Google showcases a possible reply to GPT-4o forward of I/O

What you have to know

• OpenAI at the moment launched GPT-4o, which is a brand new AI mannequin that may work together with customers by means of audio, imaginative and prescient, and textual content.
• Google seems to have a counter for GPT-4o able to go for Google I/O 2024, which begins tomorrow.
• In a teaser, Google previewed an unreleased multimodal AI interface, and stated to count on bulletins associated to AI, Google Search, and extra.

---

OpenAI beat Google to the punch at the moment, internet hosting an occasion the place the corporate launched a new AI model called GPT-4o. The timing of OpenAI’s occasion was not by chance — it got here a few day earlier than Google was set to announce synthetic intelligence plans and options on the Google I/O 2024 developer convention. Nevertheless, it seems to be like OpenAI would possibly spend solely a day because the chief in multimodal AI. In a teaser posted to X (previously Twitter), we obtained a take a look at an unreleased synthetic intelligence interface working on an Android telephone. 

The principle a part of the teaser is a video that exhibits some form of multimodal AI interface working on an Android telephone. Visually, it seems to be sort of just like the Pixel Digital camera app, and the demo exhibits somebody asking it questions on their environment. On this case, it is the Google I/O stage. After accurately figuring out that setup was occurring for an occasion, this AI instrument additionally acknowledged the Google I/O wordmark and defined the small print of the developer convention. Nevertheless, since it is a pre-recorded video, outcomes need to be taken with a grain of salt. 

Alongside the video, Google says to count on “the newest information about AI, Search, and extra” at Google I/O 2024. The principle keynote is about for 10 a.m. PT, and we’ll be monitoring all of the developments in our stay weblog.

Gemini has already appeared as a voice assistant, and as a chatbot that may eat pictures, screenshots, and extra. Nevertheless, the brand new a part of Google’s teaser is that the interface now helps imaginative and prescient as an alternative choice to voice and textual content enter. By wanting by means of your machine’s digital camera, this AI interface can reply questions and particulars about your environment. It is much like what the Humane AI Pin and the Rabbit R1 tried to do with standalone units. 

Nevertheless, there’s nonetheless lots we do not know, like whether or not this performance could be constructed into Gemini or another app. The demo appears spectacular sufficient {that a} supercharged model of Gemini is likely to be prepared to exchange Google Assistant, however that is solely hypothesis for now. 

We’ll acquire extra readability on what Google has been engaged on throughout Tuesday’s keynote, however for now, this teaser appears to substantiate a GPT-4o competitor is coming.