Seeing is Not Believing
Nothing is True and Everything is Possible
Imagine scrolling through your feed and seeing what appears to be breaking news footage of the president declaring martial law.
The video looks authentic. The news anchor has natural speech patterns and movements, and her audio moves in sync with her lips. But in fact the clip has been entirely fabricated using Google’s Veo 3 video generation tool. Welcome to the realm of the hyperreal.
Veo 3 is an impressive technical achievement, to be sure. From a single text prompt, the newest video generation model from Google DeepMind can produce fairly compelling scenes - so far, limited to eight seconds in length, though you have to think that will be expanded, at least for premium subscribers. What seems to set it apart from previous models, however, is audio syncing. What do I mean by this? With previous iterations of tools like OpenAI’s Sora, visuals were impressive, but any audio would have to be added manually in post-production.
With Veo 3, you can sync audio to characters’ mouth movements, and even add sound effects at just the right time.
In addition to the announcement of Veo 3, Google also announced Flow, a tool for building stories from multiple scenes that uses Veo, Imagen (its image-generation model), and Gemini’s generative AI.
As I’ve watched many of these generated videos on social media, it appears that Google may have significantly reduced the notorious issue of humans showing up with six fingers on a hand. Occasional visual glitches persist - weird subtitles, characters moving unnaturally, and so on.
Nevertheless, this still feels like a quantum leap forward. What could this mean for our already fractured reality?
The World of the Hyperreal
"The simulacrum is never that which conceals the truth—it is the truth which conceals that there is none. The simulacrum is true.” - Jean Baudrillard
[Simulacra and Simulation was reportedly required reading on the set of 1999’s The Matrix.]
The French cultural theorist and philosopher, Jean Baudrillard, seemed to anticipate this moment when he wrote 1981’s Simulacra and Simulation. The book is very dense, as you might expect, but for our purposes, we need only focus on his concept of the simulacrum.
A simulacrum, in Baudrillard’s telling, is no mere copy of something that exists in nature. It is a creation that has either lost its connection to the original or for which there never was an original in the first place. We’re moving beyond copies of reality to reality replacement.
Think, for example, of a modern-day influencer with their carefully curated self-representation. Each photo, each Reel or TikTok, is designed for public consumption - a performative spectacle - and may bear little to no resemblance to the messy, imperfect person behind the screen. This is perhaps taken to its logical conclusion with the rise of virtual influencers like Lil Miquela, for which there is no original “real” self at all.
Tiktok failed to load.Enable 3rd party cookies or use another browser
There is an argument that tools like Veo 3 and Flow may help further democratize filming. But this accessibility cuts both ways. The introduction of a new technology is often a double-edged sword, and we must exercise caution and consider possible ramifications while the technology is still in its early days. With these tools, users may construct scenarios with full audio immersion to fit preferred narratives or bespoke realities. Consider that clip of a fake martial law announcement that started this piece.
Or consider the example that researcher Henk ven Ess came up with, which involved a fictional local news story where the mayor had decided to convert a school into a yacht manufacturing facility. The clip is complete with chanting protestors, a first-person POV clip of a student describing the scene, a TV news host, man-on-the-street interviews, and even an interview clip with the mayor.
Future scenarios to consider might include:
Politics - the obvious target, and this has already been happening even before these uber-realistic fakes. Hyper-local fakes could be generated without getting too much attention.
Plausible Deniability on Demand - this is known in the research as the liar’s dividend. “That wasn’t me - that’s obviously generated by AI.” By simply claiming 'this could be deepfaked,' bad actors can weaponize our uncertainty, turning the technology's existence into a shield against real accountability.
Effect on Relationships: Think weaponized gaslighting. A breakdown of social trust.
Baudrillard warned us that the simulacrum might become more real than reality itself. With tools like Veo 3, we're not approaching that future - we're living in it. When everything might be a simulation, nothing demands our attention, and we continue to fracture our shared reality. Real issues and real suffering become moments to scroll past, dismissed as likely fakes. And - even worse for an empath like me - we become desensitized to it all. We don't just lose our shared reality - we numb ourselves, losing our capacity to care about it.
Additional reading:
This Week’s Recommendations
📖 What I’m Reading: Continuing to make my way through Red Dead’s History, which continues to be surprisingly informative about Gilded Age America. And I’ve started Jason Keath’s More Bad Ideas: The Counterintuitive Guide to Creativity (full disclosure - I first met Jason years ago in his role as co-founder of the marketing conference series Social Fresh). Jason is a thoughtful speaker and writer, and I hope to gather insight from his new book.
🎵 What I’m Listening To: Came across the UK singer YUNGBLUD’s new song “Zombie” by accident and really love it. It’s an emotional song about mental struggles and burnout and feelings of self-worth. The video also has Florence Pugh (!), one of my favorite young actors.
Until next time,
Jonathan