Google DeepMind

We’re working with the UK Department for Science, Innovation and Technology, the Ministry of Housing, Communities and Local Government and the Incubator for Artificial Intelligence on a new AI housing application planning prototype. 🏡 By cutting down the time spent on repetitive tasks, it could help planning officers focus their attention on complex projects and reduce processing times by up to 50%. → https://goo.gle/4xAl1Fm

How can we develop AI for the public good, and overcome its risks and complexities? By ensuring the people who undertake AI research reflect and represent the society it serves. The Google DeepMind Research Ready programme was created to encourage students from less advantaged socio-economic backgrounds into AI research. The programme provides fully funded research placements at a range of top UK universities. In summer 2025, 132 current and recent students took part to build their AI research knowledge, skills, networks and employability. Over 85% of participating students said they gained valuable, hands-on and practical experience of AI research methods and careers. To ensure we can maintain the success of the scheme we commissioned an evaluation of it by ImpactEd Group. With suggested improvements ranging from level of industry involvement to enhanced community building activities, see the full set of recommendations and the Academy's response by visiting: https://lnkd.in/etayYEB5 We now have 60 participants placed at 11 host universities for 2026. Upon completion of their placement, they will get to attend our celebration event and share their experiences in September. This would not have been possible without our funding partners Google DeepMind and The Hg Foundation, with support from the Department for Science, Innovation and Technology.

Robotics is one of the most exciting frontiers of AI, and we’re helping shape its future. Our Accelerator brings together 15 AI-driven startups across Europe who are creating the future of physical AI. This three-month program connects them with access to our AI stack, and our breakthrough Gemini Robotics models - underpinned by hands-on support from our teams. Meet the startups: 🤖 3D-Components AS Acumino Adapta Robotics Automated Architecture (AUAR) Bubble Robotics Danu Robotics Deltia Embodied AI Extend Robotics FORGIS Generative Bionics  Qualia  Robeauté Staer Touchlab Limited We look forward to mentoring and helping these companies scale their work responsibly. Find out more → https://goo.gle/4a9k25h

We’re teaming up with S.E. 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. ⚽ To map the pitch, TacticAI uses graph neural networks - treating all 22 players as individual nodes and their physical interactions as connections. This allows the club's data science department to virtually drag and drop players to test different defensive setups in real time.
 Why does this matter beyond sports? A live match is a masterclass in partial observability and multimodal data. Solving these continuous spatial problems could translate to leaps in robotics, computer games, and more. Find out more about TacticAI → goo.gle/4uv87Gc

We’re entering a new era of AI where millions of agents will soon interact, negotiate, and transact online. 🌐 To help understand how these systems behave safely and predictably, we’ve partnered with Schmidt Sciences, the Cooperative AI Foundation, Advanced Research + Invention Agency (ARIA) and supported by Google.Org to launch a new $10M technical research fund. By empowering a global network of independent researchers, we aim to address the complex, system-wide safety risks that arise when multiple models interact. → https://goo.gle/4e1kxAE

In Sierra Leone, a surging student population is outpacing available teachers. Our latest research explores how AI can act as a partner to support educators in these environments –  amplifying their reach without replacing their essential expertise and skills. We evaluated AI’s impact by looking beyond test scores to behavioral shifts. Over eight weeks, results suggest students were using AI to understand concepts, not just find answers – with Gemini queries about how to tackle problems rising from 68% to 90%. Find out more → https://goo.gle/43nM2hF

DiffusionGemma is our new experimental open model with up to 4x faster output on dedicated GPUs. Instead of predicting word-by-word, it generates entire blocks of text simultaneously. This lets the model self-correct and format complex markdown in real time. Find out more → https://goo.gle/4v4lVbL  

Say hello, hola, 你好 to Gemini 3.5 Live Translate: our latest audio model built for fast, cross-language communication. 🌐 It can convert speech into over 70 languages and processes it as it’s streamed - while keeping tone, pace and pitch intact - allowing for more natural conversations. Try it in the Google Translate app or build in preview via the API in Google AI Studio. Find out more → https://goo.gle/3QaHfgk

Co-Scientist is our Gemini-based multi-agent system able to generate, debate and evolve novel hypotheses for complex scientific problems. It can produce thousands of research directions. To find the most useful ones, the system holds debates to refine and rank ideas. In turn, it can also verify claims backed against technical literature, and use web search to bring in new knowledge. We’ve collaborated with global scientific experts to evaluate Co-Scientist on a range of complex open problems, with the system already assisting by: 🔘 Identifying new targets for liver fibrosis 🔘 Uncovering fresh approaches to tackling Amyotrophic lateral sclerosis (ALS) 🔘 Digesting decades of literature to propose novel genetic leads for reversing aging Co-Scientist is now available to individual researchers through Hypothesis Generation, as part of Gemini for Science: a collection of experiments and tools exploring the future of AI-powered scientific discovery. Find out more → https://goo.gle/4uLmtmv