
Google shows off its next-gen AI assistant that can control your Android phone
Google
TL;DR Google's Project Astra aims to be a 'universal AI assistant' that understands context, devises plans, and acts on your behalf within Android.
A demo showcased Astra assisting with a bike repair by navigating a manual, finding a YouTube tutorial, and potentially contacting a shop.
This was powered by an AI agent that controls Android apps by simulating screen inputs, indicating an advanced but still developing capability.
Google's Gemini chatbot has come a long way since its initial debut at the end of 2023. At first, the chatbot struggled with even basic tasks its predecessor was well equipped to handle, like setting a reminder. Today, it can not only chain actions across multiple services but also answer questions about what's showing on your phone's screen or through your phone's camera. In the future, Gemini may even be able to control your Android phone for you, allowing it to search through your phone's documents and open your apps to find the information you're looking for.
During Google I/O 2025 earlier today, Google showed off its vision for a 'universal AI assistant.' A universal AI assistant is not only intelligent but can understand the context you're in, come up with a plan to solve your problems, and then take action on your behalf to save you time.
For example, if you're having a problem with your bike's brakes, you can ask the assistant to find the bike's user manual online, have it open the manual, and make it scroll to the page that covers the brakes. Then, you can follow-up and tell the assistant to open the YouTube app and play a video that shows how to fix the brakes. Once you've learned what parts you need to replace, you can ask the assistant to go through your email conversations with the bike shop to find relevant information on part sizes or even have it call the nearest bike shop on your behalf to see if the right parts are in stock.
There's currently no AI assistant that can do everything I just mentioned without some manual user intervention, but Google does offer various independent AI features that, if chained together, make this feat possible. Google's latest prototype of Project Astra, the code-name for its future universal AI assistant, demonstrates exactly that. Today's demonstration shows a man asking Astra on his phone how to fix a problem with his bike's brakes, with the assistant doing every step I just described in the previous paragraph.
What's particularly interesting about this demo is that it shows off an AI agent that Google developed to automate actions within Android apps. We've known that Google has an AI agent called Project Mariner that can control a web browser, but this is the first we've heard of an AI agent from Google that can control an Android phone.
In the demo, we see Google's Android AI agent open a PDF, scroll the screen until it finds the page requested by the user, open the YouTube app, and scroll through the search results until it finds a relevant video.
Google
When Astra is controlling the phone, we can see a small, circular overlay on the left. When it scrolls the screen, we can see the tap and swipe inputs it sends, showing that Astra is simulating screen inputs. Judging by the screen recording chip in the top left corner and the glowing overlay around the edges of the screen, it seems that Astra is reading the contents of the screen and then deciding where to tap or swipe.
Google
We don't know if Astra is doing these actions on device; it would certainly be possible through the use of the multimodal Gemini Nano model, but we can't determine if it's being used in this demo. What we do know is that Google has some work to do before it can roll out this Android AI agent. The portions of the video showing the agent were apparently sped up by a factor of 2, suggesting that it can be quite slow at taking action.
Still, we're excited to see Google get closer to achieving its vision of a universal AI assistant. Every update we get on Project Astra makes Gemini a more appealing product, as we can expect new Astra features to eventually trickle down to the chatbot. The new capabilities Google demoed today might not be available for some time, but they'll eventually go live in some form, and when they do, they might be even more impressive than what we're seeing today.
Got a tip? Talk to us! Email our staff at
Email our staff at news@androidauthority.com . You can stay anonymous or get credit for the info, it's your choice.

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles


Android Authority
13 minutes ago
- Android Authority
Google Translate is testing a revamped UI for its powerful new features (APK teardown)
Aamir Siddiqui / Android Authority TL;DR Google is preparing significant UI updates to the Translate app, changing things around to accommodate its upcoming AI-driven features better. We can spot a new translation results screen, with AI features like Insights, Regional variants, and Ask a question, but the screen no longer highlights the main result immediately. The upcoming UI's focus on AI raises concerns about the potential loss of user trust due to generative AI inaccuracies. Google Translate is an excellent app that has saved my day many times. It's dead simple to use, making it very easy to recommend to non-techies. Just like with the rest of its app portfolio, Google has been working to upgrade Google Translate with AI-driven features. It seems the company also intends to use the opportunity to revamp the simplistic UI, for better or worse. Authority Insights story on Android Authority. Discover You're reading anstory on Android Authority. Discover Authority Insights for more exclusive reports, app teardowns, leaks, and in-depth tech coverage you won't find anywhere else. An APK teardown helps predict features that may arrive on a service in the future based on work-in-progress code. However, it is possible that such predicted features may not make it to a public release. Google Translate v9.10.70 includes code for a substantial UI change. We managed to activate the upcoming UI to give you an early look before its release. Current Google Translate UI Upcoming Google Translate UI The landing page could receive a small refresh, mainly by adding the upcoming Practice button. When that comes in, the voice input button will shrink a bit in size and move to the right instead of occupying the center position in the row. We also see a bookmarks button at the top for Saved content, alongside the star button for favorites, though both may not co-exist when the bookmarks button does roll out. Current Google Translate UI Upcoming Google Translate UI Current Google Translate UI Upcoming Google Translate UI As you can see, the translation results page is on track to get a completely different look. While the current UI highlights the closest translation with a bigger text size, the upcoming UI presents a list of possible translations, with no particular highlights standing out. The autocomplete bar can show up in the upcoming UI, though. Once you click on a result, the current UI presents that as a highlight, with more clarifying meanings for the word. On the other hand, selecting any result in the upcoming UI marks it as a highlight and presents a slew of AI tools, powered by Gemini of course, in a bottom sheet. These tools include buttons for Alternative translations, Definition, Regional variants, and Ask a question. Current Google Translate UI Upcoming Google Translate UI While 'Alternative translations' and 'Definition' already exist in the current UI, the 'Regional variants' and 'Ask a question' features are new. When you look up a phrase, the Alternative translations button becomes the Insights button, which gives us more insight into its usage. Overall, the upcoming UI prepares Google Translate to incorporate AI-driven features meaningfully. However, in the process, the app loses its simplicity, which is something that many of us appreciated about Translate. There's also the disclaimer at the bottom of AI-driven results about how generative AI can make mistakes — we've seen Google Search's AI Overviews give confident yet wrong meaning to gibberish idioms, and we fear the same fate for Google Translate. Leaning too hard into imperfect AI can cause a loss of trust from consumers, and we hope Google has adequately accounted for that with these upcoming changes, if and when they roll out widely. Got a tip? Talk to us! Email our staff at Email our staff at news@ . You can stay anonymous or get credit for the info, it's your choice.
Yahoo
17 minutes ago
- Yahoo
Hackers abuse modified Salesforce app to steal data, extort companies, Google says
By AJ Vicens (Reuters) -Hackers are tricking employees at companies in Europe and the Americas into installing a modified version of a Salesforce-related app, allowing the hackers to steal reams of data, gain access to other corporate cloud services and extort those companies, Google said on Wednesday. The hackers – tracked by the Google Threat Intelligence Group as UNC6040 – have 'proven particularly effective at tricking employees' into installing a modified version of Salesforce's Data Loader, a proprietary tool used to bulk import data into Salesforce environments, the researchers said. The hackers use voice calls to trick employees into visiting a purported Salesforce connected app setup page to approve the unauthorized, modified version of the app, created by the hackers to emulate Data Loader. If the employee installs the app, the hackers gain 'significant capabilities to access, query, and exfiltrate sensitive information directly from the compromised Salesforce customer environments,' the researchers said. The access also frequently gives the hackers the ability to move throughout a customer's network, enabling attacks on other cloud services and internal corporate networks. Technical infrastructure tied to the campaign shares characteristics with suspected ties to the broader and loosely organized ecosystem known as 'The Com,' known for small, disparate groups engaging in cybercriminal and sometimes violent activity, the researchers said. A Google spokesperson did not share additional details about how many companies have been targeted as part of the campaign, which has been observed over the past several months. A Salesforce spokesperson told Reuters in an email that 'there's no indication the issue described stems from any vulnerability inherent in our platform.' The spokesperson said the voice calls used to trick employees 'are targeted social engineering scams designed to exploit gaps in individual users' cybersecurity awareness and best practices.' The spokesperson declined to share the specific number of affected customers, but said that Salesforce was "aware of only a small subset of affected customers," and said it was "not a widespread issue." Salesforce warned customers of voice phishing, or "vishing," attacks and of hackers abusing malicious, modified versions of Data Loader in a March 2025 blog post. Sign in to access your portfolio


TechCrunch
19 minutes ago
- TechCrunch
TC Sessions: AI Trivia challenge for tickets ends tonight
Tonight's your last shot: Take on the TechCrunch Sessions: AI Trivia Countdown and walk away with two tickets for just $200. The window closes tonight at midnight — go! Think you know AI? Can you name the company that created the AI assistant known as Siri before Apple acquired it? Or the name of Google's large language model family released in 2023? If so, now's your time to shine. Ready to compete? Prove your AI smarts and win your spot at TC Sessions: AI Tackle a short round of AI trivia before tonight ends, and you could earn a special offer — check your inbox if you win. How to play Step 1: Take today's AI Trivia Countdown quiz. Step 2: Check your inbox to see if you've scored the special code. Step 3: Use the code to grab 2-for-1 tickets to TechCrunch Sessions: AI. Join the future of AI at TC Sessions: AI — tomorrow at UC Berkeley's Zellerbach Hall. Play the trivia. Score the deal.