Slashdot Mirror
US Lawmakers Say AI Deepfakes 'Have the Potential To Disrupt Every Facet of Our Society' (theverge.com)
Yesterday, several lawmakers sent a letter to the Director of National Intelligence, Dan Coats, asking him to assess the threat posed to national security by deepfakes -- a new type of AI-assisted video editing that creates realistic results with minimal effort. The Verge reports: The letter says "hyper-realistic digital forgeries" showing "convincing depictions of individuals doing or saying things they never did" could be used for blackmail and misinformation. "As deep fake technology becomes more advanced and more accessible, it could pose a threat to United States public discourse and national security," say the letter's signatories, House representatives Adam Schiff (D-CA), Stephanie Murphy (D-FL), and Carlos Curbelo (R-FL). The trio want the intelligence community to produce a report that includes descriptions of when "confirmed or suspected" deepfakes have been produced by foreign individuals (there are no current examples of this), and to suggest potential countermeasures. In a press statement, Curbelo said: "Deep fakes have the potential to disrupt every facet of our society and trigger dangerous international and domestic consequences [...] As with any threat, our Intelligence Community must be prepared to combat deep fakes, be vigilant against them, and stand ready to protect our nation and the American people."
They're obviously lying. Politicians are gonna love the proliferation of deepfakes. That way, the next time they say something stupid in an interview, they can say the clip is a deepfake. They're just starting to blame them as an upcoming problem now so they can start using it as an excuse ASAP.
Robert A. Heinlein imagined this problem, and in his novel Stranger in a Strange Land he described a new profession: "Fair Witness"
A Fair Witness is a person who is trained to observe and remember without jumping to any conclusions. A Fair Witness should be able to describe in court what he/she saw, and only that. As an example, a Fair Witness would say something like "I observed a house, and the side I saw appeared white" rather than "I observed a white house." It's possible that other sides of the house, not seen by the Fair Witness, could be a different color; and it's possible the house was repainted after the Fair Witness saw it... the Fair Witness keeps such things in mind.
Surprisingly, Wikipedia doesn't seem to think that the idea of "Fair Witness" is notable. I Google searched for a reference, and I found a reference that claims to be quoting Wikipedia, but I can't find it on Wikipedia now.
http://dlkphotography.com/fair-witness/stranger-in-a-strange-land
I found the "Fair Witness" idea to be one of the most interesting things in the book, and I have long wondered if we would one day see that profession in real life.
lf(1): it's like ls(1) but sorts filenames by extension, tersely
What could be possible:
* Cryptographic signatures on raw data leaving a camera, or Cryptographic signatures on the default recording app as the videos/photos are taken
Probably won't help that much though, but might help to identify unedited footage.
* Give images/videos timestamps signed by a third party immediately as they are taken
This can prove that a piece of information existed no earlier than that time.
None of these can thwart recording a video screen playing back pre-edited video.
And what gives them power? Laws criminalizing things. There is no way in this universe they can stop this. But they can profit from it and they are certainly trying to.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
We used to have sources that the majority of people believed in. Professors. Clergy, civic leaders. Not saying the trust was deserved, but at least people mostly agreed on facts. (even if they were wrong)