Google has been accused of falsifying a video demonstrating Gemini's AI capabilities
Google published a video demonstrating the work of a new multimodal model of artificial intelligence Gemini. The video quickly gained a million views, but it soon became clear that it is actually a production.
Here's What We Know
In the video, Gemini responds to voice commands, recognises images and gestures, which looks impressive for an AI system. However, journalists found that the model responded to the questions asked with pre-prepared text, meaning the interaction was fabricated.
A Google spokesperson said that the video shows the real results of Gemini's work. At the same time, he admitted that edits were made to improve the demonstration. Nevertheless, according to experts, the video misleads viewers and does not conform to the stated "Hands-on" format.
The incident has shaken the tech giant's credibility. So far, Google is lagging behind its competitors in AI, so the company's reputation as a reliable partner has come into question, experts said.
According to observers, such productions will only hurt Google. Instead of honestly demonstrating the current capabilities of Gemini, the company went for falsification. This could have a negative impact on Google's image and trust in its AI developments.
Go Deeper:
Source: TechCrunch