Google “admits” having edited the demonstration video of its Gemini AI but they defend its capabilities

It seems that Google is not good at presenting its new technologies. We already saw it with Bard, which had a series of errors in the presentation of functions shown, and it has also happened with its Gemini artificial intelligence.

After an impressive presentation of the new artificial intelligence Geminiwhich will compete directly against OpenAI’s GPT-4, those from Mountain View did not hesitate to affirm that their technology was far superior.

No further from the truth, just a few hours later the Bloomberg site pointed out that one of the videos that showed the benefits of AI was doctored.

This controversial video shows how Gemini works when answering questions in real time and describing drawings.

But this video was edited, stating that this demonstration was not done in real timenot even with voice, but it seems that they used frames of still images from the filming and text instructions.

Now after the controversy, the co-director of Google, Oriol Vinyalsstated that this particular video was intended to “inspire developers,” showing the potential of user experiences powered by Gemini.

“The video illustrates what multimodal user experiences created with Gemini could look like. We made it to inspire developers.”

Thus It admits some editing but on the other hand it also defends the technology by stating that everything seen in the demonstration video can be done by AI.

While there is some disclaimer in the description of this video demo on YouTube, many interpreted what was seen as being entirely in real time.