4 ms·
Their promise not to train on the conversations seemed to imply that conversations were private. It is disconcerting that they were not. The extent to which you
by ryao 2y ago
Their promise not to train on the conversations seemed to imply that conversations were private. It is disconcerting that they were not. The extent to which you need to look for loopholes in privacy guarantees these days is amazing.
- neom 2y agoTheir privacy policy is EXTREMELY easy to interpret: https://www.anthropic.com/legal/privacy https://www.anthropic.com/legal/privacy They use personal data “to improve the Services and conduct research.” Your chat interactions (that is, your "Inputs and Outputs") are included in the data they collect. and: "If you include personal data in your Inputs, we will collect that information and this information may be reproduced in your Outputs." You don't need to look for loopholes, it's spelled out plainly.
- falafels 2y agoI listened to an interview a while back with the researchers did this analysis. They developed privacy-preserving techniques, avoiding having to read user conversations directly. This is Anthropic we're talking about, they're rightly recognized as the 'ethical' AI company.
- simonw 2y agoThey explained the privacy-preserving mechanism they use for this in detail a few months ago: https://www.anthropic.com/research/clio https://www.anthropic.com/research/clio