Yes, what is proposed is breaking the end-to-end security model, not breaking the encryption itself.

Effectively it causes the same loss of security and trust as if they broke the encryption, but it allows them the fig-leaf of pretending that you're still secure because they "haven't broken the encryption".