Google Faces Criticism After Revealing Gemini's First Demo Video

Google is facing its first setback for its latest AI model, Gemini, just moments after the company released its first demonstration video.

Google Shows Off Gemini's Capabilities

During its launch, Google presented a six-minute video following a conversation between a user and a Gemini-powered chatbot. The short demonstration boasted the AI model's capability to recognize and analyze information based on pictures, physical objects, and more.

"For the purposes of this demo, latency has been reduced and Gemini outputs have been shortened for brevity," the demo description read on YouTube as it showed various ways that Gemini can be used.

Google Deepmind introduced Gemini as the first model to outperform human exports on Massive Multitask Language Understanding (MMLU). The video fully demonstrated its capability to do various tasks and analyses through various types of prompts.

Google Admits Demo was not Conducted Real-Time

Later on, the company confirmed to Bloomberg that the demo was not conducted in real-time and that the team used still images and text prompts so that Gemini would respond. Meanwhile, Google defended that it was just a "depiction of possibilities of interacting with Gemini."

"We look forward to seeing what people create when access to Gemini Pro opens on December 13," the company added. Earlier this month, reports revealed that Google dropped its plan to do an in-person launch for Gemini and settled for a virtual event only.

Gemini Pro will be available starting December 13 while Gemini Nano is only open for a sign-up for early preview. On the other hand, Google is still finalizing the safety checks for Gemini Ultra.