Google’s I/O: News and Updates

Search Engine Optimization and technology

By Emil Bjerg, journalist and editor

Just 24 hours after OpenAI’s spring update, Google held its annual developer conference, I/O, in Mountain View, California. The competition between OpenAI and Google was further emphasized by the similarities in their recent innovations. Here we’ve gathered the most essential news and updates.

Project Astra

The big news first. At I/O Google unveiled Project Astra, an AI assistant, that is designed to assist users in a wide range of contexts. Demis Hassabi, CEO of Google DeepMind, said that he had long had a vision of a “universal assistant. It’s multimodal, it’s with you all the time.[…] It’s that helper that’s just useful. You get used to it being there whenever you need it.”

Commentators were fast to draw comparisons between Project Astra and the new GPT-4o. As Wired writes Google’s “vision for the future of AI is strikingly similar to one showcased by OpenAI on Monday.” They share a lot of the same features: the models are natively multimodal, meaning that they are built to handle input and give output in different formats: a camera, voice, text, and even imagery. This allows Project Astra to understand your environment and requests through various means.

AI Overview in Google Search

Google aims to continuously dominate search with their  ‘AI Overviews’, taking up the competition with AI-powered search engines like Perplexity. A Gemini model specialized for search will summarize answers so users can get insights without using Google’s famous 10 blue links.

Google hopes that AI-powered overviews can provide users with more concise and relevant information, reducing the need for extensive browsing through search results. Critics and website owners on the other hand fear that the overview feature will dramatically reduce traffic to websites given that users will now be able the find information directly on Google. The feature will roll out in the coming weeks in the US and in the EU by the end of the year.

Gemini updates and integration

Several updates to Google’s large language model Gemini were announced.

Google announced a new Gemini model: Gemini 1.5 Flash. The new multimodal model is just as powerful as Gemini 1.5 Pro, but it’s optimized for designed for tasks that need quick responses, like answering frequent questions or completing simple actions. Additionally, Google has made updates to the base Gemini 1.5 model, enhancing its capabilities in language translation, logical reasoning, and coding.

An overarching theme was Google’s drive to tightly integrate AI across all its products and services. From Project Astra as a potential universal assistant to Gemini’s integration into various Google products, Google clearly aims to make AI a seamless part of the user experience. In line with that Google will roll out Gemini 1.5 Pro in their Workspace products: Docs, Sheets, Slides, Drive, and Gmail.

Gemini 1.5 Pro will be available to paid Google Workspace subscribers next month, offering an AI assistant that can access and summarize information from your Drive, draft emails using relevant content, and remind you to reply to emails.

Google Chrome

Google also announced that it will roll out Gemini Nano to the desktop version of Chrome. The inbuilt AI can help with writing emails, social media posts or product reviews directly from Chrome.

With Google’s Project Astra closely resembling OpenAI’s GPT-4o and with OpenAI rumored to enter the search market we can expect the competition and the fast AI innovation to continue.

LEAVE A REPLY

Please enter your comment!
Please enter your name here