Last May, Google CEO Sundar Pichai said the company would use artificial intelligence to reimagine all of its products.
But new generative AI technology comes with risks, including the spread of misinformation, so Google is wary of applying it to its search engine, which is used by more than 2 billion people and generated $175 billion in revenue last year. Ta.
At Google's annual conference in Mountain View, Calif., on Tuesday, Pichai showed how the company's aggressive AI efforts have finally trickled down to its search engine. Starting this week, users in the U.S. will see AI Overviews, a feature that generates summaries of information on top of traditional search results, he said. By the end of the year, more than 1 billion people will have access to this technology.
The introduction of AI could heighten concerns that web publishers will see less traffic from Google search, further increasing pressure on an industry reeling from rifts with other technology platforms. Google allows users to view long summaries on topics, potentially reducing the need to visit separate websites, but Google downplayed those concerns.
“Links included in AI summaries receive more clicks from users than when they appear in traditional search results,” Liz Reid, vice president of search at Google, said in a blog post. “We remain focused on sending valuable traffic to publishers and creators.”
The company also announced a number of other initiatives, including lightweight AI models, new chips and so-called agents to help users perform tasks, in a bid to gain an edge in the AI battle with Microsoft and ChatGPT maker OpenAI. . .
“We're in the very early stages of transitioning to an AI platform,” Pichai said Tuesday at Google's I/O developer conference. “We want everyone, including developers, startups and the general public, to benefit from what Gemini can do.”
When ChatGPT was released in late 2022, some in the technology industry believed it was a serious threat to Google's search engine, the most popular way to find information online. Since then, Google has been actively working to regain its dominance in the AI space, releasing a technology family called Gemini that includes new AI models for developers and chatbots for consumers. We also brought this technology to YouTube, Gmail, and Docs, allowing users to create videos, emails, and drafts with less effort.
Meanwhile, the tit-for-tat race between Google and OpenAI and its partner Microsoft continued. The day before Google's conference, OpenAI announced a new version of ChatGPT that resembles a voice assistant.
(The New York Times sued OpenAI and Microsoft in December for copyright infringement of news content related to AI systems.)
At a Silicon Valley event, Google showcased how it's embedding AI more deeply into users' lives. It introduced Project Astra, an experiment to see how AI can act as an agent that chats with users by voice and responds to images and videos. Some of the features will be available to users of Google's Gemini chatbot later this year, Demis Hassabis, CEO of Google's AI research lab DeepMind, said in a blog post.
DeepMind also announced Gemini 1.5 Flash. It's an AI model designed to be fast and efficient, yet lighter than Gemini 1.5 Pro, a mid-tier model that Google has deployed across many consumer services. Dr. Hassabis wrote that the new model had “very strong” reasoning abilities and was good at summarizing information, chatting, and captioning images and videos.
The company announced another AI model, Veo, that generates high-resolution video based on simple text prompts, similar to OpenAI's Sora system. Google said some creators will be able to preview Veo, and others will be able to join a waitlist for access to Veo. Later this year, the company plans to bring some of Veo's features to YouTube Shorts, a competitor to his video platform TikTok, and other products.
Google also unveiled the latest versions of its music generation tool Lyria and image generation tool Imagen 3. In February, Google's Gemini chatbot came under fire from users on social media for refusing to generate images of white people and displaying historically inaccurate images. numbers. The company said it is blocking its ability to generate images of people until the issue is resolved.
The company says more than 1 million users have signed up for Gemini Advanced, a version of Google's chatbot available for a $20 monthly subscription, in the past three months.
In the coming months, Google plans to add Gemini Live. This provides users with a way to interact with chatbots through voice commands. Google says the chatbot will respond with a natural voice and users can interrupt Gemini to ask clarifying questions. Later this year, users will be able to use their camera to view the physical world around them in Gemini Live and have conversations about it with a chatbot.
In addition to the AI overview, Google's search engine displays an AI-organized search results page with generated headings that highlight different types of content. The feature will start with results for meals and recipes, and later extend to shopping, travel, and entertainment queries.
Search chief Reid said in an interview before the conference that he expects the search updates will save users time because Google “does more of the work for you.”
Pichai said he expects the majority of people to use Gemini AI technology through Google's search engine.
“We're going to make it increasingly seamless for people to interact with Gemini,” Pichai said in a pre-conference briefing.
