Chris Lehane argues that stronger AI capabilities require stronger safety evidence, shared standards, and durable policy action while the policy window remains open.
Trust is a central obstacle to realizing the potential of AI agents at work. Employees are unlikely to give agents meaningful autonomy if they are uncertain about their reliability, intentions, or
Imagine this scenario: You’re a manager testing out a new AI agent “assistant” for your team, and as you set it up it asks for permissions. Three options appear:
Today, a single individual enabled by AI can—at very low cost—perform many activities needed to establish a startup quickly and in parallel: generating product ideas, simulating
Almost two decades ago, the entrepreneur and author Eric Ries launched a methodology for the “lean startup.” In this approach, a new business could be launched with just capital, a minimum viable product, a team, extensive discussions with potential customers, and a process of repeated pivoting as the startup scaled. The lean startup approach combined Toyota’s lean manufacturing method with Stanford professor Steve Blank’s approaches to customer development. Because founders typically had few resources at their disposal, the lean approach took hold.
How predicting the molecular impact of every possible single-letter DNA variant in the human genome will help accelerate our understanding of biology.
Today, we are introducing AlphaGenome Atlas: a platform (detailed in our paper) containing predictions for the effects of 9 billion single-nucleotide variants — every single-letter change possible — in the human genome. It is the most comprehensive catalogue of how genetic mutations affect molecular biology, and it is available for academic research through an intuitive and free-to-use website portal.
DNA is the language of life. Mastering it is a grand challenge that could transform our ability to understand biology and treat disease. But progress has been limited by a fundamental problem: interpreting how genetic variations impact biology at a molecular level. With roughly 9 billion possible single-letter mutations in the human genome, testing each one in the lab is practically impossible.
Google DeepMind has already made progress on this challenge with AlphaGenome, an artificial intelligence (AI) model that can predict how genetic variants impact biological processes. AlphaGenome is helpful for analyzing specific variants and has found widespread use in research, but we wanted to show researchers a big-picture view of variants across the entire genome.
By precomputing AlphaGenome’s predictions at scale, we have created an easily accessible resource that vastly expands the model’s reach. Just as an atlas is a collection of maps, linking together features of the land like altitude and location, AlphaGenome Atlas charts the molecular effects of DNA variants across the genome.
To help scientists quickly find the most impactful genetic changes, we are also releasing the AlphaGenome Variant Impact (AVI) score. The AVI combines the strengths of AlphaGenome and AlphaMissense — our model for predicting the impact of protein-altering DNA variants — condensing both models’ predictions into a single number. Now, researchers can rapidly rank variants and interpret their molecular effects at the same time.
Our trusted external collaborators have already used AlphaGenome Atlas to identify and experimentally verify key variants in unsolved rare disease research and find rare variants associated with common traits.
AlphaGenome Atlas is a massive 1-petabyte dataset, more than 30 times larger than the AlphaFold Database. When we expanded the AlphaFold Database in 2022, we grew the 3D structure information available from around 190K experimental structures to more than 200M structure predictions — covering nearly all catalogued proteins known to science. The database provided a portal that researchers with no coding experience could use, providing intuitive visualizations and making it easier to do large-scale protein structure analysis. It quickly became a crucial resource that drove discoveries across the life sciences and continues to accelerate researchers’ important work in countless fields.
In building AlphaGenome Atlas, we also aspire to make predictions more accessible and give scientists an intuitive way to explore a vast dataset.
AlphaGenome Atlas provides several powerful, interconnected resources, allowing researchers to link variants directly to the functional DNA sequences they disrupt.
Molecular effect predictions Atlas contains thousands of molecular effect predictions for each variant, across multiple important aspects of gene regulation, spanning hundreds of human and mouse cell types and tissues. This serves as the starting point for further resources.
AVI score A single number describing the impact for each genetic variant.
AVI feature attributions Each AVI score is also linked to distinct biological features driving it, such as the aspects of gene regulation predicted by AlphaGenome or the protein impact score from AlphaMissense.
DNA sequence motifs A comprehensive collection of over 2,500 recurrent DNA sequences — the “words” of the genome — and their locations.
Together, these resources support researchers for a wide range of genetic research tasks, from rapid variant ranking to deep dives into variant functions.
Extensive community collaboration guided the design of AlphaGenome Atlas. The AVI score helps researchers rapidly score and rank variants based on their potential impact. Crucially, it works for both coding regions (the 2% of the genome that codes for proteins) and non-coding regions (the remaining 98%), which orchestrates gene activity and houses most trait-associated variants.
Our testing shows that the AVI score provides best-in-class performance across many variant pathogenicity and rare disease benchmarks. To help interpret these scores, we also calculated AVI feature attributions that highlight which molecular processes — like RNA splicing or gene expression — are predicted to be most disrupted by each variant.
Overview of the AlphaGenome Atlas.(1) Precomputed effects are generated genome-wide for over 9 billion single-nucleotide variants. (2) From this, an allelic-resolution AlphaGenome Variant Impact (AVI) score is derived for each variant. To facilitate variant interpretation, AlphaGenome Atlas then decomposes the AVI score into additive feature contributions across interpretable categories such as chromatin accessibility, splicing, and conservation. (3) The precomputed variant effects, AVI score and the AVI feature attributions are linked, together with a compendium of genome-wide de novo motifs, which enables high-resolution mechanistic insights into variant function.
Real-world impact: From rare diseases to population genetics and molecular biology
AlphaGenome Atlas provides a high-resolution, global view of the genome. These large-scale predictions become most useful when applied to targeted research questions. By translating this data into actionable biological insights, our academic partners are already uncovering links between genetic variation and disease.
Understanding unsolved rare diseases. A major hurdle in understanding rare diseases is the daunting task of pinpointing the few causal variants hidden among thousands of candidates. In collaboration with the GREGoR Consortium, researchers applied the AVI score to prioritize these needle-in-a-haystack genetic variants for unsolved rare disease research. When Laura Covill and Anne O’Donnell-Luria from the Broad Institute and their colleagues used the AVI score to prioritize variants, driving a rare disease, that were overlooked in previous research, the team discovered a variant affecting a gene called DNM1, which is strongly linked to epileptic encephalopathy.
Crucially, the AlphaGenome predictions underlying the AVI score showed exactly how the variant functioned: it created an incorrect splice site (a mistake in the cell’s genetic instructions) that led to an abnormal extension of the resulting protein. Experimental screens validated the research prediction and found nearby variants with similar effects, showing that Atlas is a powerful tool for understanding impactful genomic variation.
Mapping rare variants associated with protein levels and complex traits. Moving beyond individual rare disease research, AlphaGenome Atlas can help uncover the genetic architecture of common traits in the general population. Identifying which rare, non-coding variants are associated with a specific trait or disease is notoriously difficult because the sheer volume of harmless genetic changes creates a statistical ‘background noise’.
To test how AlphaGenome Atlas can improve our ability to find non-coding variants affecting human traits, Gareth Hawkes, a Medical Research Council fellow at the University of Exeter, applied AlphaGenome Atlas to whole-genome data from over 54,000 UK Biobank participants, which made these elusive signals more obvious. By grouping rare variants based on their predicted molecular effects, Hawkes uncovered 22% more non-coding genetic associations, which would otherwise have not been detectable in the statistical noise. This let Hawkes pinpoint specific regulatory variants driving the abundance of critical proteins circulating in the human body, including PLA2G7 (linked to aging) and EGLN1 (a vital cellular oxygen sensor).
Taking this approach even further, Hawkes used AlphaGenome Atlas to look at how hundreds of millions of non-coding variants in the UK Biobank might be linked to body mass index. By focusing on the 1% of non-coding variants which Atlas predicts to be most impactful, he identified 19 genetic regions, which could help direct the next stage of targeted research into this trait.
Identifying the regulatory ‘words’ of the genome. Atlas can also be used to identify which recurring short sequences, or motifs, are driving different molecular processes in different cell types for different genes. These motifs can provide key clues, such as locating binding sites of transcription factors (proteins that turn genes on or off) and providing additional interpretation of non-coding variants. Julia Zeitlinger and Melanie Weilert at the Stowers Institute for Medical Research used this resource, for example, to categorize which transcription factors only affect the accessibility of DNA versus which ones are also able to turn genes on and off.
Accelerating genomic discovery
With AlphaGenome Atlas we are creating new layers of information that will help further our understanding of the human genetic code. We hope that this will be a valuable resource for scientists, but we also view it as a baseline rather than an endpoint. As our AI models like AlphaGenome improve, our maps of the entire human genome will become increasingly comprehensive and precise.
AlphaGenome Atlas is powerful in isolation, but it also represents a step towards our vision of a broad, unified solution for biologists. Its resources can be integrated into our broader agentic systems, like Google Antigravity, to help enhance end-to-end scientific workflows.
It is also important that AlphaGenome Atlas’ scientific knowledge is widely available, so we have made it accessible for non-commercial use through our website from today, as well as for commercial use on Google Cloud soon. (The AlphaGenome base model is already available for academic use on GitHub and via the AlphaGenome API, and also is available for commercial use on Cloud via Model Garden).
Together, these tools will enable researchers and industry partners to accelerate the pace of biological discovery: finding novel therapeutic targets, better understanding genetic disorders, and driving the next wave of targeted experimental validation.
Acknowledgements
We are grateful to our research collaborators at University of Exeter, Broad Institute, Boston Children’s Hospital, Stowers Institute for Medical Research, Harvard University, Memorial Sloan Kettering Cancer Center, Center for Genomic Medicine at Massachusetts General Hospital, and the University of Kansas Medical Center.
This work was done thanks to the contributions of Jun Cheng, Kyle R. Taylor, Lauren Nicolaisen, Joshua Pan, Clare Bycroft, Matteo Perino, Tom Ward, Raina W. Thomas, Natasha Latysheva, Gareth Hawkes, Laura E. Covill, Melanie Weilert, Maile J. Hirschmann, Xi Dawn Chen, Robin N Beaumont, V Kartik Chundru, Michael N Weedon, Simon Bourdareau, Hoyin Chu, Dhavi Hariharan, Thais Kagohara, Lucas Tenório, Yosuke Ushigome, Amanda Stafford, Courtney A. Shearer, Barbara Ikica, Ada Fang, Mouad Naciri, Victoria Johnston, Richard Green, Elisa Lai Hong Wong, Vincent Dutordoir, Anne Mottram, Adam Gayoso, Eirini Arvaniti, Guido Novati, Heidi L. Rehm, Fei Chen, Caleb A. Lareau, Caroline F Wright, Anne O’Donnell-Luria, Julia Zeitlinger, Pushmeet Kohli, Žiga Avsec.
We wish to thank our teammates and collaborators for their technical support, feedback, including Kathryn Tunyasuvunakool, Alexander Karollus, Risha Patel, Francesca Pietra, Alisha Eastep, Doga Fadillioglu, Charlie Taylor, Raphael Aboyeji, Uchechi Okereke, Gemma Gibbs, Olufemi Duduyemi, Juan Mateos-Garcia, Mariana Felix, Sahar Abdulrahman, Antonia Mould, Rachael Tremlett, Chang Yun, Salil Deshpande, Anshul Kundaje, Samantha Bryen, Greg Findlay, Phoebe Dace, Kinga Bujakowska, Emma Sherrill, Aubrie Soucy Verran, Boxun Zhao, Tim Yu, Francesca Pietra, Brendah Namugamba, Cassie Gray, Daniel MacArthur, Lingyi Wang, Marc Mansour, Mohamad Hajjari, Mounica Vallurupalli, Philip Montgomery, Phoebe Dace, Roisin Sullivan, Sam Bryen, Teresa Niccoli,
Finally we wish to thank Evie Gray, Adriana Fernandez Lara, Alex Wilkins, Danielle Breen, Mariana Montes, Inês Ayer, Ryan Smith, Ross West and Gaby Pearl for their expertise in communicating this work.
AlphaGenome
AlphaGenome: AI for better understanding the genome
In the September 7 edition of The Insider newsletter, managing editor Gretchen Gavett shares HBR’s plans for covering AI over the next month, new insights from the creators of blue ocean
For the better part of two decades, the story of industrial disruption had a familiar shape. Nimble entrants ran circles around the incumbents, unburdened by legacy portfolios, tangled supply chains, or installed bases measured in decades. The advantage belonged to whoever carried the least.
AI is inverting the story. The same weight that slowed large manufacturers—the sprawling product histories, the regulatory scar tissue, the decades of hard-won engineering intellectual property (IP)—is precisely the raw material that makes AI valuable.
In discrete manufacturing, the winners of the AI decade will not be the startups with the cleanest slate. They will be the Goliaths who finally turn their bulk into leverage.
The Asset No One Can Buy
Here is the uncomfortable truth for every well-funded newcomer: The one thing they cannot acquire is history. A general AI model can learn everything on the public internet and still know nothing about how your product is designed, built, changed, and certified. That knowledge lives in the engineering systems established manufacturers have run for years—the requirements, designs, bills of materials, and change history that constitute the essential IP of a business, along with the rules governing who may see it and how it gets approved.
That is exactly the context AI needs to reason well: what it is working with, who is allowed to use it, and how similar problems were solved before. Strip that away and even a capable model is reduced to guessing.
Owning It Isn’t the Same Thing as Wielding It
There is no partial credit here: Either a manufacturer’s product data is connected, governed, and curated or it isn’t. And for most, it isn’t. When a manufacturer’s product data sits fragmented across systems that don’t talk, modeled inconsistently across functions, locked in PDFs and the memories of engineers who may be five years from retirement, that is not a company with a slightly weaker AI position. It is a company with no AI position at all, pointing retrieval-augmented generation at a pile of disconnected documents and calling the output intelligence.
Take a routine scenario: A supplier discontinues a component and a replacement has to be qualified.
In one company, an engineer spends weeks tracing where that part lives: which assemblies, which configurations, which customer variants, which open orders, and which regulatory filings reference it. In another, an AI agent reaching the live product record through an open standard like the Model Context Protocol returns the full ripple in minutes, flags the two variants where the substitution breaks a compliance requirement, and surfaces a part already qualified for the same purpose.
Same change, same models, same ambition. But one company absorbs the cost quietly; the other never pays it.
And that is one scenario among dozens that play out every week, such as:
A field failure comes in from a customer. One company reasons across service records, the as-built configuration, and the change history to find the root cause in an afternoon; the other convenes a task force and waits.
A new program goes out to bid. One company knows in hours which existing designs, parts, and lessons it can reuse; the other quotes as if starting from scratch and either overbids or wins work it will lose money on.
An auditor asks for the complete design and approval trail on a shipped product. One company produces it on demand; the other sends people digging through inboxes and shared drives for a week.
A design engineer starts a new part. One company’s AI surfaces the three interchangeable parts already in the catalog; the other creates a fourth and pays to carry it for its entire lifecycle.
None of these is a dramatic, board-level crisis. Each is a small, ordinary moment of product work. And that is the point. Here, the math tips toward the giant: the larger the portfolio, the deeper the installed base; the longer the history, the more of these moments occur every day and the more a connected foundation compounds across them. The same scale that was a liability for 20 years becomes the multiplier. On disconnected data, that scale only multiplies the drag. Those returns don’t come on a sliding scale. They belong entirely to the manufacturers whose data is ready to be reasoned over, and not at all to the ones whose data isn’t—a divide decided long before the first model is deployed.
A Strategic Decision, Not a Software Purchase
Closing the gap is less about buying new software than about making a decision: to own and govern product data as an enterprise asset, not a by-product of engineering. Three things unlock it:
A product data foundation, anchored in the bill of materials, so AI reasons over an authoritative connected model rather than fragments that disagree.
Governed AI, where the same access rules that protect the IP constrain the model to limit what each user may see, keeping the model secure and auditable rather than a novel way to leak secrets.
AI in the workflow, running within the procedures teams already use, from engineering change to configuration to manufacturing planning, rather than hovering above them as a chatbot.
The executives who win this decade won’t be those who ran the most AI experiments. They’ll be those who asked the harder question early: Are we turning decades of engineering IP into a structural advantage or letting it depreciate in systems that don’t connect, align, or scale?
For 20 years, size was a liability. AI is making size the ultimate advantage—but only for the manufacturers who build the foundation to wield it.
The advantage is already theirs. The only question is whether Goliath picks up the stone—before David convinces him he’s too big to fight.
For years, corporate lawyers have obsessed over one question when it comes to AI: Is our data going into someone else’s training set? Most companies feel they’ve solved that. They’ve signed contracts saying the AI provider can’t train on the prompts their employees type in.
