Instagram Needs Some Introspection on Account Moderation in the Age of AI

I’m by no means a saint. Far from it according to some. But I do know how to conduct myself in the online world and don’t get myself into situations that would warrant TOS violations. You can imagine my surprise to have a note sprawl across my screen that my account was suspended because I was impersonating someone, seemed like a really bad (and late) April Fool’s joke.

On April 2nd, 2023, without warning or notice, I was informed that my Instagram account was suspended because it was suspected that I was impersonating someone else. I’ve owned the account since 2015, and since its inception I have logged in with my Facebook account since they’re both owned by the same parent corporation; I was doing cross-posting to share between my paltry number of Instagram followers and Facebook followers/friends, but mostly? concerned that the security posture for using separate creds wouldn’t really change so I decided to leave all (then) Facebook Corp.’s services to auth with my Facebook account. My Facebook account is protected with 2FA for as long as it was offered, whether SMS texts or OTA (app-based) rotating codes, passwords upwards of 50 characters—I followed all the best practices to keep my account’s security tight. I am exactly who I’ve always said I was on the platforms, and never attempted to be anyone else. (I gave up on that a long time ago when I was struggling with my identity as a gay man.)

I slowed my roll with posting pics and content given the recent controversies around how Facebook/Meta wheeled-n-dealed with, much like I did with my main Facebook account. I pivoted to actively using another social media platform without the toxic links and bad press, because I wanted to show off cool places and things I’d been, shots from my drone, etc.. I refused to be the fuel that powered Meta’s billion-dollar ad business and perpetuate the toxicity that swirled around most social media platforms.

Given the most interaction I had with Instagram leading up to April 2nd was following accounts, occasionally commenting on posts, even less occasional DMs, mostly just following other accounts (some blue-check types, mostly others just out showing off their lives) there wasn’t any real content to be flagged, or any content that’d show I was trying to be someone else. However, there it was, I was flagged for pretending to be someone else.

Okay, so what do I do now? I followed the prompts, had a special code sent to my registered 2FA phone number, followed the instructions clearly with hand-writing my name, username, and another special 6-digit code on a blank sheet of paper, the selfie picture should include my hand holding the sheet on a plain backdrop. It was going to be reviewed within a few days and a decision would be made then. Okay, guess it’s time to be patient and let the process work itself out. Should be simple, right? I auth’d in with my Facebook account, did the two-factor code thing, got confirmation with e-mail accounts—what more was left?

Well, months roll by and the same “we’ll get back to you in a few days” message was still there. Was there any way to reach out to Instagram to ask for help? Nope, the online resources pointed me in circles, no help desk tickets to file, and as long as I followed the steps, I could regain access to my account. Was I missing a step in doing the code-selfie-pic thing? Or was it simply buried in a queue so deep that no reasonable team of humans could look at it? I guess I needed more patience and would just wait it out. My network of professionals didn’t really have any contacts at Instagram’s moderation team to help figure out what it’d take to get this resolved—again being accused of impersonating someone else could be offensive in a way so I wanted to resolve this and reclaim my online identity! I assumed the review queue was too long and just had to dig in for more patience.

So fast forward to October 2nd 2023, and in my junk mail folder is a notice from Instagram that said, in summary, “Your account has been permanently disabled because your account, or activity on it, violated the TOU. As it’s been more than 180 days since we suspended your account, you can no longer request a review of this decision.”

Uh, wut? Let’s unpack the message here:

“My account or activity on it violated the TOU”

No, this is not why my account was (wrongfully) suspended. Unless you actually mean lack of new picture posting is a reason to axe my account, this isn’t why I was told my account was suspended. It was because I was accused of impersonating someone else and needed to validate my identity.

“As it’s been more than 180 days…”

I responded to the prompts to prove my identity on April 3rd around 2 PM PDT. I have the original photo from my phone to prove this. I responded quite promptly, thank you, it was Instagram’s failing to review the submission within 180 days, so this timeout elapsing is not my fault.

“…you can no longer request a review of this decision.”

I—already did? But now I have no further way to tell someone, anyone that I complied with the request and that it timed out waiting for Instagram to get through their backlog of reports. I’m being penalized for inefficiencies completely outside of my control.

Further, this entire thing is built on a lie because this wasn’t about my account activity at all.

For s&gs, I decided to log on to Instagram to see if the site had anything additional to say.

“Why this happened: Your account, or activity on it, doesn’t follow our TOU.”

Okay, got that part covered above. Oh, and thanks for the “read more about this rule” that’s effectively useless with a permanently disabled account.

“What this means: No one can see or find your account and you can’t use it. All your information will be permanently deleted. As it’s been more than 180 days…” (repeats the same thing in the e-mail.

Same old story as above. Fine. But what’s new is a section here called “How we made this decision.”?

“Our technology found your account, or activity on it, doesn’t follow our rules. As a result, a member of our team took action.”

“We use the same rules around the world, and for everyone on Instagram.”

Okay so you mean to tell me that your “technology” found my account and decided I was impersonating someone else, flagged it for suspension, and made me jump through a bunch of manual steps to prove I “owned” the account, yet conveniently at precisely the 180th day (4,320 hours to the moment) a “member” of “your team” took action?

This is a blatant falsehood. No one looked at my account. The request aged out. If a human had actually reviewed my account they would have matched my photos/videos with my selfie submission, aligned the codes, and unlocked my account.

If you were honest, the post would say “our AI models accidentally flagged your account as a false positive and we’re restoring your account to its original state. Our sincerest apologies.” But no, instead you’d rather just take the easy path out and blame “your technology” and a member of “your team” even though literally no member of your team was involved in this decision process.

If you’re going to take on “AI” as doing some moderation work on your behalf, you HAVE to train it to do better. I committed no sin here, yet I’m being deplatformed for an error in the algorithm with literally no way to reach out to figure out what happened, what I can do to fix it (assuming the sin was mine to begin with), or request an escalation. I get that humans are expensive to employ to go through these queues, but I should be afforded the opportunity to face my accuser. Yet, this does beg the question that if this is the way you’re running your business and moderating people, how often is this repeated across other innocent individuals with literally no recourse?

Is AI actually capable of doing this scale of moderation with this degree of failings safeguarding the user base from false accusations by an algorithm? What’s being put into place to ensure moderation actions are being reviewed, and requests to prove my identity or file a grievance being put through for an actual review?

Or, is it that you really do not care about deleting people from your platform erasing their history? Sure, I could easily go and create a new account and start over, but if this is how y’all want to run your business on top of the ethical problems Meta is facing, why would I want to do that? Would you rather have an army of bots running rampant on the platform instead?

If you want to use AI in this form, you have to do better, friends. Time for some Insta-Introspection to occur.?