I wonder when chat gpt being included in everything will backfire. How many requests are being sent to a multiple servers that people are unaware of and how secure are those servers?
Yeah my company enabled the use of copilot with our corporate Microsoft accounts. I don't understand how you can open such a massive can of worms to ALL your users. It's pretty much begging for information to completely leave corporate control. It's absolutely insane.
Believing you're immune makes you particularly vulnerable, because it may hinder you from noticing that you're wrong. None of us are above deception or manipulation, and to assume otherwise is to let down our guard.
I mean they've been doing multitenancy in Azure for a while.. I'd be pretty surprised if the data input from copilot was not handled in the same manner.
Local government is also having this issue. Lots of cities don't have the resources so will probably throw a lot at gpt...which means a very large attack vector.