fawkes VS mia

Not sure if its coincidence but one of the professor's pictures appeared on the Fawkes page as a demo picture.

https://sandlab.cs.uchicago.edu/fawkes/

There is a great tool called Fawkes that allows you to modify pictures of your face in a visually undetectable way, enough to thwart AI algorithms and mitigate the risk of your face being used for fraudulent purposes. Well, I'm looking for a tool of the same kind that does the same thing with your voice, in real-time, during calls, using a separate app from the one doing the calls. I know there are plenty of voice-changer apps, but Fawkes was created because simply modifying an image using simple tools is not enough, and the same applies to the vast array of voice-changer apps available. I'm looking for one that is specifically designed to combat AI, as with Fawkes, and doesn't change your voice in any way that a human can detect. I'm not looking for a tool that anonimizes my voice from other people, but from AI. Does such an app exist? Context: The AI Dilemma, Tristan Harris and Aza Raskin, Center for Humane Technology (YT)

Fawkes might be worth a look for you: http://sandlab.cs.uchicago.edu/fawkes/

Leaving a link to Sandlab's Fawkes tool for confusing facial recognition models. If you have any pics of yourself on the internet, chances are they either will scrape or already have scraped them. Feed pictures into this before you post them.

Projects on privacy in machine learning is what a lot of companies are moving towards. Here’s an end to end project by google called ‘Membership Inference Attacks’ (around 2017) that was one of the first papers on how someone can extract information about people from trained models - https://github.com/spring-epfl/mia Of course Membership Inference Attacks have come a long way since then, but this repository shows you how you can create datasets and train models for such attacks. It’s a really good project but can also be very challenging!