According to foreign media reports, it is claimed that the current humanStrongestartificial intelligence (AI)Large ModelGemini 1.0 has only been released for a day and is being questioned about fakery, while theGoogleIt's been acknowledged, too.
To confirm how powerful its own artificial big model is, Google went live yesterday with a video demonstrating Gemini's multimodal capabilities, which has since garnered millions of views, though some have questioned the performance as "fake.
From what the media has questioned, Gemini's ability to quickly determine physical objects in the frame and give human-like voice feedback highlights the amazing multimodal capabilities, which, from actually being hands-on, is simply not achievable.
Subsequently, the head of Gemini denied intentional fakery, claiming that the reaction time was simply shortened for brevity, admitting that the video was not in real time, but used still image frames from the original footage, and then wrote text cues.
All this post-processing by Google only shows that they are trying too hard to make the Gemini look so much better than the competition, but to do so would be outright false advertising.