Google has revealed a sneak peek at its next annual I/O conference, revealing that its Gemini AI has picked up some new skills.
The Gemini AI models developed by Google and their integration with Workspace and Chrome were a major topic of discussion throughout the event.
This Google I/O presentation had no mention of any hardware announcements or teases, in contrast to prior ones. Though none of it materialised, it was previously thought that we may get an early peek at the Pixel 9 series or the Pixel Fold 2, which some are referring to as the Pixel 9 Pro Fold.
In case you missed it, have a look at OpenAI’s GPT-4o demo, which showcases remarkably similar abilities: a real-time AI conversing with a person and discussing their environment. An AI arms race is about to begin, particularly with Apple aiming to include ChatGPT into iOS 18 (Gemini has been chosen by Samsung, Oppo, and OnePlus).
According to updates from Google after the event, Gemini 1.5 Flash, a new AI model has been added to its portfolio. Although the new multimodal model is tailored for “narrow, high-frequency, low-latency tasks,” it is just as powerful as the Gemini 1.5 Pro.
This improves its ability to produce quick answers. Google claims that the modifications it made to Gemini 1.5 will enhance its reasoning, coding, and translation capabilities.
Additionally, Google claims to have quadrupled the context window (or amount of data) of Gemini 1.5 Pro from 1 million to 2 million tokens.
Some of the updates include:
- Project Astra: Google recently introduced Project Astra, an AI assistant of the future that responds to your queries contextually by using voice recognition and phone-captured video.
- Google Workspace: Gemini, which can condense emails that are part of larger email chains, is being extended out to even more well-known Google services, such as Gmail.
- Google Veo: Based on your request, Google Veo can produce realistic, detailed 1080p films using generative AI. Imagen 3 can produce images in the interim by responding to text commands.
- Google Search AI: This week, Google will make “AI Overviews”—a mouthful that was previously known as “Search Generative Experience”—available to all US users. From now on, a “specialised” Gemini model will create and add condensed responses from the internet to results pages.
- Android AI-powered scam detection: Google claims that Android phones will be able to assist you in avoiding scam calls by using on-device Gemini Nano AI intelligence to scan for warning signs, such as typical scammer talking patterns, and then display in-the-moment alerts such as the one above.
- Introducing Trillium: Google announced the 6th generation of TPUs, called Trillium which was claimed to be high performant and efficient TPU to date.
Other announced updates include updates in its SynthID AI watermarking, Google Chrome AI assistant, Circle to Search math problems solver, etc.
“We will keep making the investments necessary to advance AI innovation and deliver state-of-the-art capabilities,” Google wrote on its official page.