Google is under scrutiny due to the recently launched explanatory video of the artificial intelligence model Gemini.
The company introduced a new artificial intelligence model considered the largest and most powerful, named “Gemini,” weeks before the end of the year, and provided an explanatory video for the media and the public.
The video, lasting six minutes, includes verbal conversations between the user and Gemini-enhanced chatbot. The video also showcases Gemini’s ability to recognize visual images and physical objects and distinguish between them.
In the video clip, Gemini demonstrates its ability to express clearly and loudly about duck drawings and describing the difference between a real duck drawing and a rubber duck, among other examples.
The company described itself on YouTube in a short sentence saying: “We have reduced the access time in the explanatory presentation, and summarized Gemini’s outputs for brevity.” However, this description does not include a disclaimer in the video itself.
After the launch, the company confirmed to Bloomberg agency that it did not execute the explanatory presentation in real-time using spoken words but instead used static images and written words Gemini system responded to.
A Bloomberg article pointed to a significant difference from what Google suggested, as the company proposed in its video that a person could have a smooth voice conversation with Gemini while browsing and responding to the world around in real-time.
The company’s statement read: “The video serves as a means to illustrate the interaction capabilities with Gemini, based on realistic requests and test results. We look forward to seeing what people will do once access to Gemini Pro opens on December 13.”
Gemini’s video altered Google’s previous showcase memories, as explanatory demonstrations usually get altered to avoid any technical obstacles direct presentations might face.
Earlier this year, Google faced criticism due to the rushed and failed demonstration of its AI chatbot, coinciding with Microsoft’s plans to integrate Bing with ChatGPT in the same week.
A previous report this month indicated that Google had canceled several personal events plans for launching Project Gemini and instead decided to launch it virtually.
Google is in strong competition with OpenAI’s GPT-4, an advanced and successful model to date.
This week, Google released a technical report claiming that Gemini Ultra outperforms GPT-4 in several criteria.
