Google DeepMind

When AI-generated content has become so realistic, how do we verify what's made by human, machine, or nature? We explore the vital role of provenance – and how imperceptible watermarking can safeguard both digital media and biological designs – on our podcast. → https://goo.gle/4detQwj

SynthID Bio is our new family of watermarking methods made for AI-generated biological designs. 🧬 In a world first, we can now embed an imperceptible signature directly into protein sequences without affecting their biological function. Acting like a digital ID card, the watermark can help DNA labs verify synthetic proteins and keep sequence databases accurate. As these requests increasingly scale, verifying if designs come from AI with built-in safeguards will help labs automate and strengthen biosecurity. As AI accelerates scientific discovery, we’re helping researchers keep pace by quickly authenticating if biological designs are created with built-in safeguards, without stalling crucial work. We’re releasing these tools on an open basis for research use to strengthen safeguards → https://goo.gle/4yBysVC

Today we’re introducing Gemini 4 Argon, our next era of frontier intelligence. It delivers frontier performance in complex workflows across real-world software engineering, knowledge work, and cybersecurity defense with an industry-leading 1M token output limit. Gemini 4 Argon is rolling out to an initial cohort of cyber defenders through our Fairwind Program so they can leverage its full frontier-level cybersecurity defense capabilities. We'll continue to gather feedback from early testers as we iterate on guardrails before making Argon available to developers, enterprises, and consumers as soon as possible. Learn more → goo.gle/4AZLPRt

We’re hiring — right now — for Research Engineers to join Polaris, a new team at Google DeepMind. As synthetic data pipelines hit their ceiling, the hardest engineering problem in AI is translating the full depth, nuance, and complexity of real-world software engineering into rigorous evaluation frameworks and training benchmarks. That’s why we’re building Polaris. Rather than generating a large number of tasks using a synthetic data pipeline, engineers on the Polaris team focus on creating a smaller number of tasks that more closely capture the complexity of real-world software engineering — taking roughly a week to complete, from ideation to final approval. Alongside building these benchmarks, you will have the autonomy to launch entirely new projects within DeepMind aimed at advancing Gemini’s training data and evaluations. You’ll create experiments, prototype implementations, design new architectures, and tackle real-world problems across AI, NLP, compilers, search, and hardware/software performance analysis all while staying connected to the wider research community through university partnerships and publishing papers. Great talent comes from anywhere. We're hiring across the entire spectrum of experience to find the best people in the world. We want to hear from you if you are: • A Competitor or Hacker: You thrive in competitive programming, math olympiads (IMO, IOI, Putnam, USAMO), or hackathons, and love constructing deeply challenging problems. • An AI-Native Builder: You already live in AI coding agents and LLM-based workflows to ship software at high velocity. • A Polyglot Engineer: You can parachute into unfamiliar programming paradigms, tools, and technical stacks and master them rapidly. If you want to build with a team at Google DeepMind, or know someone who belongs in this room, join us. Apply now → https://goo.gle/3V5Kacp

From robotics to biomedicine and data engineering, there are many directions a postgraduate degree in AI could take you. We've partnered with University of Cambridge, The University of Edinburgh, University of Oxford and UCL to offer masters and PhD scholarships in a range of AI and Computer Science disciplines. These partners are home to globally renowned departments where researchers are working at the cutting edge of computer science research. You could be joining a world-leading department next September, pursuing AI research alongside current and future AI pioneers. Find the course that is right for you and apply now: https://lnkd.in/egc6gDDm

I’m incredibly honoured to share that I have been elected a Fellow of the Royal Academy of Engineering - an accolade I’d dreamed of, yet never fully dared to imagine receiving. To have this happen during the Academy’s 50th-anniversary year makes it feel even more surreal. When the news came through, my mind went to the first engineer I ever knew - my late father, Shawki Ibrahim. He was orphaned in Lebanon, and then immigrated to the US with very little, aside from the dream of becoming an engineer. A core childhood memory of mine is him sitting at his drawing table late into the evening, completely absorbed, sketching away with a set of colored pencils. As a kid, I just thought he was making art. It took me a long time to understand that those drawings were patent-worthy designs for heart pacemakers, and critical safety devices for cars & airplanes. Engineering gave him both social & economic mobility, while having a pioneering impact in the world. It gave us our foundation as a family, and it made it feel entirely natural for both my brother and me to follow his footsteps into the semiconductor world.  Going through the nomination forced me to sit down and map out my career all the way back to my first engineering job in 1990! Seeing 35+ years laid out in front of you is a strange exercise. You realize very quickly how little of it was a master plan. The honest truth is that much of my path was built on raising my hand for the unknown, unclear, tough jobs nobody else wanted - simply because they seemed interesting to me (& maybe because I’ve never been very good at walking away from a problem that needs to be solved)! None of it felt like "pioneering" at the time. It was just a curiosity to solve hard problems to benefit the world, and alongside people who were generous enough to mentor and sponsor me. From Purdue, to Intel where I dove into Pentium processor and computer architecture, wrangled global DVD standards, spun up (multiple!) developer programs, and built computers for classrooms - to my VC time at Kleiner Perkins, co-founding the non-profit Team4Tech, leading engineering teams through Coursera and DeepMind, and being at the frontier of AI for the past 8+ years. It was always just about trying to build responsibly & make technology useful for real people. I carry my dad with me every single day, & today I feel his presence especially close. This Fellowship means the world to me, and at its heart, it’s a tribute to him at that drawing table. It’s a reminder of what engineering can do for a human life, and a commitment to what comes next. As a Fellow, I’m committed to continue bridging the gap between AI breakthroughs and real-world societal impact, and helping the next generation of curious builders find their own way in the world. https://lnkd.in/ef_6tMpB

Is the world prepared for a future pandemic? COVID-19 exposed how vulnerable we are to infectious disease outbreaks. According to an analysis by the Center for Global Development, it's unlikely to be the last pandemic we face during our lifetime. To support our biological understanding of viruses, AI-predicted protein complex structures for 2,800 viruses are now openly available in the #AlphaFold Database. The dataset prioritises proteomes from viral families known to infect humans. This work was made possible thanks to a collaboration between EMBL-EBI, Google DeepMind, NVIDIA, Seoul National University, MRC-University of Glasgow Centre for Virus Research (CVR), SIB Swiss Institute of Bioinformatics, CEPI (Coalition for Epidemic Preparedness Innovations), and Sungkyunkwan University. This dataset release coincides with the United Nations General Assembly High-level meeting on Pandemic Prevention, Preparedness and Response, which takes place on 25 September at the UN headquarters in New York. The meeting will review the progress made towards commitments made in 2023 to strengthen global pandemic preparedness. Read the full story here: https://lnkd.in/gdwQHDZJ Explore the data in the AlphaFold Database: https://lnkd.in/dzdzkeX

Really exciting milestone today: we’re expanding Private AI Compute with secure, server-side memory, setting the stage for AI that remembers your context across devices. The linked blog post discusses how our updated architecture keeps data secure: combining hardware-isolated enclaves with device-derived encryption keys so cloud memory remains truly private. This takes us beyond temporary chats toward continuous, cross-device AI that works over time – while ensuring your data stays inaccessible even to Google. I’m incredibly proud of the teams across Google DeepMind, Core, Platforms & Devices, and Cloud for driving this important work. Bridging the gap between persistent AI utility and verifiable privacy is essential if we're going to build this tech responsibly. Read more about the update in our blog, which links out to our updated technical brief and  third-party security audit: https://lnkd.in/gQG8XEYF

Create and deploy custom audio with our new text-to-speech models: 🔵 Gemini 3.8 Flash TTS: Design unique voices with distinct accents and characteristics. 🔵 Gemini 3.8 Flash-Lite TTS: Built for efficiency and scale, choose from your created styles or our expansive production-ready library. Fine-tune the delivery line by line, shaping pacing, emotion, and cues like laughs or pauses. All generated audio is watermarked with SynthID so it can be reliably identified as AI-generated. Start building with the Gemini API via Google AI Studio. Find out more → https://goo.gle/4hEiOmt

Over 3 billion people are still offline, and advanced AI usually requires expensive hardware and fast internet. The digital inclusion means meeting people where they are. That’s why I am very proud of my team’s work with Viamo . Through their "Ask Viamo Anything" platform powered by Gemini, people in low-resource communities across Rwanda and Nigeria have received answers to over 2 million questions—using basic feature phones over standard phone lines. Hearing how farmers like Mukakalisa use voice-based AI in their local language to access information and make farming decisions shows what's possible when the ecosystem—telcos, model builders, and governments—collaborates. Google has joined a new Gates Foundation-backed global coalition dedicated to closing the AI language gap for 3 billion people over the next 5 years. Watch Mukakalisa’s story: https://lnkd.in/ez7xcEYD