We are aware of an ongoing wave of subscription spam throughout the fediverse.

We are investigating avenues to mitigate this, but they will likely involve deep changes in the sign-up flow that may break sign-up from existing applications.

In the meantime, you can find helpful resources at about.iftas.org/2026/09/11/boo

#MastoAdmin

about.iftas.org/2026/09/11/boom-protocol-spam/
en

Replying to @⁨MastodonEngineering@mastodon.social⁩

@MastodonEngineering

Please for the love of God put the captcha before the first email is send.

And please also make it configurable to put a manual approval before the fist email is send, since captchas are circumvented today for fractions of a cent.

It would be great if the signup process generated a random verification code that users must write on paper and photograph. This would allow us to verify that accounts are not created by AIs, residential proxies, or malicious bots in general.

Replying to @⁨pauamma@mstdn.social⁩

@pauamma @MastodonEngineering

Please read lgbtqia.space/@m/1173207135827

Please provide a alternative that makes AI signups more detectable or keep your criticism. Criticism is cheap and does mean nothing if you cannot come up with a single solution instead.

As it looks currently mastodon signups don't work at all because you cannot differentiate between a human and a bot.
And this is the most terrible solution.

✨ Mia’s Simulacrum 🏳️‍⚧️✨@m@lgbtqia.space

@shanie @MastodonEngineering

First of all it increase cost.
And than you can detect these AI images with quite high precision and reliability.

This is not about the numbers on the paper, it is about the pixel data itself, the code is just a tool to make it impossible to use a single image over and over again

Replying to @⁨m@lgbtqia.space⁩

@m @pauamma @MastodonEngineering for what it's worth, disabled people being frustrated and hampered by capcha is a fairly common thing, I find :cat_frown: another common concern I've heard from other disabled people is that our accessibility needs are ignored and requests for accessibility treated as nuisances :blob_frown:

I know admins are getting swamped and having to do even more with than they normally do, but it would be sad to see people rush to solve it by tromping over the needs of disabled people as happens far too often :kitty_sad:

Replying to @⁨m@lgbtqia.space⁩

@m @MastodonEngineering I'm confused. Do you not believe that AIs can write numbers on a piece of paper? I believe the opposition is a bit underestimated.

We'll have to be a little more innovative than that.

A piece of lined blue paper with the number "43455" scrawled on it with pencil. The image is overblown from flash. It's AI generated. — Open image attachment ⁨1⁩ of ⁨1⁩
A piece of lined blue paper with the number "43455" scrawled on it with pencil. The image is overblown from flash. It's AI generated.

Replying to @⁨samir@m.fedica.com⁩

@samir @gunchleoc l'll offer a few things in response.

I want an open social web. More friction is bad friction.

Many people expect "the app". They live on their phones. They are mobile-first.

There are a ton of industry standard mitigations that can be applied. Spam is an old problem. Most of this can be mitigated with a mix of time and money. We ask a lot from a very small, overworked team struggling to fund itself.

#GoodNotPerfect needs to be more acceptable to the community at large.

Replying to @⁨MastodonEngineering@mastodon.social⁩

@MastodonEngineering Thanks for raising this, we're dealing with the same wave of spam signups on our side. We tested a soft captcha at registration, but it doesn't trigger reliably every time, so we still end up manually blocking somewhere between one and five accounts a day regardless. We also tried a hard captcha on every signup, it does cut down spam, but it noticeably discourages real people from registering too, so it's not really a clean tradeoff either.

What we're running now is a different approach on the server side, a local model that runs on a schedule every couple of hours and checks new accounts for spam signals. If something looks off, the account gets frozen, and then once a day we manually review whether it's a real person or a bot and unlock or remove accordingly.

Not a perfect solution, but it's keeping things manageable for now. Good luck to everyone dealing with this.