In recent days Google did it the official presentation of Gemini AIthe first Artificial Intelligence system created by the company which promises to intervene on different types of information and content, such as images, videos, audio and much more, and to be compatible with a large number of different devices, including smartphones.
The launch, however, was not without perplexity and controversy. In particular, some details of a video shown during the presentation did not go unnoticed. This is about six minutes of content in which Gemini’s multimodal capabilities were demonstrated. It was the newspaper that raised the doubt Bloombergwho stated that Gemini’s capabilities shown in the content appear to be high, but not very credible. Google’s response came immediatelywhich sounds like an admission of guilt.
Google, doubts after the presentation video
In the launch video Google highlighted Gemini’s impressive capabilities. Among other things, the content demonstrates how the system is able to verbally describe some drawings, highlighting details and peculiarities.
In the video description of YouTubethe platform on which the content was published, a notice was inserted: “For the purposes of this demo, Gemini’s latency and outputs have been reduced for brevity”. Disclaimer not shown anywhere else within the video itself and that, according to Bloombergwould be far from reality.
Gemini then appeared as a perfectly capable AI-based system interact in a natural and fluid way with the surrounding worldresponding to stimuli quickly and consistently.
Google’s response to the controversy is an admission: the footage was not in real time
After submitting the video for investigation, Google confirmed to the newspaper that the demo is not the result of real-time shooting with the use of vocal suggestions, but it is the result of combining still images from an initial shot, in which Gemini responded to suggestions and text messages, subsequently edited.
The purpose of the videoGoogle leaders told CNBC, it was purely illustrative and just wanted to show the public the future possibilities of interaction and user experience. If without the announcement Gemini seemed capable of responding fluidly, but after Google’s clarification, made necessary by criticism, the presence of some limits is now hypothesized.
The real strength of Gemini and its ability to compete with advanced Artificial Intelligence systems, including OpenAi’s ChatGPT, at this point, it will only be understandable after the first tests by userswhich will be able to highlight its potential, weaknesses and improvements to be made.
To know more: Artificial Intelligence: what it is and what it can do for us