News for hackers. Stuff that matters.

A comment on Docusign just admitted that they use customer data to train AI. Parent comment

Re:Docusign just admitted that they use customer data to tra... by 65a on Friday March 01, 2024 @01:47AM (#39559236)
https://en.wikipedia.org/wiki/K-anonymity Basically keeping only chunks of text which have are not unique and contributed by many users.
[ Parent | on HN ]
Re:Docusign just admitted that they use customer data to tra... by bastawhiz on Friday March 01, 2024 @05:14AM (#39560379)
> Given person-specific field-structured data

That's not what this data is and that's broadly not what LLMs are trained on.

[ Parent | on HN ]
Re:Docusign just admitted that they use customer data to tra... by 65a on Friday March 01, 2024 @02:16PM (#39565645)
By chunking and binning the data by user contribution count, you are structuring it as described. Pretraining is often accomplished on similar chunked text. k-anonymity is not perfect though, or state of the art, the wikipedia explains several known attacks.
[ Parent | on HN ]