Comment by yalok
12 hours ago
yes, it was not easy to mentally give that access, but then I heard from friends how much work was put in into securing it and separating LLM/agents from the API tokens access (also saw this detailed blog post from the lead developer - https://research.meta.ai/blog/security-and-safety-for-ai-age...), plus I had the pressure to get the problem resolved - so I decided to risk and try it out.
Eventually, though, Muse Confidential VM is an even more locked-in option coming, per that blog post...
Ah, your concern was about an LLM messing something up, vs. about becoming a (tiny tiny) piece of training data for Zucky / getting slightly more targeted ads (or price discriminated against or something in the future)?
Mm, I see a small note on the second half of a bullet ( https://about.fb.com/news/2026/09/introducing-muse-personal-... ) that you can opt out of training. I know I’m a privacy weirdo since this didn’t make the FAQ in the OP:
Guess you can opt out on the free tier? (I’m not much [read: at all] trusting of the company but I know many are fine either way.) Not saying I’m happy about having my emails on Google’s servers btw or would be comfortable connecting email to Anthropic/OpenAI.
Wonder if opting out of training means they’d avoid using information gleaned from our emails to better serve their advertisers. Without knowing I guess I’d be concerned about even correspondents of mine using the service.
I was just answering a question above about the email access concerns. What you mention is another concern, but personally, if they are going to be showing ads to me, I'd actually prefer them to targeted for versus some totally random stuff.
As for training, I'd be highly surprised if they use the data they get from APIs for training. But the conversation itself & thinking trajectories probably will, unless you opt out, like you pointed out.
What if they use this data to target kids in your family in their vulnerable moments, like they did in the past?