Google I/O 2024 kicks off in just over a day and, ahead of the event, Google has shown off a pretty impressive new conversational Gemini prototype in action which seems to use live video.
AI chatbots have mostly focused on text and image-based prompts to date, but the dream for these multimodal assistants so much bigger. In a new demo, Google shows off a new version of Gemini that will presumably be detailed more fully during tomorrow’s keynote.
Apparently filmed during I/O’s setup, this demo shows Gemini on a Pixel using live video along with spoken prompts to give information.
Gemini is asked “what do you think is happening here,” to which it replies, correctly, that it is looking at a stage for a large event being set up. From there, Gemini prompts the question “is there anything in particular that caught your eye,” which naturally pushes the conversation forward. When asked about the letters on screen, Gemini responds saying they’re for Google I/O and offers a brief description of the event.
The demo as a whole is rather impressive, not just because of the multimodal use of voice and video in the prompts, but also in just how naturally the conversation is carried.
That said, it’s worth noting that Google previously showed a very similar conversational Gemini demo that was later detailed to show that it was a little too good to be true. It’s unclear if the same thing is happening here, but the UI shown on screen clearly shows that this is using video, and Google says that this is a “prototype.”
It’s pretty easy to see why Google released this teaser today. The video was uploaded to Twitter/X less than an hour before an OpenAI event where ChatGPT picked up the same functionality Google teased, all for free.
Stay tuned for 9to5Google’s full coverage of Google I/O this week, where we’re expecting plenty of new announcements around Gemini.
More on Gemini:
- Gemini’s ‘Memory’ feature inches toward launch, will remember things about you
- Are you subscribed to Google One AI Premium for Gemini?
- Gmail adding voice input, Gemini for Google Chat, Meet ‘Translate for me,’ & more
Follow Ben: Twitter/X, Threads, and Instagram
FTC: We use income earning auto affiliate links. More.
Comments