Google Releases Veo 3, AI Video Generator Now Equipped With Auto Audio

JAKARTA - Google officially launched Veo 3, the latest AI-based video generative model, in this week's Google I/O event. One of the flagship features of Veo 3 is its ability to generate automatic audio, both background sound, dialogue, and voice effects without requiring voice input from users. This marks a major leap in generating technology, while giving rise to various reactions from technology observers.

In a review written by total Johnson, tech journalist with more than 10 years of experience, Veo 3 is referred to as an "AI stamp engine" because of its highly realistic ability to generate visual and audio content, although it is not always relevant or to the user's liking. Johnson noted that as he tries to make videos using simple text prompts, Veo 3 can add never-so-asked dialogues with body movements and quite convincing atmospheres.

For example, when making a video about a fire at Space Needle, this AI not only displays visual disaster, but also adds a news anchor who submits reports of the incident in a realistic voice and voice background.

The same thing was also done by Alejandra Caraballo, an instructor at Harvard Law School, who succeeded in making fake videos of news carriers announcing the death of US Defense Secretary Pete Hegseth even though the figure was still alive.

Google claims to have implemented restrictions and guardrails in the use of Veo 3. For example, users cannot make videos about the president falling, killing public figures, or CEOs of technology laughing in the middle of the rain of money. However, Johnson insists that even without loopholes or special tricks, users can still create potentially misleading videos, such as fake natural disasters or fictitious events that appear to be delivered by official news agencies.

However, Veo 3 cannot be used to create personal deepfakes directly. When Johnson tries to make videos using his own with certain dialogues, the system refuses to process them. However, for simple content such as children's cartoon videos, Veo 3 is very effective. It is capable of creating videos similar to YouTube Kids's content. Monster trucks that glide into colorful paints, complete with music and voice effects in a matter of minutes.

Johnson's biggest concern arises when trying to make a video of two cartoon cats fishing. Without including dialogue in the prompt, AI still generates conversations between cats that sound natural. From here came the big question: if you make a short video as easy as this, how much longer will it take for people to start producing long videos containing misleading information only with AI?

For now, videos that want to be extended in duration will be returned to the Veo 2 system, which does not support automatic audio features. However, with the speed of technology development by Google, many believe that the full AI video will soon become a reality.

Google itself shows the positive potential of this technology by displaying the collaborations of Eliza McNitt and well-known director Darren Aronosky, who is developing films with AI video elements.

However, total Johnson closed his report in a critical tone. According to him, instead of producing high-quality cinematic works, Veo 3 will most likely be used to flood the internet with generic and bland content that is now made easier with images, movements, and AI-made sounds.