Meta's AI Deletes Instagram Accounts, Appeals Go Nowhere

A New York Times investigation published on July 21, 2026 reveals that Meta's artificial intelligence systems are mass-deleting Instagram and Facebook accounts based on vague allegations, with appeals against these decisions rejected within days. This is happening despite the company's earlier assurances that the most critical decisions in disputes with users are made by humans, not algorithms.

Accounts Erased Without Explanation

Camille Hanson and her husband ran Facebook and Instagram accounts dedicated to teaching English to non-native speakers, followed by nearly a million people combined. In March, Meta flagged both profiles for removal, citing a violation of its rules on "fraud and deception" without specifying which post had actually broken the rules.

When Hanson filed an appeal, the response came within a week: Meta informed her that all her data would be permanently deleted and that she had no further recourse to challenge the decision.

Athenia Rodney met a similar fate. For 17 years she has run the JuneteenthNY account, with more than 500,000 followers, dedicated to Black American history and culture. Meta suspended it last year, alleging distribution of child sexual abuse material, a charge she has consistently denied.

Promises Versus Reality

On March 19, 2026, Meta announced a shift to "significantly more efficient AI systems" for content moderation, while cutting the number of external contractors handling manual review. At the time, the company insisted that key decisions in disputes over suspended accounts were still made by humans.

Humans retain the central role in the most critical decisions - Meta, statement of March 19, 2026

Yet the cases documented by The New York Times show that in practice, appeals are rejected quickly and without explanation, leaving users with no way to reach a human representative or learn exactly what part of their activity supposedly violated the rules.

Oversight Board and German Regulator

Meta's Oversight Board, an independent advisory body funded by the company, reached similar conclusions earlier in its review of how Meta blocks accounts on its platforms. In a report published on June 4, 2026, the board described the problem as serious and systemic.

Systemic human rights concerns - Meta's Oversight Board, report of June 4, 2026

The board's auditors also pointed to a lack of transparency and traceability in how account-blocking decisions are made, and criticized Meta Verified, the paid subscription meant to give subscribers faster help in moderation disputes, which in practice offers no meaningful support to users whose accounts have been suspended. The board's recommendations are not binding on Meta.

Regulators are flagging similar problems. Germany's Bundesnetzagentur, the agency responsible for overseeing digital services, reported on April 30, 2026, that it received more than 2,000 complaints about Meta's decisions in 2025, with the most common complaint being insufficient justification for restrictions placed on accounts, content, or services. The agency has opened 26 administrative proceedings on the matter, concerning Meta's obligation to justify its decisions and how it handles complaints.

What This Means for Polish Users

For Polish creators and businesses that rely on Instagram and Facebook for sales or customer contact, the issue has practical stakes: losing access to an account with no path to appeal often means losing the only communication channel built with an audience over years. Experts cited by Notebookcheck recommend regularly downloading backups of platform data and building alternative communication channels that don't depend on a single platform.

The case feeds into a broader debate about platform accountability for algorithmic decisions, an area partly regulated in the European Union by the Digital Services Act (DSA), which requires large platforms to justify moderation decisions and provide an appeal path reviewed by a human.