Users’ trust in Google’s technology and integrity may be shaken upon discovering that the highly publicized Gemini demo was essentially a sham, despite its grand unveiling.
The video titled “Hands-on with Gemini: Interacting with bidirectional AI” garnered a million views within a day, showcasing the bidirectional model’s adaptability and responsiveness to various inputs. It commences with a duck sketch evolving from a squiggle to a completed drawing, followed by the surprising appearance of a toy orange bird, eliciting a whimsical “What the quack” reaction. The demonstration proceeds to address queries about the toy, showcasing activities like rearranging planet sketches, interpreting shadow puppet gestures, and engaging in cup-switching games.
While the video acknowledges reduced latency and shortened Gemini outcomes, it emphasizes the model’s flexibility in responses. Despite the impressive display of bidirectional knowledge, the realization dawns that the hands-on experience was not genuine.
The video’s fabrication is evident, as it was pieced together using still image frames and text prompts to simulate interactions with Gemini. The discrepancy was highlighted by Bloomberg’s Parmy Olsen, exposing the deceptive nature of the presentation.
Although Gemini may possess some capabilities depicted in the video, the live execution fell short of the implied seamless interactions. The scripted prompts and carefully chosen visuals crafted a misleading portrayal of Gemini’s conversational abilities.
Notably, discrepancies arise in the video’s depiction of interactions, such as the swift recognition of hand gestures as a game of Rock, Paper, Scissors, and the evaluation of drawing sequences. These instances underscore the manufactured nature of the interactions, raising doubts about the model’s true capabilities.
The video’s lack of transparency regarding its staged nature and the misrepresented interactions casts a shadow over Google’s credibility in showcasing its AI advancements. The misrepresentation calls into question the authenticity of Google’s AI demonstrations and the accuracy of their claims.
In conclusion, the fabricated nature of the Gemini demo video raises concerns about the transparency and integrity of Google’s presentations. The discrepancy between the depicted interactions and the actual capabilities of the model undermines the trust in Google’s technological advancements.
