instance being actively hijacked during an absence of admins. There, apparent, fake Nazi accounts were created and began mass following users outside of the instance while pretending to be Ukrainian citizens.

A tech.lgbt user noted and warned about this, and later edited their message to include “#fediblock”.

This post then received a report saying “Piggybacking on a good fediblock rec with a bunch of unhinged Russiagate bullshit”.

At the same time, a then-mod from eldritch.cafe posted about the fact that this Ukrainian mastodon instance was, in reality, a fascist instance.

Links to both of these (the report and the post from eldritch.cafe) were posted in our internal chat quite close to each other.

Another then-moderator of ours did not properly differentiate between these different posts, confusing the linked post to be from the report. They later said that the mention of “russiagate” made them “see red”, and effectively tunnel-visioned into anger, and reported the linked post (from eldritch.cafe) for reasons wholly unrelated to their actual contents. This was not the correct response, not only that a single moderator performed this action without consensus from other moderators, but also because emotions took over rational and reasonable judgement.

Another moderator (the one who posted the two links in the internal chat) apologized in their stead, and tried to explain the situation. We have since enacted checks in the review process to prevent further mistakes like this. The first moderator who made the wrong judgement call and took improper actions stepped down over the incident.

We had, at the time, shared correspondence with admins from other instances who inquired about these events, and explained our position. Weirder.earth was among those we sent explanations, but they have not replied to our email.

Mastodon.art was one of the servers that asked asked about our situation and wanted to confirm if the allegations had merit. We addressed them as best we could at the time and this became the basis of our restructuring process, which was detailed a bit in our previous blog post.

Mastodon.art enthusiastically responded to our explanation, which is able to be observed here.

We note that mastodon.art’s most recent fediblock (made in the last month) still links to the same accusations.

Resulting Moderation Amendments

Resultant of these actions, and as explained in our last blog post;

We are still not perfect, but we have improved efficiency and accuracy greatly in these past 7 months, and will continue to improve in time. This effort allowed most of our connections who were announcing the severance of ties to rescind those actions, including mastodon.art.

To explain the recent event with the abuse of our reporting system to harass a user on rage.love again; Our user had no prior history of harassment or offenses, and used our reporting system to demean and harass a user and admin of rage.love. This included an element of racism, and subsequently this became racist in nature. We received an inquiry about the report, as it was forwarded to rage.love, and we responded in agreement that it was an abuse of the reporting system and that we will issue our form of consequence to the user. Rage.love understood this and wanted to confirm the user’s identity to issue a fediblock warning on them. We understood from our previous experiences about the impact of a fediblock on a single user, and we did not want to subject our user to the same harm of public ridicule given that the user had only this instance of such harassment on record. We failed to consider the perspective that protecting the user will empower them to harass BIPOC users again in the future, and in this instance it was a mistake. Our efforts to rehabilitate the user failed after they continued to bring up the situation publicly and try to justify their actions without truly understanding their fault or the racist nature of their report.

Now we must be clear. We summarily denounce racism in all forms, we denounce Antisemitism, we denounce harassment, and we denounce abusing our systems to perpetuate any of these offenses. We deem the nature of all issues of the weirder.earth fediblock post to now have been addressed fully. We acknowledged our failures, learned their nature, and structurally reorganized our processes to resolve them.

While we have much more principled moderation now, we are still not done improving.

Recent Events

Understand that when we say that moderation is not easy, we truly mean it. Many times, there are nuances that one person alone cannot see, and sometimes stress, bias, and urgency can cloud judgement. Facts can be hard to verify and connections to other things may not be easily recognized. We do not blame anyone in this position for having guilt for a wrong decision, and we truly believe it is possible to learn from mistakes. We take our hard learned lessons and do better next time.

It may take time and a cool head to come to a decision, especially when we need to come to consensus on a serious action, but as a result, it is not made haphazardly at the behest of one person.

This is why our individual moderator public accounts do not bear weight on tech.lgbt as a whole in regards to our moderation, our individual actions are only representative of our individual selves and views, and we are still held accountable to the rules of our own server.

This all leads to the recent situation that has needed time to be thoroughly reviewed to be properly addressed. One of our moderators, @shadowjonathan, has been alleged of starting a racist harassment campaign against the owner of thebad.space (TBS), so in the effort of integrity in this thorough investigation, we corroborate the materials that Jo provided, such as screenshots or reference material (links), to public statements and other materials that can be brought up by other moderators. Jo was otherwise not involved in the original drafting of this document, only reviewing for confirmation of details before posting.

While we wish to keep the sanctity of privacy in private matters, we feel that we must provide certain evidence to substantiate our findings in this investigation, due to the severity of the allegations.

We have compiled a timeline of events, this timeline was moved to the bottom of the page, to keep the blog post from becoming unwieldy to read. We suggest readers to first go through it, before returning to this paragraph, as the next sections will provide our responses to these events.

Our assessment

It’s important to first acknowledge the context of these events: the fediverse has a problem of systemic racism and white supremacy, reflecting the same issues in the US and Europe. Even white, anti-racist instance moderators who actively block Nazis and overt racists, have difficulty recognizing and acting upon covert racism and dog whistles. This has created an atmosphere that supports racist harassment, where white instance moderators fail to recognize the severity of covert racism and act upon it. This is evidenced by the prior harassment campaign against the PlayVicious instance.

We see that, while we are not a wholly white moderation team, tech.lgbt’s white moderators are included in this recognition failure, and the issue is not exclusive to any single portion of the fediverse.

Controversy regarding thebad.space and racist harassment

Our summary of the events is as thus;

A bug in thebad.space incorrectly listed multiple servers, either limited or suspended by single sources, without additional corroborating sources, on the official blocklist after the addition of new servers to the trusted sources.

The effect of this bug was discovered by several admins and users hours prior to @shadowjonathan becoming aware. Upon discovery, @shadowjonathan spent approximately 15 minutes trying to find information that was not present on thebad.space site and felt the effect of this bug (unaware that it was a bug) pointed to a larger problem with shared blocklists in general, whereby large adoption of blocklists experiencing major technical errors could negatively impact federation in the long term.

Therefore, they decided to make a public post to raise awareness of the issues they had found while researching the site. @shadowjonathan was unaware that thebad.space was a project developed by Ro, a black fediverse user who had been the victim of a previous harassment campaign. They did not know that their post would add to the preexisting racist atmosphere in the fediverse that would result in a new harassment campaign, and they missed various critical but obscure details that would have led them to see that being a potential outcome.

Shortly after making the public post, they decided to start a discussion about the project in the fedimins channel of Mastodon.art’s discord.

People in the fedimins channel who were participating in the project and fully aware of its history did not acknowledge or understand that @shadowjonathan was unaware of its history, nor did they help acknowledge the critical lack of pertinent information on the site itself.

After Ro (@Are0h@ubiqueros.com) read through @shadowjonathan’s criticisms, he did not consider it to be a good-faith awareness post, but rather a defamation attempt. This is understandable because, regardless of @shadowjonathan’s good intentions, Ro most likely anticipated a wave of racist harassment against him as a result. This discussion led to a private discussion in the fedimins group which was then brought to public attention by Ro and several other admins in the fedimins channel who vaguely referenced @shadowjonathan and tech.lgbt as a whole.

After this, @shadowjonathan made an earnest apology to Ro in private, and rectified their public post, as they now understood the context around thebad.space, and had learned to understand various other ways that racism complicates these interactions, and about assuming bad-faith while trauma may be a prominent factor in others’ actions.

@shadowjonathan also offered to help solve the technical shortcomings they had found in the site.

The apology was half-accepted by Ro, and technical assistance was rejected.

Rectifying the thread within the hour of issuing the apology, as well as offering assistance again days later during the DDoS attack on thebad.space site, shows understanding on @shadowjonathan’s part of the situation, and was a meaningful attempt to make amends.

However, we see that @shadowjonathan’s post was not the origin of the speculations, as initial public posts were what informed them, and, given their (@shadowjonathan’s) post had limited reach, with fewer than 158 boosts and 207 favorites at the time (these numbers were only first recorded on 16 Sept 2023; the specific window of 28 hours, 20 minutes between the original publication and the rectification is likely much lower), we deem it unlikely that the majority or even moderate impact can be attributed to @shadowjonathan directly, though we acknowledge that their part in this situation was not nil.

We state this, as many allegations made by Ro, WelshPixie and admins of the fedimins group afterwards, accuse tech.lgbt of being the sole instigator of the following fallout. (In referring to tech.lgbt, we assume they would prominently be referring to @shadowjonathan.)

While we don’t believe that @shadowjonathan was the origin of the speculation or had any intent of malice, their good reputation meant that their post unwittingly provided additional credibility for a racist portion of the fediverse who were spurred on to enact a racist harassment campaign against Ro and thebad.space. This is not to say that criticizing thebad.space or other blocklists is inherently racist, as good-faith discussion is necessary to improve and balance them, but criticizing these topics unfortunately became ample grounds for racist fediverse users to organize against, especially considering the history Ro already had with these groups.

This is not to say that Ro’s own statements and actions have not contributed to this fallout. We note that, on multiple occasions, Ro has posted escalating language about tech.lgbt and @shadowjonathan that increasingly diverge from fact, not dissimilar to inflammation, while rejecting criticisms, even in cases where the person was attempting to communicate in good faith, seemingly ignoring the perspective or background of the persons making the statement.

We see this as a potential angle of continued harassment, as the language that Ro uses also leads to further harassment and racism. We condemn users of the wider fediverse for using said angle to further harassment and racism.

While investigating these claims and getting to the root cause of this incident, one of our moderators, @sudaksis, asked WelshPixie (@welshpixie@mastodon.art) in private for a transcript of the channel at the time.

Our good faith attempt at conversation and remediation, through @sudaksis’s communications with the fedimin group and Ro, through WelshPixie in private, was halfway accepted only in the form of a redacted transcript of @shadowjonathan’s statements during that conversation. Having discovered that it had been altered, with intentional bias, by WelshPixie, we can only confirm that @shadowjonathan’s words, confirmed in accordance with WelshPixie, are in whole truth. All other statements in the transcript made by others, including WelshPixie’s verbatim recorded statements, cannot be confirmed or denied in any meaningful way. This act on WelshPixie’s part damages the trust tech.lgbt has placed in good faith communications with mastodon.art and subsequently the fedimin group as a whole.

None of the statements made by @shadowjonathan are determined to be provably intentionally racist, nor are they definitively proven unintentionally racist, given they went out of their way to reduce the reach after they became aware of the developing situation. We also do not think there was intentional racism as alleged in the fedimin discord channel, as their interpretation of the discussion at hand (given their panicked and possibly tunnel-visioned mental state) was one of airing and acknowledging concerns, together with the subsequent lack of resolution of those concerns, resulted in a dire miscommunication, one which was not intended to be had in the first place.

Tech.lgbt has made no official statements about the situation in any capacity other than referencing this upcoming response in the 15 Sept 2023 blog entry regarding the simcha.lgbt allegations. Any comments from tech.lgbt moderators before this publication are not official statements that speak for the whole moderation team.

Our Response

First and foremost, we, the moderation team of tech.lgbt, fully understand and empathize with the mission to make the Fediverse safer for BIPOC communities. There is absolutely a racism problem and it must be addressed. We believe that there are ways we all can make our spaces safer for all people, either through principled moderation practices or verified tools for the Fediverse that can prevent harm.

It is entirely understandable that past issues on the fediverse have caused severe harm, particularly in regard to the PlayVicious incident, which is likely why Ro believed that @shadowjonathan’s post was an attack on his project and on his character, regardless of its contents.

The Oliphant blocklist was only made known to us from the request for our active participation by WelshPixie in July, to which tech.lgbt declined.

It was the lack of knowledge of the thebad.space project and Ro, with the lack of available information about the list on the site itself, that coalesced into unfettered panic that resulted in the awareness post within those few 15 minutes.

What happened in the fedimins channel was that this fear reaction was understandably mischaracterized as racism by Ro, and intentionally perpetuated by the admins within the group.

When those admins found out that @shadowjonathan had a limited understanding of the history of thebad.space and PlayVicious, they did not offer support to resolve the root concerns for negative trends that an unverified blocklist could perpetuate, and instead ridiculed them for not knowing the history and context of this blocklist.

There was no conflict de-escalation provided in the discussion and only pressure was applied to @shadowjonathan into acceptance of their arguments. What happened here because of those other admins in particular was inappropriate and should be condemned.

We, tech.lgbt, do not condone any harassment that has been directed towards Ro and others involved, including the DDoS attack that took down the list site for several days. That is a shame on our communities, and we all should be doing better. We also do not condone any conspiracy theories about Ro or thebad.space project. We do not support any claims that thebad.space was designed with the intent of harming the trans community or to further any political element, nor do we believe there was willful misconduct on part of Ro or @shadowjonathan prior to the attempt to make amends on 11 Sept 2023.

We can all be reactive when put under scrutiny, and we do not fault Ro or @shadowjonathan for having reacted the initial ways that they did. What we all can do is ask for forgiveness for our mistakes and make amends.

We believe that @shadowjonathan has given adequate reflection, retraction, and attempted to make amends. Learning comes from experience.


As a repetition of one of our points, and to be entirely clear, the tech.lgbt moderation team as individuals do not constitute instance-wide support or condemnation of any projects or people. All official responses from the team as a whole, though rigorous procedure, will be made from the @mods account and/or listed on our blog page.

What we all need now most of all is to finally put aside our egos and stop the name-calling, harassment, and dishonesty. We need to self-reflect, admit our faults, and work to rebuild our trust in each other. Not these callouts for perceived sleight and unproven claims of poor character and misconduct. The only way to heal is to treat each other with empathy and respect; to understand each other as equals. We have the power to make the Fediverse work for ourselves and our communities, and it is our responsibility as moderators to do so with kindness and trust.

Do not use this response to further any harassment to Ro, to WelshPixie, to the fedimins group, or anyone else who has alleged misconduct to the tech.lgbt mod team over this event. We can have discussions of disagreements and concerns in ways that do not require name-calling, ad-hominem attacks, and most important of all, without bigotry or racism.

We will continue to monitor the situation, but we hope that this outline of our view of this situation will help members of the fediverse gain some more detailed insight. We hope that, if this results in a discussion, that discussion is not a repeat of the fallout that occurred a month ago.

With this full detailing of the specifics of the incident, please use your best judgment to have rational conversations and make meaningful choices of how to invest your time on the Fediverse. Thank you for understanding.


The Timeline

Note, everything marked in italics are editorialized comments made by the tech.lgbt moderation team as a whole.

(Also note that, due to a change in how mastodon technically displays web pages in newer versions, we could not get mastodon pages to display properly on some archival websites. We’ve tried our best to capture these pages, but some digging in the website’s source code (of the archived web page, through “inspect element”) might be needed to look behind CW markers.)

One or a few days prior to 10 Sept 2023

10 Sept 2023

(Below minutes, except 13:39 EDT, are the notable recorded events from the transcript provided by WelshPixie until the time when @shadowjonathan is revoked access to the fedimin channels.)

11 Sept 2023

12 Sept 2023

13 Sept 2023

14 Sept 2023

15 Sept 2023

16 Sept 2023

17 Sept 2023

18 Sept 2023

19 Sept 2023

20 Sept 2023

21 Sept 2023

22 Sept 2023

Additional events outside of this timeline:

tags: