The end-to-end encryption that Meta is removing from Instagram direct messages by May 8, 2026, was never truly universal. From its introduction in 2023, it was an opt-in feature that required users to deliberately activate it. The vast majority of Instagram users never did. In practical terms, most Instagram DMs have never been end-to-end encrypted. What is changing in May 2026 is that the option for those who wanted it — and used it — is being taken away entirely.
This context matters for understanding the scale of the change. Privacy advocates are correct that the removal of even a limited privacy feature is significant and worth resisting. But framing the removal as the loss of universal encryption is inaccurate. What is being removed is the option for a minority of users who chose to activate an opt-in privacy feature. The privacy loss is real for those users — but it is not the sudden removal of encryption from a universally protected system.
The more important point is why the feature was never universal. It was opt-in by design — a design choice that reflected the institutional and political pressures Meta faced when introducing the feature. Making encryption the default would have been a more meaningful commitment to user privacy, but it would have generated more intense opposition from law enforcement and regulators. The opt-in design was a compromise that satisfied neither privacy advocates nor their opponents.
Now that compromise is being retired. Meta is using the low adoption of the opt-in feature as justification, but the low adoption was partly a product of the opt-in design itself. The feature that was never truly universal because Meta designed it not to be is now being removed because it was never truly universal. The circularity of this logic is the clearest evidence that the official explanation is incomplete.
For users who did activate the feature, the loss is genuine. For the majority who did not, the May 2026 change formalizes a status quo that already existed. The more important question — why the feature was never made default — is one that has not been adequately addressed by Meta’s public communications.