Seriously, What the Fuck are We Doing Here?
Google’s Veo 3 and the death of truth as we know it.
Seriously guys, what the actual fuck are we doing here? I know Google has long since removed the ‘don’t be evil’ tagline from their bios, but I want to know what the actual fuck is going through their heads right now. I’m talking, of course, about Google’s decision to launch Veo 3, the latest version of their video creation tool that utilises artificial intelligence to generate hyper-realistic video AI slop. Now with dialogue!
The Verge reported this week on the launch of the product, finding out (concerningly quickly) that although there are some guardrails (you can’t make a fake video of Joe Biden falling over, for instance - wow, just in time! Turns out he was in great shape the whole time) there are ways that you can trivially manipulate the software to create incredibly dangerous video content. From eweek:
Veo 3 renders highly realistic video content, including synchronized dialogue, music, and ambient sound, with results that are often difficult to distinguish from live-action footage.
According to Axios, Veo interprets complex prompts with precision, simulates real-world physics, produces accurate lip movements, and generates human figures with lifelike proportions, including correct hand anatomy and smooth motion.
In a promotional video released by Google, filmmaker Dave Clark said, “It feels like it’s almost building upon itself,” expressing the creative freedom the tool offers.
‘Haha, it’s almost like it’s spinning out of control. I’ve never seen a science fiction movie before, it always goes so great when AIs do that, right?’
The software can now generate footage that is indistinguishable from news footage. Here’s a fun (needless to say, fake) clip of a phony anchor sitting in front of an image of the Pentagon talking about Pete Hegseth’s untimely demise after “drinking a litre of vodka” at the behest of RFK JR. Which, don’t get me wrong, is a sentence that, while incredibly funny, isn’t true (yet).
Or hey, look at the totally necessary video of the Space Needle in Seattle being the site of an imaginary terror attack, no way this could affect financial markets or people’s safety! Let's make a CNN anchor announce a nuclear first strike on Moscow. What could go wrong?
Fun! And necessary!
The biggest problem with this isn’t the hyper-realistic video (there are still some telltale signs that the video isn’t real, if you know what to look for); the problem is more insidious: we’ve created plausible deniability for reality.
Did Peter Dutton say that shitty thing about Lebanese migrants in the 1970’s (yes, he did) or was it a deep fake? Did Trump say ‘grab em by the pussy?” (YES) or is it all just fake news? The problem isn’t that Veo 3 makes videos that are so well-faked they look real - the problem is that they make real things look fake.
What is the actual purpose of this? We know all the reasons why this is terrible and going to poison the minds of our already lead-addled, conservative, boomer parents with no critical reading skills, but I’m begging you to tell me; what’s the upside?
I use AI as a professional writer to help draft and refine proposals. I bounce ideas off it, create outlines, and get it to make suggestions for improvement. It is a valuable tool in my arsenal to pull together vast amounts of research, or synthesise large amounts of unstructured data and turn around my output faster. The thing is, it’s still (mostly) my writing. It’s a helper, and it has the utility of increasing my productive capacity. I’m more of a conductor than a player in the orchestra, but it’s still my piece of music that we’re playing.
This? I don't know what the possible benefit of this is. What could the benefits be that would make the obvious risks to society worthwhile? I was chatting this over with Gemini, when I was putting the outline of this article together, and she (it) suggested that perhaps there were upsides of the technology that I hadn’t considered, like the creative freedom that filmmakers would now experience, due to being able to create content much more cheaply, or teachers that could communicate complex educational topics for their students through the wonders of costless AI slop video.
‘At the expense of the entire filmmaking and teaching industries…’ I responded.
Slight pause.
“That’s a valid concern…”
Yeah, no shit. The idea that we will have teachers, art, filmmaking, human interaction, and deeply researched pedagogy replaced with depersonalised, AI video slop is a deeply discomforting thought. I don't know why we're hurtling towards this brick wall of a future without even pretending to stop to think about the consequences. It’s time for that line from Jurassic Park that, evergreen, applies to the entire first quarter of this century:
Not to harp on this point, but again, what is the practical utility value of this software other than to break the information space? What actual advantages are created for society by giving everyone access to ready-made deep fakes, all the time, at no cost? I genuinely can’t think of any.
We could regulate it. We could put up bigger, better guardrails, but the way machine learning works, that would be surprisingly tricky to implement. As far as I’m concerned, anything to do with public figures or news, politics, health information, finance (and other important things I’m not thinking of, I’m certain) needs to be banned from generative video software. Anyone allowed to use it should only use it for discrete, specific business purposes. And it should be digitally watermarked and openly listed as AI content. It should be as regulated as handling uranium. But Google won’t do that; there’s too much money to be made. Instead, it’s going to be a bad actor’s field day; at least, Putin will save some money on his troll farms.
Yes, the video output of Veo 3 is (currently) limited to eight seconds, but this is the worst it will ever be. It’ll only get more realistic, more accessible to more people, and become more indistinguishable from reality. What about Veo 4, Veo 20 and Veo 200? The good news is, it’s also destroying the planet, so that’s fun.
I don’t know, tell me what I’m missing here - how is this good for anyone?