Google is gearing up to share a lot of AI and search news at its I/O Developer Conference on May 14, and we’re expecting a ton of announcements regarding Gemini, Android, search and more. The company couldn’t even wait for its own keynote to let us know what it’s up to, and has already teased us with a video of an intriguing camera-based AI feature on its social accounts.
From what we know so far, this is shaping up to be an hour-long launch event filled with announcements. If you can’t watch the Google I/O 2024 keynote for that long, or if you prefer a text and images approach, we have the live blog for you. Our reporters Karissa Bell and Sam Rutherford will be at Shoreline Amphitheater to bring you the news live, with support from the entire local Engadget team. Come back around 12 p.m. ET on May 14 to stick with us as we cover all the Google I/O 2024 updates!
Live129 updates
Gemini will be accessible in the side panel on Google apps like Gmail and Docs
Google
Google is adding Gemini-based AI automation to more tasks in Workspace. In its Google I/O keynote on Tuesday, the company said its advanced Gemini 1.5 Pro would soon be available in the Workspace side panel as “connective tissue between multiple apps with workflows powered by the AI” as AI gets smarter, learns more about you, and further automates your workflow.
Read the full story here.
Thank you for joining us. We’ll try to take a closer look at some of the things Google announced today at I/O.
Oh, and that Sundar increased the AI’s final tally to 121 once again.
Who wants to bet if Google will surpass this figure next year?
Honestly, I’m just glad it didn’t last longer than that. While Sam and Karissa check out the demos that might be available at I/O 2024, come join me and lead reviewer Devindra Hardawar on the Engadget YouTube channel. I’m sure you have thoughts and feelings to share about everything Google just announced!
And that’s the end of the keynote, it was *a little* less than 2 hours of all things Gemini and AI.
Even Google makes jokes about how many times people have talked about AI. (Photo by Sam Rutherford/Engadget)
120 mentions of “AI” in this keynote, according to Sundar (who has just returned to the stage) and Gemini. In fact, I would have guessed a little higher.
Gems will use the new LearnLM model, based on Gemini. (Photo by Sam Rutherford/Engadget)
Google, like most of its AI peers, plans to use watermarking to increase the transparency of AI-generated content with SynthID. These watermarks will extend to AI-generated video (remember previous Veo demos?) and text, which is particularly interesting because AI-generated text is much more prevalent (and in some ways more difficult to detect than images or video).
SynthID is a tool created by Google to easily detect AI-generated content. But it also looks like something from the Blade Runner universe that people can use to detect replicants.
We cover Building AI Responsibly, with an overview of Google’s approach to red-teaming, the threat hunting process, “problematic” results, and other issues that could cause trouble for Google (and maybe the rest of us). ). After the Gemini image generator fiasco earlier this year, I’m actually surprised they waited this long to bring this all up.
James Manyika at Google I/O 2024 talks about AI ethics and responsibility. (Photo by Sam Rutherford/Engadget)
So it looks like we’ve reached the developer-focused section of the I/O keynote. Karissa and I were just talking about the fact that for non-developers, it can be difficult to contextualize the quantities of tokens available in the different Gemini plans.
We delve deeper into Gemini Flash, who we first heard about at the top of the talk. Flash is the “lighter” Gemini model, which has lower latency and is more efficient. It appears this is aimed at developers and organizations who don’t need all the power of Gemini 1.5 Pro but still want to take advantage of the multimodal capabilities.
Yes, I think the scam alert feature is really nice because it can give you real-time insight when you might feel like a call seems a little suspicious.
Google Gemini can power a virtual AI teammate with their own Workspace account
Google
Google’s Gemini AI systems can do a lot, judging from today’s keynote/release. This includes the ability to set up a virtual teammate with their own Workspace account. You can configure Teammate to perform specific tasks, such as monitoring and tracking projects, organizing information, providing context, identifying trends after analyzing data, and playing a role in team collaboration.
Read the full story here.
Josh Woodward is back to talk about the different Gemini models available to developers.
Google offers another solution to endless scam calls. If you pick up and it detects that the call is “suspicious” or likely a scam, an important alert may appear directly on your phone, saving you from purchasing thousands of dollars of gift cards or transfer funds to a malicious account. Many of these scams seem obvious, but people still fall for them and this might help.
TalkBack Updates (Photo by Sam Rutherford/Engadget)
For example, will we ever see the day when you’re allowed to bring a phone to take the SATs? Because otherwise, you are potentially giving students help that they may not be able to use on important assessments, which could end up being a disadvantage.
Karissa, I think Circle to Search’s support formulas are weird because to older guys like us (or at least me) it seems like cheating.
