Google Confirms Its Awesome Gemini AI Demo Was Staged And Explains Why

To the ire of some, Google has confirmed that its Gemini AI hands-on demonstration video it showcased along with Gemini’s launch was somewhat misleading. While the tech giant said it wanted to inspire developers and did nothing wrong, some were quick to reply that the demo was disingenuous.

As it turns out, the actual interaction with Gemini did not include any voice interaction, nor was it happening in real time. The video was actually made by “using still image frames from the footage, and prompting via text.” This has led some to question just how advanced Gemini actually is, and calling into question Google’s integrity on the matter.

Many of those replying to Vinyals’ tweet were not buying it. One commented, “If you want to inspire developers then why don’t you post factual content? The prompts can’t be ‘real’ and shortened at the same time. It was disingenuous and misleading.” Another replied, “Why would you demonstrate with video and audio input, when it was actually image and text input? Seems disingenuous.”

In fairness to Google, the video does have a disclaimer attached that reads, “For the purpose of this demo, latency has been reduced, and Gemini outputs have been shortened for brevity.” However, this still includes nothing about the misconception that voice prompts were being used instead of text prompts.


In an op-ed on the situation, Bloomberg writer Parmy Olson remarked, “That’s quite different from what Google seemed to be suggesting: that a person could have a smooth voice conversation with Gemini as it watched and responded in real-time to the world around it.”

For those curious to how the Google Gemini interaction actually occurred, Google posted a “How its Made: Interacting with Gemini through multimodal prompting” post on its website.