At the Google I/O 2025 event on May 20, Google introduced Veo 3, a novel AI video generation model designed to produce 8-second videos. Immediately following its launch, AI artists and filmmakers started displaying strikingly realistic videos. You might have encountered some of these clips on social media without realizing they were created by AI.

In simple terms, Veo 3 is groundbreaking. It’s remarkable, it’s disconcerting, and it’s bound to evolve further.

Experts have frequently cautioned that we would arrive at a stage where it would be impossible for the average individual to differentiate between AI videos and genuine ones. With Veo 3, we have stepped into a new era where AI-generated videos are indeed a reality.

Although other AI video tools are available, such as Sora from OpenAI, the clips generated by Veo 3 are particularly notable. Veo 3 brings forth numerous innovations, including audio and dialogue creation, photorealism, soundscapes, and interactive conversations. It ensures consistency in characters across clips, and users have the ability to modify camera angles, framing, and movements in novel ways. Many users on social media are in awe of the outcomes.

Veo 3 is currently accessible through Google’s paid AI subscriptions and can be reached via Gemini, Google’s AI chatbot, and Flow, an AI filmmaking resource for creators.

AI filmmakers have already begun utilizing Veo 3 to produce short films, and it’s only a matter of time before a feature-length film is produced using Veo 3.

On platforms such as X, YouTube, Instagram, and Reddit, users are posting impressive videos made with Veo 3. If you’re not careful, you might fail to question the authenticity of these videos.

The short film “Influenders,” crafted by Yonatan Dor from The Dor Brothers, is being widely circulated. It features influencers reacting to an unforeseen catastrophe. The video has received hundreds of thousands of views.

“Yes, we exclusively utilized Google Veo 3 for this video, but to bring it to life, we required additional sound design, creative editing, and some upscaling,” Dor commented in an email to Mashable. “The complete piece took about 2 days to finish.” Dor also noted, “Veo 3 marks a huge advancement, by far the most sophisticated tool available to the public at this moment. We’re particularly impressed with its dialogue and prompt adherence abilities.”

Similar videos have gained traction, with artists like Alex Patrascu and Impekable exhibiting Veo 3’s strengths. A Wall Street Journal reporter recently produced a short film featuring a virtual version of herself using Veo 3, all in just 10 days.

In “Influenders” and other videos, certain clips and characters exhibit more realism than others. Many still retain the shiny aesthetic and jerky character motions commonly associated with AI videos, a clear hint reminiscent of the ChatGPT em dash.

A few years back, AI creations with anatomical flaws were prevalent. If advancements continue at this pace, there will soon be negligible differences distinguishing real videos from AI-generated ones.

Google highlights its collaborations with artists and filmmakers, such as Darren Aronofsky, in the promotion of Veo 3. Veo 3 has the potential to drastically lower the costs associated with animation and special effects. However, it also poses risks for exploitation by content farms and malicious actors generating fake news.

Google has established measures such as digital watermarks in Veo 3 video clips. “It is crucial that individuals have access to provenance tools for videos and other content they encounter online,” a representative from Google DeepMind informed Mashable. “The SynthID watermark is incorporated into all content produced by Google’s AI tools, and our SynthID detector was launched to early testers last week. We aim to broaden access soon, and we’re introducing a visible watermark for Veo videos.”

Furthermore, Google complies with AI safety protocols and intends to “assist individuals and organizations in responsibly creating and identifying AI-generated content.”

But do average users question the authenticity of images and videos appearing in their feeds? As the viral emotional support kangaroo illustrates, they frequently do not.

AI videos are projected to become increasingly prevalent on social media and video applications. This will encompass both subpar AI content and videos with more nefarious intentions. Despite existing safeguards, proficient AI artists can produce deepfakes involving celebrities and public figures. TV news anchors addressing the camera have emerged as a recurring theme in Veo 3 videos, raising concerns about the integrity of the online information landscape.

If you haven’t begun to question “Is this real?” when viewing video clips online, now is the appropriate time to start.

Or, as many are remarking on X, “We’re so cooked.”

Disclosure: Ziff Davis, the parent company of Mashable, initiated a lawsuit against OpenAI in April, alleging it infringed Ziff Davis copyrights while training and operating its AI systems.