Elena Kindyni
@elenakindyni.bsky.social
Leading the digital editors for the Council of the EU #ITLaw | #techpolicy| #bigtech | Web & legal geek. Doesn't get bluer than Greece. Views are my own
If the findings hold, it looks like it is surprisingly easy to extract training data from supposedly sophisticated and closed source large language models systems like ChatGPT, showing how vulnerable these systems are, especially in terms of privacy
New Study Suggests ChatGPT Vulnerability with Potential Privacy Implications | TechPolicy.Press
Tech Policy Press is a nonprofit media and community venture intended to provoke new ideas, debate and discussion at the intersection of technology and democracy. We publish opinion and analysis.
techpolicy.vercel.app
December 5, 2023 at 6:59 PM
If the findings hold, it looks like it is surprisingly easy to extract training data from supposedly sophisticated and closed source large language models systems like ChatGPT, showing how vulnerable these systems are, especially in terms of privacy
Trying to overcome the ‘perfect first post anxiety’, I am sharing the book I am currently reading & the setting in which I am enjoying it. From the first email to the pop up ad this is a great selection of essays of how a fee lines of code changed our lives
December 3, 2023 at 10:43 AM
Trying to overcome the ‘perfect first post anxiety’, I am sharing the book I am currently reading & the setting in which I am enjoying it. From the first email to the pop up ad this is a great selection of essays of how a fee lines of code changed our lives