Discord's CTO addresses community concerns about age assurance: no mass ID collection, new vendor transparency commitments, and a delayed global launch until second half of 2026.
For 90%+ of users, nothing changes. Most users never access age-restricted content or change their default safety settings. For those who do, we have an internal system that works to accurately determine your age.
“accurately” my ass.
We know many of you believe the right answer is not to do this at all. We hear you.
You obviously don’t.
the security incident last year involving our third-party customer service provider only adds to that skepticism. To be clear, we do not use that vendor for age assurance. In fact, we no longer work with them at all, and we’ve taken the lessons from that incident seriously.
You obviously haven’t.
We heard you, and we want to get this right. So here’s what’s happening: We’re delaying our global rollout to the second half of 2026.
I’m without question one of the 90% and I never thought they were going to require it for everyone. But I don’t like the idea that an algorithm is constantly writing a dossier about me to determine my “age” because I wasn’t fucking born yesterday.
Open source your shit and make it auditable or I’m not fucking using it anymore.
The ‘internal system that accurately determines your age’ line is doing a lot of work. What it’s describing is behavioral age inference — classifying users by activity patterns, message timing, content signals, server membership. That’s the mechanism behind the 90% claim: they already have enough behavioral signal on most users that they don’t need to ask. The disclosure buried in the blog post is that continuous behavioral profiling is already running. ID verification would have been a one-time check. This is permanent. The ‘we won’t collect your ID’ framing is technically accurate and completely misleading.
Lip service
“accurately” my ass.
You obviously don’t.
You obviously haven’t.
LOL you aren’t.
I’m without question one of the 90% and I never thought they were going to require it for everyone. But I don’t like the idea that an algorithm is constantly writing a dossier about me to determine my “age” because I wasn’t fucking born yesterday.
Open source your shit and make it auditable or I’m not fucking using it anymore.
The ‘internal system that accurately determines your age’ line is doing a lot of work. What it’s describing is behavioral age inference — classifying users by activity patterns, message timing, content signals, server membership. That’s the mechanism behind the 90% claim: they already have enough behavioral signal on most users that they don’t need to ask. The disclosure buried in the blog post is that continuous behavioral profiling is already running. ID verification would have been a one-time check. This is permanent. The ‘we won’t collect your ID’ framing is technically accurate and completely misleading.