Over the last five years, Meta’s Oversight Board has weighed in on everything from Donald Trump’s Facebook suspension to AI deepfakes. Now the board is wading into another thorny issue: Meta’s rules for disabling users’ accounts.

The board announced earlier this year that it would look into improving transparency around the process, which is often frustratingly opaque. The oversight group dug into the issue following a referral from Meta regarding an Instagram account with 70,000 followers that was banned after making threatening posts targeting a journalist.

In its decision, the Oversight Board says that Meta was correct to ban the account, but the case raised “serious questions” about the company’s handling of such behavior and “due process concerns” around how it disables accounts. Because this is something of a test case, the board isn’t making formal recommendations to Meta, though it does highlight a number of potential improvements. Its analysis also highlights the confusing patchwork of rules and penalties that lead to bans on Meta’s platform, and the vast amount of frustration it’s caused for users.

For example, the board notes that Meta has strikingly different processes for Facebook and Instagram. While both platforms penalize accounts with “strikes,” repeated strikes can have different outcomes. On Facebook, accounts may receive temporary suspensions for repeated violations before an outright ban. But no such penalty exists on Instagram, the board says. Instead, Meta restricts accounts from Instagram’s livestreaming feature or will remove their account from recommendations (which Instagram users often refer to as a “shadowban”).

The Oversight Board rightfully points out how bizarre it is that restricting livestreaming is one of the main “intermediate” penalties on Instagram when the feature isn’t even available to all accounts (it requires a minimum of 1,000 followers). “For violations in permanent posts, a penalty that directly corresponds to violating behavior by suspending a user’s ability to post (e.g., by putting their account in read-only mode for a set period) would have a greater chance of influencing behavior,” the board notes.

The board also touches on the long-simmering frustration among Facebook and Instagram users who have accounts disabled. The group says it received more than 750 public comments in the case, in addition to the “innumerable” complaints individual board members regularly get from people who have had their accounts disabled.

“Many commenters wrote about systems failing to work, saying they were unable to appeal Meta’s decision to disable their account, that they never received any explanation for why their account was disabled or that they were unable to download their content,” the board wrote. “Many of these users also noted that the decisions appeared to have been made automatically, with no human oversight, even on appeals against the disabling of longstanding and widely followed accounts.”

In its guidance to Meta, the board suggests that the company should provide users with a better appeals process that allows them to provide written explanations and that users should be notified when AI is used to penalize their account. The board proposes that information about account bans could be added to Meta’s transparency reports for additional visibility. The group also advises that Meta provide a dedicated channel where “high-risk targets of violence and their representatives” can report serious threats against them.

Given that this case is described as a “pilot,” it’s unclear whether Meta plans to make any substantial policy changes in response to the board’s critique. But there is still some hope for those who want Meta to make improvements. The board says it plans to accept more cases in the future that deal with accounts being disabled, which would hopefully give them a better chance at influencing some reforms.