Artificial intelligence (AI) has gone viral over the past several months, spurred on by the launch of ChatGPT last November. The ensuing AI arms race has pushed AI into the headlines, with the technology enjoying a renaissance.
Nowhere has that been more acutely obvious than in the halls of big tech. Microsoft ignited the battle with its $13 billion investment in ChatGPT creator OpenAI and the ensuing integration with its Bing search engine.
Not willing to cede years of experience, Alphabet's (GOOGL 0.72%) (GOOG 0.81%) Google kicked off its 2023 I/O developer conference with a multitude of announcements and more than 35 blog posts and press releases regarding the infusion of AI into everything the company debuted. Here's a rundown of the AI-laden barrage of announcements.
New and improved generative AI
Large language models (LLMs) provide the foundation for generative AI, with the technology fueling next-generation chatbots and a host of other AI capabilities.
Google announced the latest version of its Pathways Language Model (PaLM) LLM, which has been "heavily trained" on multilingual text. This new and improved PaLM 2 has the ability to "understand, generate, and translate nuanced text -- including idioms, poems, and riddles." Perhaps most importantly, the new model has been trained on mathematical and scientific papers, as well as many more languages, imbuing it with better reasoning capabilities and more common sense. It also has the ability to pass "language proficiency exams at the mastery level."
The company also created a scaled down version of PaLM 2 that can run on its smartphones.
A facelift for Bard
PaLM is already being used by Bard, the company's chatbot, to write code, debug existing code, and even explain code. Now, Bard is getting a makeover. Google's flagship chatbot has been upgraded to the new PaLM 2 LLM, which will soon provide even more images -- alongside text -- in response to queries. The company is also integrating Google Lens directly into Bard, which will soon allow users to include images into their prompts, resulting in more robust and creative responses.
Furthermore, after a limited release in March, Google is removing the wait list and lifting the restrictions to use Bard, opening it up to the public in more than 180 countries and territories.
Infusing AI into everything
Google has infused its underlying AI technology into 25 products and features.
The tech powers a new feature in Workspace called "Help Me Write," which will help draft responses in Gmail; write sales pitches, essays, and job descriptions in Google Docs; and fill out spreadsheets in Google Sheets. Users can create images from text descriptions in both Google Slides and Meet. Maps will be able to render not only the route, but also traffic conditions and weather. A new Google Photos tool, Magic Editor, will help reposition subjects in photos while expanding images that were only partially included in the original shot.
The company also announced a host of new AI-powered tools coming to the cloud. One of the most important of these was Duet AI for Google Cloud, a collaboration tool that uses machine learning to help cloud users program and write code, offer suggestions, and help review and inspect the completed code. This is just the beginning, as Google promises additional tools in the coming months.
New and improved Search
Improvements will also be coming to Google Search. Over the next few weeks, the company will debut a new feature called Perspectives. When a user searches for a topic that "might benefit from the experiences of others," the Perspectives filter will show up in the results. Tapping the filter will provide videos, social media and discussion boards posts, and short- and long-form videos, to help provide a more rounded view of the topic.
Google will also begin testing a souped-up version of its classic Search, powered by generative AI. The public will be able to sign up to test the new capabilities in Search Lab, which will soon be opened for early access. Users can ask complex questions like this example provided by Google: "What's better for a family with kids under 3 and a dog, Bryce Canyon or Arches?" The response will include pros and cons for both options as well as next steps to further explore the topic.
Search will also help with complicated decisions like product choices and purchases, by including images, product descriptions, ratings, reviews, and prices.
Access to Search Labs will initially be offered via Chrome Desktop or the Google App in the U.S.
AI for Android and Pixel
Finally, Google is working on a host of generative AI tools for its Pixel and Android-powered devices. The company is introducing a new collection of generative wallpapers, custom emojis, and a cinematic mode that creates short movies from users' photos. The aforementioned Magic Compose will also offer AI-fueled suggestions for responding to texts, resulting in entire sentences, rather than just words or short phrases. In all, the company said it redesigned more than 50 apps to enhance the user experience, with AI at the center of the improvements.
Google said the updates will be rolled out across 3 billion active Android devices, including smartphones, watches, tablets, cars, and even TVs.
A whole new world
The debut of ChatGPT captured the public imagination and the infusion of its technology into Microsoft's Bing led some investors to conclude that Alphabet had fallen behind in the AI race. This ignored the company's long history and existing AI expertise.
Yesterday's host of announcements illustrates that when it comes to AI, investors shouldn't be too quick to dismiss Google, as the company has a deep roster of AI in its pipeline.