Google is relying on its very personal GPT-4 competitor, Gemini, a lot that it staged elements of a current demo video. In an opinion piece, Bloomberg says Google admits that for its video titled "Hands-on with Gemini: Interacting with multimodal AI," not solely was it edited to hurry up the outputs (which was declared within the video description), however the implied voice interplay between the human consumer and the AI was truly non-existent.
As an alternative, the precise demo was made by "utilizing nonetheless picture frames from the footage, and prompting by way of textual content," quite than having Gemini reply to — and even predict — a drawing or change of objects on the desk in actual time. That is far much less spectacular than the video needs to mislead us into considering, and worse but, the shortage of disclaimer in regards to the precise enter technique makes Gemini's readiness quite questionable.
It comes as no shock that Google denies any wrongdoing right here, because it referred The Verge to an X publish written by Gemini's co-lead, Oriol Vinyals, which says "all of the consumer prompts and outputs within the video are actual," and that his workforce made the video "to encourage builders." Given the trade and authorities' consideration on AI currently, maybe the tech large ought to be extra delicate about its displays on this subject.
Actually completely happy to see the curiosity round our “Fingers-on with Gemini” video. In our developer weblog yesterday, we broke down how Gemini was used to create it. https://t.co/50gjMkaVc0
We gave Gemini sequences of various modalities — picture and textual content on this case — and had it reply… pic.twitter.com/Beba5M5dHP
— Oriol Vinyals (@OriolVinyalsML) December 7, 2023
This text initially appeared on Engadget at https://www.engadget.com/google-admits-that-a-gemini-ai-demo-video-was-staged-055718855.html?src=rss
Trending Merchandise
