Hacker News Reader

Comment by LinuxBender | Hacker News Reader

America May Not Be Ready for the Looming Tsunami of Deepfakes

How is this unique to America?

And it's too late for policymakers in the United States or the EU to stomp the brakes on the development and commercialization of deepfake technologies.

Even the body of the article agrees this also impacts the EU. I would suggest it impacts the entire internet. Is there an international body that would regulate this or will each nation have to implement it's own regulations once deepfakes reach some level of critical mass?

Are we about to reach a point where video and picture evidence will be dismissed from courts? Can deep fakes yet provide the quality required to allow someone to record a video of them breaking into their own home and then make it appear to be a neighbor they dislike? Could a government agency afford to create deep-fake evidence of a bank robbery to put someone away or silence them?

View on HN

In most western democracies I think mass surveillance will counter most efforts by private individuals to get off using deep fakes, I'm much more concerned about police using it to get evidence against people they "know" are guilty. The old scandals of police taking mixing evidence from the scene with dna collected from the suspect, using the latter twice then pointing to it as evidence comes to mind.

I really dont think any country is really ready for what is coming, but I also think its exactly the repressive countries that are in the most trouble, not democracies. Lets say some mid level bureaucrat sees a faked video of Xi telling everyone to spray their airport runways with disinfectant. Does so, cracks down on anyone critiquing it, even if its on reasonable basis such as saying Xi said airports, not airport runways, that they would stop. Lets say the video really gets wide spread, and is something even more stupid, what is Xi going to do? Admit he failed to control the internet, and tell people to think critically about the orders he gives?

Hmm, now I'm going to go tell 4chan to get me the script for the dumbest vaguely plausible thing to have Xi say. A tip btw, you want to create videos of politicians saying something and moving reasonably for a longer period of time, use one of the classic/old face to face transforms, then use that as prior for the deep fake.

American (and other British common-law style) legal systems are not ready for deep fakes. Probable cause based on deep fakes is too easy. A criminal record in America is deeply disenfranchising, even if there is no conviction. The legal process sucks the life out of people.

I think it might have to do with how much more available the tools are here (GPUs and such) and how it would be harder to repress disparate fake videos here than, for example, a unitary country with a state monopoly on news media (to those of you looking to read my comment in as stupid a light as possible that is not a commendation of such systems). So then it goes back to the original point, which I believe is that the US is "uniquely susceptible" to the blowback of destructive deepfakes which is concerning to policymakers and think-tanks. The real thing I want to be a pedant about here is how the suggestion solution is to "halt the commercial development of deepfake technology". First of all, you can't keep that cat bagged forever, and secondly a lot of such developments aren't commercial, they're Foss projects. You can't really mettle with those too much without cracking down on them in messier ways.

Reply on news.ycombinator.com