With this open source deep fake detection tool, Delta Chat could recognize when a photo is true or false and alert the user.
In my opinion it is an interesting addition. More and more users have been bombarded with images manipulated for various purposes. A deep fake recognition mechanism is an aggregating feature.
really interesting idea, but nothing for the near future (because it is not an rust crate that we could simply integrate and i think it needs too much cpu performance)
On messengers? I guess this is more common on social media platforms such as facebook, twitter and instagram.
If you were to develop this tool, ideally it would have to work only on the desktop. I do not think that the vast majority of smartphones have performance for this without compromising performance.
In social media it is more common, but given the movement of facebook and others to fight fakes, the environment in messengers is practically uncontrolled. I believe there will be a migration from the type of falsehoods to messaging applications because they cannot be censored.
I am not the type to preach or believe in the need for corporate and government surveillance on messengers. This is absurd to me. But if there are tools that make it possible to identify a deep fake in a way that guarantees privacy above all, for me it’s fine.
Someone could make a bot for this using the delta chat python bot! https://github.com/deltachat-bot/deltabot
Seems like user (you) could more easily download the image yourself to feed it to a tool of choice.
