Football managers spend countless hours analyzing corners, free kicks, and player positioning in search of tiny competitive advantages. Google DeepMind believes artificial intelligence can make that process significantly faster, and its latest project, TacticAI, is designed to do exactly that. TacticAI is a football-specific AI assistant capable of modeling player movement, forecasting future play dynamics, and even recommending tactical adjustments for corner kicks. One of its standout abilities is predicting player trajectories up to eight seconds into the future using only broadcast-style visual data.

TacticAI was built with Liverpool FC and validated by football experts

Unlike general AI models, TacticAI focuses specifically on football tactics. Using geometric deep learning, the system analyzes the positions and interactions of players during corner kicks before generating predictions about what could happen next and suggesting alternative player arrangements that may improve outcomes.

Perhaps more importantly, the model wasn’t just tested in a lab. Google says its usefulness was evaluated through a qualitative study with football experts at Liverpool FC, who compared the AI’s recommendations against real match scenarios. According to the published research, experts preferred TacticAI’s suggested tactical setups 90 percent of the time over the original match configurations, highlighting the system’s practical value rather than just its statistical performance.

The benchmarking results are equally impressive. TacticAI outperformed existing baseline models in predicting both the likely receiver of a corner kick and whether a shot would occur afterward, while also generating realistic alternative player layouts that closely resembled genuine professional match situations.

This could be much bigger than football

The research is already moving beyond the lab. Google DeepMind has now announced a partnership with Brazilian football club Palmeiras, making it the first team to meaningfully build on TacticAI to simulate on-field scenarios and predict open-play dynamics up to eight seconds in advance. If successful, it could mark the beginning of AI becoming a genuine tactical assistant on the sidelines, not just another analytics tool running in the background.

We’re teaming up @Palmeiras, the first football club to meaningfully build upon TacticAI: our AI system that can help simulate field scenarios and predict open play dynamics up to 8 seconds in advance. pic.twitter.com/M3Krejk9Er

— Google DeepMind (@GoogleDeepMind) June 11, 2026

What’s more, is that the underlying technology has applications far beyond sports. Similar predictive models could one day assist autonomous robots, traffic systems, logistics planning, or any environment where understanding and forecasting coordinated movement is critical. And perhaps that’s the most fascinating part of TacticAI. On the surface, it looks like an AI built to help coaches win football matches. Underneath, it may be quietly laying the groundwork for machines that understand and anticipate complex real-world interactions before they unfold.

Google DeepMind built an AI that can predict football plays before they happen. TacticAI uses geometric deep learning to model player movement, forecast dynamics up to eight seconds into the future, and recommend tactical adjustments, all from broadcast-style visual data. Brazilian club Palmeiras is the first to use it for live open-play analysis.

The system was originally developed with Liverpool FC and validated through a qualitative study with the club’s football experts. They compared TacticAI’s recommended tactical setups against real match configurations. The experts preferred the AI’s suggestions 90% of the time. The published research, in Nature Communications, showed TacticAI also outperformed existing models in predicting corner kick receivers and whether a shot would follow.

The Palmeiras partnership, announced at Google’s Brasil event on June 10, marks a significant step. TacticAI was previously limited to set-piece analysis, specifically corner kicks. Palmeiras is the first club to use it for open-play dynamics. The club’s data science team uses a drag-and-drop interface to virtually reposition players and observe how changes affect the collective behaviour of both their team and the opponent.

That means a coach can ask: what happens if we push the left back five metres higher? TacticAI simulates the downstream effect on the entire defensive structure. It quantifies tactical options that were previously gut feeling. Google also partnered with Brazil’s football confederation CBF to use AI in World Cup preparation.

The underlying technology has applications well beyond sport. Predicting coordinated movement from visual data is the same problem autonomous robots, traffic systems, and logistics planners need to solve. TacticAI’s geometric deep learning approach, which treats players as nodes in a dynamic graph and models their spatial relationships, is architecturally closer to physical AI systems than to a standard language model.

Football has been slower than other sports to adopt AI-driven tactics. Baseball has Statcast. Basketball has Second Spectrum. Football’s continuous, 22-player dynamics make it harder to model. TacticAI’s 90% expert preference rate suggests the gap is closing, and the growing role of AI in professional sport is moving from analytics in the background to tactical recommendations on the sideline.

Google is continuing its push to make Gemini a central part of Android by giving the AI assistant deeper integration with Google Contacts. A newly discovered update suggests Gemini may soon handle contact-related tasks more directly, potentially turning it into a more capable personal assistant for calls, messaging, and everyday communication.

According to a report by 9to5Google, the latest Google app beta includes references showing that Gemini integration with Google Contacts is expanding beyond basic assistant functions. The feature appears designed to let Gemini interact more naturally with saved contacts, helping users quickly find people, initiate communication, and manage relationship-based tasks through conversational commands.

The update reflects Google’s broader strategy of positioning Gemini as more than just a chatbot. Instead of acting as a standalone AI tool, Gemini is increasingly becoming deeply embedded across Android services, apps, and system functions.

Google wants Gemini to feel more like a real phone assistant

The integration could significantly improve how users interact with Android devices. Rather than manually searching for contacts, opening apps, or navigating menus, users may simply ask Gemini to perform actions involving specific people stored in their contact list.

That could include tasks such as finding a person’s details, starting calls, sending messages, sharing information, or interacting with contacts through connected Google apps. Reports suggest the functionality may rely on Gemini’s understanding of relationships and conversational context rather than requiring rigid voice commands.

This matters because AI assistants have historically struggled to feel genuinely useful in day-to-day smartphone usage. While voice assistants can already place calls or send texts, they often require precise wording and still feel disconnected from broader workflows. Google appears to be trying to make Gemini more proactive and context-aware by giving it tighter access to personal data and Android services.

For users, the convenience could be substantial. A more intelligent contact-aware assistant may reduce friction when multitasking, driving, or quickly trying to communicate with people. Instead of opening multiple apps manually, users could rely on conversational requests handled directly by Gemini.

At the same time, deeper contact integration may raise familiar privacy concerns. Access to personal relationships, communication patterns, and contact data gives AI systems significantly more insight into users’ lives. While Google positions Gemini as a productivity-focused assistant, expanding its reach into sensitive personal information will likely invite scrutiny from privacy advocates.

Gemini’s role inside Android keeps expanding rapidly

The Google Contacts integration is part of a much larger shift happening across Android. Over the past year, Google has steadily replaced or reworked traditional Google Assistant features with Gemini-powered alternatives.

Gemini is already appearing inside Gmail, Drive, Docs, Photos, Search, and Messages, while Android itself is becoming increasingly centered around AI-powered experiences. The addition of deeper Contacts integration suggests Google wants Gemini to evolve into a true operating-system-level assistant capable of managing communication, organization, and productivity tasks seamlessly.

The feature has reportedly been spotted in development and has not officially launched yet, meaning details could still change before release. However, its discovery strongly suggests Google is preparing even tighter Gemini integration across core Android apps.

As AI assistants become more deeply woven into smartphones, the line between operating system and personal AI companion continues to blur – and Google appears determined to make Gemini the center of that experience.

SpaceX has lined up another compute deal ahead of its historic IPO, this time with Google. The company announced the deal in a regulatory filing on Friday.

Under the terms of the deal, Google will pay SpaceX $920 million per month from October 2026 through June 2029 for access to “approximately 110,000 NVIDIA GPUs, CPUs, memory, and other related components.”

The deal is similar in length and scope to the one SpaceX announced with Anthropic in late May. As part of that deal, Anthropic agreed to pay SpaceX $1.25 billion per month through 2029 to rent all the available compute from its Colossus 1 data center near Memphis, Tennessee, that xAI — now part of SpaceX — originally built for its own artificial intelligence efforts.

Google’s deal appears to be paying for roughly half the amount of compute that Anthropic has access to at Colossus 1. SpaceX didn’t say which specific data center Google would be using. CEO Elon Musk has previously suggested his company would reserve the Colossus 2 data center for xAI.

Anthropic was significantly limited in its compute capacity prior to its deal with SpaceX, raising usage limits on the same day the deal was announced. Google is in a very different position, with some estimates naming it as the world’s largest single owner of AI compute.

In a statement, a Google representative described the deal as a result of unexpected demand for its recently launched AI products. “Google Cloud and SpaceX are long-time partners,” Google said in a statement. “This is a short-term, timely agreement to ensure we have bridge capacity to meet surging customer demand for our agent platform, Gemini Enterprise, which has been even higher than we expected.”

But its parent company Alphabet is on a spending spree. Alphabet has already committed to more than $180 billion in capital expenditures this year and has said it expects that to “significantly increase” in 2027. To help with that, Alphabet recently announced an $80 billion equity sale.

Also like the Anthropic deal, the agreement with Google includes a cancellation clause. Both SpaceX and Google have the option to terminate the agreement with 90 days’ notice after December 31, 2026. Google’s access to the data center will ramp up “through September at a reduced fee,” according to the filing.

“If we fail to deliver access to the committed amount of GPUs by September 30, 2026, then following a one-month grace period, Google may immediately terminate the agreement or accept the number of GPUs provided” with a reduction in the monthly fees, it reads.

SpaceX announced the deal just one week before the company’s stock is expected to start trading on the Nasdaq exchange. Paperwork filed with the Securities and Exchange Commission shows the company is aiming to raise around $75 billion at a valuation of around $1.75 trillion — making it the largest in history.

Google is a longtime investor in SpaceX. Its stake in Musk’s company is expected to be worth more than $100 billion after the IPO. The companies are also reportedly in talks to try to build orbital data centers — a major component of SpaceX’s future plans post-IPO.

If you have ever sat through a Google Meet call and later struggled to figure out what was actually decided, Google just made your life a little easier. The company has rolled out a meaningful upgrade to its “Take notes for me” feature, giving you more control over what gets captured and making it easier to track outcomes after the meeting wraps up.

Meeting notes just got a serious upgrade.

You can now toggle specific sections (like Decisions or Next Steps) with Take notes for me in Meet. Plus, our new “Decisions” tracker helps you categorize outcomes like Aligned or Needs Further Discussion so nothing falls through the… pic.twitter.com/vuot3EBI97

— Google Workspace (@GoogleWorkspace) June 5, 2026

What’s new in “Take notes for me” feature?

The biggest addition here is the ability to toggle specific sections of your meeting notes on or off during the call itself. From the in-call menu, you can now choose which of the four sections you want included: Summary, Decisions, Next Steps, or Details. These toggles only apply to your current meeting and do not change your default settings for future calls.

The Summary section has also been tightened up. It is now designed to be more concise and scannable, so you can catch up on the key points of a meeting quickly without having to wade through a wall of text.

What is the new Decisions section and how can it help you?

The Decisions section is the most interesting part of this update. Rather than just listing what was discussed, it explicitly tracks the outcome of each decision with a status label. Those labels are Aligned, Needs Further Discussion, Disagreed, and Shelved.

So instead of scanning through a long notes doc to figure out where things landed, you get a clear snapshot of what was resolved and what still needs work. For now, the Decisions section is only available in English. We can expect more languages to be added in the future.

The update is available for Enterprise Standard and Plus, Business Standard and Plus, Frontline Plus, and Google AI Pro for Education subscribers, as well as Google AI Pro and Ultra subscribers.

Google announced on Tuesday that Android is launching fake call detection to protect against AI deepfake impersonation scams. The feature is rolling out globally in Phone by Google to Android 12+ devices this month, starting with Pixel devices.

As people increasingly refuse to answer calls from unknown numbers, scammers are shifting their tactics by spoofing trusted phone numbers and using AI deepfake technology to sound like authority figures, family members, or employers.

For example, a person may receive a phone call showing the caller ID “Mom,” and the voice may sound exactly like her, but the caller is actually a scammer using AI tools to impersonate her and request money for a fake emergency.

The new feature is on by default and works automatically behind the scenes.

Google explains that the new feature works kind of like a “digital handshake between devices.” When a contact calls you, and you’re both using Phone by Google, their phone sends a silent confirmation signal to your device to verify the call is legitimate and actually coming from their phone.

“If a scammer tries to impersonate your trusted contact, that initial confirmation signal will be missing,” Google explained in a blog post. “Your device will instantly notice this and ping your contact’s actual device to double-check. If their real device says, ‘I’m not making a call right now,’ you’ll get a warning on your screen advising you to hang up immediately.”

The tech giant notes that it built this feature on top of Rich Communication Services (RCS), making it possible for other apps and companies to adopt the technology.

The launch of fake call detection was announced alongside other updates from Android, including a new Google Photos feature that lets users mix and match outfits and try them on virtually. The new “wardrobe” feature catalogs the clothes you’re wearing in your photo library by turning them into snapshots you can browse on your phone. The feature is rolling out next week to eligible users in the U.S., India, and Brazil with Android 10+.

Additionally, Google Play Books is getting a new “Catch me up” feature that lets users jump back into a story with a recap. Users can also highlight a passage to ask questions. These features are rolling out today for select English titles.

Google is also making it possible to search entire outfits with its “Circle to Search” feature. Now the feature will be able to find every item in an outfit at once, getting rid of the need to search piece by piece. This update is now available on all Android 14+ devices that have Circle to Search.

TL;DR

Alphabet’s Verily is seeking EPA approval to release 64 million Wolbachia-carrying male mosquitoes in Florida and California over two years  to suppress Aedes aegypti populations. Earlier Debug trials achieved up to 95% reductions in biting mosquitoes in Fresno and 70%+ dengue reduction in Singapore.

Verily, the health and life sciences subsidiary of Google parent Alphabet, has asked the US Environmental Protection Agency for permission to release up to 64 million lab-bred mosquitoes (up to 32 million per year) across Florida and California over two years. The proposal, submitted under the company’s Debug initiative, would deploy male Aedes aegypti mosquitoes carrying Wolbachia bacteria into areas where the species transmits dengue fever, Zika virus, yellow fever, and chikungunya. The EPA is reviewing the application under an experimental use permit, with a public comment period running through early June.

The approach is elegant in its biology. Wolbachia is a naturally occurring bacterium found in roughly 60% of insect species, but not in Aedes aegypti. When lab-reared males carrying Wolbachia mate with wild females that lack it, the resulting eggs do not develop. The males do not bite, do not transmit disease, and cannot reproduce with each other, making the release self-limiting. Each generation of released males reduces the wild population without introducing permanent genetic changes or synthetic chemicals into the environment.

The results so far

Debug has been running field trials since 2017. In Fresno, California, the programme achieved reductions in biting female Aedes aegypti populations of 68% in its first year, over 95% in 2018, and 84% in 2019. The results put it among the most effective biotech-driven population control methods tested at scale in the United States.

In Singapore, where the government has deployed Wolbachia-based mosquito control as part of its national dengue strategy, releases of sterile males achieved 80% to 90% suppression of Aedes aegypti populations and a more than 70% reduction in dengue cases within six to 12 months. The World Mosquito Program, which operates independently of Verily, has deployed Wolbachia across 14 countries using a different approach that infects both male and female mosquitoes to establish self-sustaining Wolbachia populations in the wild.

How Verily scales it

The challenge with mosquito releases is not the biology but the logistics. Releasing millions of male mosquitoes requires breeding them at scale, sorting them from females with near-perfect accuracy (releasing females would increase biting), and distributing them across target areas at the right density and timing. Verily is applying its parent company’s AI capabilities to each of these steps.

The Debug programme uses AI-powered mosquito sorting systems that can separate males from females using computer vision, automated breeding technology that maintains colony health at industrial scale, and specialised release platforms, including drones and ground-based dispensers, that distribute mosquitoes across target zones. The total programme scope is 64 million mosquitoes, broken down to 16 million per state per year over the two-year permit period. Some outlets have reported the figure as 32 million, which appears to refer to the annual release across both states rather than the full two-year total.

The debate

Not everyone is comfortable with the idea of a tech company releasing tens of millions of insects into the wild. The intersection of AI, biotech, and public health raises questions about corporate involvement in ecological interventions, long-term environmental effects, and community consent. Critics have raised concerns about whether reducing one mosquito species could affect the food chain, whether Wolbachia could transfer to non-target species, and whether corporate-led biological interventions should require stricter oversight than the current EPA permit process provides.

Proponents counter that Aedes aegypti is an invasive species in the Americas, not part of the native ecosystem, and that its removal would have minimal ecological impact. The species is responsible for hundreds of millions of dengue infections annually worldwide, with cases rising sharply as climate change expands the mosquito’s range into previously temperate regions. AI-driven approaches to biological and medical challenges are producing results that traditional methods have struggled to match.

Why it matters for tech

Debug is one of the few surviving projects from Verily’s original mission to apply technology to life sciences at scale. Alphabet has restructured its health ambitions multiple times, shutting down or spinning off various initiatives. While most of Big Tech’s resources flow toward AI infrastructure, Verily’s mosquito programme represents a different kind of applied technology: using automation, computer vision, and logistics software to solve a public health problem that kills hundreds of thousands of people each year.

If the EPA approves the permit, Florida and California would become the largest US deployment of Wolbachia-based mosquito control to date. The decision is expected after the public comment period closes, with releases potentially beginning in late 2026 or early 2027.

Google Photos backups are getting a little less wasteful for people who use Takeout to keep their own copy of a photo library.

This is aimed at Google Photos users who don’t want Google to be the only place their photos live, especially anyone keeping a second archive on an external drive, NAS, or another cloud service. Google says Incremental Takeout for Photos will make recurring downloads faster and more efficient once the initial archive is complete.

The first run still includes all selected photos and albums, but later runs focus on photos and videos that were uploaded, backed up, created, or edited since the last successful backup. That’s a useful fix for anyone tired of downloading the same massive archive again and again, but the catch is built into setup, Photos has to be the only product selected in Takeout for the incremental option to appear.

Why does the first backup take everything

The first run is still the baseline, not a shortcut around the biggest download. Google says it includes all selected photos and albums, so longtime Photos users should expect the initial export to remain the heavy lift.

The payoff comes after that baseline is finished. Once Takeout has a successful backup to compare against, the next recurring export can skip unchanged files instead of packaging the whole library again. That should mean fewer duplicate downloads, less wasted drive space, and a backup process that feels less punishing over time.

Why does setup have one catch

Incremental Takeout only works when Photos is the lone product selected for export. Anyone who usually bundles photos with other Google data will need to create a separate recurring export just for the library.

That limit keeps the feature focused, but it also narrows who gets the full benefit. Takeout can make Photos backups less annoying, while broader Google account archives still need their own setup, schedule, and storage planning.

What should local backup users do next

The practical move is to set up a recurring Takeout export for Photos by itself, then treat the first download as the baseline archive. After that, each successful run should make the next one smaller by leaving unchanged items behind.

Google hasn’t provided a broader rollout date beyond the announcement, and exact regional availability isn’t stated. During setup, the key check is simple, the incremental option should appear only when Photos is the sole selected product.