Google Veo 3’s AI video realism just broke the Internet with a street interview that never happened

On May 20, 2025, Google introduced Veo 3 at its I/O event, showcasing a major leap in AI-generated videos. Built to create hyper-realistic footage with synchronized audio, Veo 3 quickly became the talk of the internet, not for its technical specs, but for a street interview video so convincing, many believed it was real. It wasn’t.

Veo 3’s Street Interview Video That Fooled Everyone

The clip, posted by X user @minchoi, shows a man in a blue shirt interviewing two women on a neon-lit city street. Everything about it looks authentic — the dialogue, facial expressions, ambient noise, and camera work. But the entire scene was created from a single text prompt. No actors. No street. No camera.

Google Veo 3 realism just broke the Internet yesterday.

This is 100% AI

10 wild examples:

1. Street interview that never happened pic.twitter.com/qdxZVhOO3G

— Min Choi (@minchoi) May 22, 2025

That revelation sparked a wave of disbelief and concern. If an AI can fake a video that realistic, what happens when the intent isn’t harmless?

Veo 3 is part of a broader push by Google to position AI at the center of media creation. Alongside Veo, Google launched Flow, an AI filmmaking tool that combines Veo with Imagen for visuals and Lyria for music and sound. The result is an all-in-one studio that lets users generate scenes with full audio, dialogue, and cinematic effects using nothing more than prompts. At the event, CEO Sundar Pichai called this “the most exciting era of AI yet,” noting its potential to reshape creative industries.

From Filmmaking Tool to Ethical Flashpoint

But Veo 3’s realism is forcing bigger questions.

The street interview — described by @minchoi as “100% AI” — has become a flashpoint in debates over ethics, disinformation, and media trust. Oren Etzioni, founder of TrueMedia.org, warned that videos like this could be used to manipulate public opinion or mislead voters. The concern isn’t theoretical — it’s immediate.

What separates Veo 3 from earlier models is how well it handles both visuals and sound. According to DeepMind CEO Demis Hassabis, Veo 3 can interpret video scenes and sync audio automatically, making it possible to generate realistic conversations in real-time. For creators, it’s a powerful tool. For misinformation campaigns, it’s a gift.

Google says it’s taking steps to reduce the risk. Veo 3 embeds invisible watermarks using SynthID, a DeepMind-developed system designed to tag AI-generated frames. But experts remain skeptical, especially as detection tools lag behind generation tools, and bad actors find ways to work around safeguards.

Online reaction has been mixed. Some are amazed by what Veo can do. Others are deeply uneasy. “So lifelike, it’s unsettling,” one user wrote. Another said it “blurs the line between tech and art” and predicted it would shake up the film industry. That tension, awe versus anxiety, now defines the conversation around AI video.

Technically, Veo 3 builds on years of internal research. It’s better at following prompts and gives users more control over shot composition and pacing. But details on its training data remain vague. Google hasn’t confirmed whether YouTube content was used — a sticking point for critics who worry about copyrighted material being pulled into training sets without permission.

The Impact on Jobs and Hollywood’s Future

The ripple effects go well beyond entertainment. A 2024 report by the Animation Guild estimated AI could disrupt over 100,000 jobs in the U.S. film and TV sectors by 2026. With tools like Veo 3, that timeline feels less hypothetical.

Lawmakers are already reacting. California and Washington have passed laws aimed at limiting the use of AI-generated content during election cycles. Other states are drafting similar bills. But regulation is only one piece of the puzzle.

Etzioni argues that broader solutions are needed: stronger detection systems, better media literacy, and public education around synthetic content. The fake interview may have started as an experiment, but its implications are real and urgent.

Veo 3 didn’t just break the internet. It exposed how fragile the boundary between real and fake has become. The question now is how we respond — and how long that line will remain visible.