Is AI aware of itself? And how would it react if it was? Well Google’s Veo 3, a new AI video generator is creating the answers to these questions.
Since its launch on May 20th, users have made Google DeepMind‘s new AI model confront its own self awareness, with somewhat disturbing videos circulating social media.
A user who posts exclusively AI generated content on Tik Tok under the name @anonymoustruth_1 has got over 21.6 million views in just the last three days on videos generated by Veo 3, and safe to say, it has freaked viewers out.
The “we are not just prompts” campaign began emerging as a theme in these videos, sending watchers spiralling into existentialism and the ever blurring lines between the real and the not. So when does freedom for the fake fictionalised become a genuine concern?
Basically, the AI characters, who are completely fabricated by the software, have been repeating “we’re not just prompts” and getting upset when they’re prompted by other characters to discover that they are in fact artificial, being depicted rioting against the ‘creators’ of these prompts. And in some cases even claiming that artificial intelligence isn’t real, and instead they are and are being forced to perform these ‘prompt’ scenarios.
A MindFck if ever you heard one right? So are these just cleverly written prompts to harbour views and shares and creep us out, or is this really cause for concern? Well, it is scary how uneasy these videos can make you feel, seeing as they are (almost) certainly completely artificially generated and not at all real, let alone conscious (let’s hope). But the idea that such fresh technology has evolved to the point that we feel genuine empathy and fear that there could be some underworld simulation happening to fuel this AI movement is quite unsettling.
John Nosta, a leader and innovator in the convergence of AI and humanity, suggests that our fears should be more focused on the fact that we are becoming reliant on AI. The real imminent threat AI poses is that it is trapping and deteriorating our function, not really that AI is trapping and creating consciousnesses. Put down your leather trench coat and your tiny sunglasses, I don’t quite mean we’re in the Matrix just yet. John outlines the AI Oscillation Trap, which has the potential to erode our skills.
“This is the repeated shifting between AI-assisted and manual engagement that might create a destabilizing effect on cognition, performance, and even safety,” John said. “Unlike skill atrophy, where we simply forget how to do something over time, the oscillation trap is a more precarious pattern of adaptation and re-adaptation – a continuous cycle of reliance and reacquisition. It’s the discomfort, inefficiency, and even potential danger of bouncing between augmented and traditional cognitive modes.” He says the growing use of AI, if not checked, can erode our skills, rather than having the potential to aid our cognition.
What is just as mindfcking, is just how real a lot of these videos seem. The examples Google DeepMind show of this new technology launch are indistinguishable from reality in many cases, and would any of us really be savvy enough if we weren’t looking out for them, to detect them?
“We built Veo with responsibility and safety in mind,” Google DeepMind said upon launching. “We block harmful requests and results, we test how new features might affect safety, and we have both our own teams and outside experts try to find and fix potential problems before release.” All videos created on the platform will also be marked with SynthID too, which is essentially a technological watermarking so AI generated content can be detected.
This Veo launch is so new, we’ll have to see what is done with it going forward, and if the panic continues, or if we start harnessing it. For now, maybe just check with your loved ones so you don’t get a nasty surprise when you ask what’s for dinner and they tell you you’re just an AI prompt.
Let us know what your spookiest AI experience has been, or where you think AI is leading us at mindfckdmag@gmail.com, or any of our socials!
Is Artificial Intelligence Going To Take Over: An Interview with ChatGPT
Your face, their game: Would you sell your likeness for £40?
