Data Poisoning Won't Save You From Facial Recognition · HackerPapers discussions

Data Poisoning Won't Save You From Facial Recognition — mentioned by jessa0

Sun, 01 Feb 2026 15:34:59 GMT

It sounds like this is a poisoning attack, which has been shown to be pretty trivially defeated [1]. That said, while poisoning countermeasures in the facial recognition case were shown to easily generalize, we dont know yet how general of a defense could be built for a VLM. Which means holding a 0day poisoning attack on a VLM could cause a lot of trouble / deaths before an update to the model with counter-training could be deployed.. [1] https://arxiv.org/abs/2106.14851

Paper and discussion history. A mention is not an endorsement.

Data Poisoning Won't Save You From Facial Recognition — mentioned by simandl

Mon, 20 Mar 2023 15:00:39 GMT

Last year's ICLR had a paper, "Data Poisoning Won't Save You From Facial Recognition" that included the Glaze team's previous project, Fawkes. This statement from that paper is quite damning. "This paper shows that these systems (and, in fact, any poisoning strategy) cannot protect users’ privacy. Worse, we argue that these systems offer a false sense of security. There exists a class of privacy-conscious users who might have otherwise never uploaded their photos to the internet; however who now might do so, under the false belief that data poisoning will protect their privacy. These users…

Paper and discussion history. A mention is not an endorsement.

Data Poisoning Won't Save You From Facial Recognition — mentioned by belter

Tue, 29 Jun 2021 09:27:09 GMT

https://arxiv.org/pdf/2106.14851.pdf

Paper and discussion history. A mention is not an endorsement.

Data Poisoning Won't Save You From Facial Recognition — HN submission

Tue, 29 Jun 2021 09:27:09 GMT

Data Poisoning Won't Save You from Facial Recognition

4 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.