A question of follow -up on the final score was answered properly, but Gemini had the name of the marker of the first hit badly: the AI suggested that it was Johan Dotson. Dotson has been shown by obtaining a touchdown in the strengths with the scores at 0-0, but it was excluded – an example of the nuances that AI does not necessarily resume.
Gemini successfully identified himself when the Kansas City chiefs obtained their first points, and even included a time -tumor connecting directly to the YouTube clip. He also obtained the name of the marker. It seems that Gemini strongly depends on the comment of the sports clips, which is not surprising.
Summarize video content
Then we tried to put Gemini against a Behind the scenes For the Grand Budapest Hotel, directed by Wes Anderson. The clip takes place at four and a half minutes, and Gemini withdrew some almost instantly answers: he identified the name of the film spoken, and the main rhythms of the story of the clip.
However, it all depends on the audio (or transcription) again – there is no analysis of the real video content. The AI could not say who were the speaking heads in the video, even if their names were shown on the screen, and could not say who was the director (even if it was also mentioned in the description of the video).
On the positive side, Gemini did an impressive job to summarize the audio of the video. He correctly identified some of the cinema challenges that were mentioned throughout and provided them with horodatages – the search for a set to represent the great Budapest, to fill it with extras.
Summarize interviews
Finally, we tried Google Gemini with an interview: Channel 4 in the United Kingdom to speak to Charlie Brooker and Siena Kelly of the last series of Black mirror (Perhaps suitable for an article on AI). Gemini turned out to be very capable of choosing discussion points and adding horodatages, although of course the whole video talks mainly.
Again, there is no context on anything outside of audio or transcription. Gemini Ai could not say where the interview took place, or how the participants acted, or anything else on the video visuals – which is worth keeping in mind if you use it yourself.
For the videos where the answers you want are in the audio of a YouTube video, and its associated transcription, Gemini works very well to summarize and provide precise responses (provided that the commentators mention when a touch is excluded, as well as when you are noted). For any type of visual information, you will always have to watch the video yourself.
