Google is betting on Gemini – its own GPT-4 competitor – so much that staged parts of a recent demonstration video. According to Bloomberg, Google admits that in the case of the video titled “Hands-on with Gemini: Interacting with multimodal AI” it not only shortened the AI ​​responses to speed up the video (which was declared in the video description), but the implied voice interaction between the human user and the AI ​​did not actually exist at all.

The demo was realistically created “using still images from the recording and textual prompts,” instead of Gemini reacting to – or even predicting – drawing or changing objects on the table in real time. That’s a lot less impressive than the video looks at first glance, moreover the absence of a statement on the actual input method makes Gemini’s readiness somewhat questionable.

Not surprisingly, Google denies any wrongdoing here. The company bequeathed the magazine The Verge to X’s post, written by Gemini co-head Oriol Vinyals, stating that “all user prompts and outputs in the video are real” and that his team created the video “to inspire developers.” Given the recent attention paid by industry and authorities to artificial intelligence, begs the question of whether a tech giant of its size should be a little more sensitive about the ways in which it creates its own artificial intelligence presentation.

Source: engadget.com