Whether he likes it or not, Will Smith has become the face of a terrifying new age in which humanity may no longer be able to tell fact from fiction due to artificial intelligence.
We’ve developed all sorts of tests to see whether a piece of technology could masquerade as human or not and the latest of these is the ‘Will Smith eating spaghetti test’.
This became a thing a couple of years ago after people started sharing AI generated videos of the famous actor Will Smith eating spaghetti. They were obviously fake as an inhuman-looking Smith shovelled clumps of pasta into his mouth, while his entire face seemed to collapse in on itself.
Whether he likes it or not, Will Smith is the face of dodgy AI videos (Karwai Tang / Contributor / Getty)
The videos were very disturbing, but also reassuring in a way as they were evidence that people couldn’t just order an AI to vomit up a video which depicted the image of a real person seemingly performing a real task.
As long as you could look at an AI video of Will Smith eating spaghetti and tell that wasn’t really Will Smith eating spaghetti then we were fine.
We are no longer fine as sadly, according to Forbes, an AI developed by Google has passed the ‘Will Smith eating spaghetti test.’
They report that Google’s Veo 3 video generator has been put to the test and is capable of making videos which are now much harder to identify as AI at first look.
Now, I personally have never witnessed Will Smith eating spaghetti, but I would hope that the sides of his head would not pulse quite so much as they do in the video.
I would also hope that people recognise the ungodly sounds made are not the traditional noises emitted by a session of spaghetti-munching.
People are ‘terrified for the future’ (Andriy Onufriyenko / Getty Stock)
However, it’s much closer to the real thing than videos from just a couple of years ago, so who knows where we’ll be in another couple of years?
That prospect has got some people very scared, particularly since in the hands of the malicious a tool like AI video generation, which is hard to distinguish from the real thing, will be a recipe for disaster.
People watching this fake Will Smith slurp down a serving of fake spaghetti have warned that ‘AI is a plague on humanity’, and that increasing video realism is ‘utterly horrifying to me on an ethical level.’
“Misinformation about to go up ×1000,” was someone else’s verdict, while another said it left them ‘terrified for the future’ and a third warned it would have ‘disastrous effects on the world.’
Others called for making AI videos depicting real people to be illegal, with viewers saying it was ‘genuinely scary’ to see ‘how real these are getting so quickly’ and someone else called to ‘get some f**king regulations on this s**t.’
Another disturbing example of Veo 3 being used is a video of a conversation between three people that’s going viral because it never actually happened.
The AI generated footage contains a completely nonsensical conversation riddled with Gen Z slang used incorrectly, but some people were having trouble distinguishing it from the real thing visually.
Even if you can still tell these videos are AI generated the technology is clearly advancing at a rapid pace, how long until there’s an AI which can create a video even you can’t tell is fake?