Artificial intelligence (AI) has the power to revolutionise the way we approach scientific research, offering new opportunities for discovery and innovation. By leveraging AI, scientists could access tools that may help unlock some of our most complex and persistent scientific problems.
This is where Science Digital comes in. This program brings together digital science experts from CSIRO's Data61, National Collections,, Agriculture & Food, Health & Biosecurity, Environment, Manufacturing, and Early Career Research (CERC) Fellowship program.
Together we are building a platform and community of practice that aims to become the one-stop-shop for our scientists wanting to harness data and AI to unlock new discoveries in their field.
Let's take a closer look at how it all comes together.
The Challenge: meeting our AI potential
Narrow AI, trained for specific tasks, has been used by scientists for years to great effect. It's one of the most prevalent forms of AI and performs targeted functions with high proficiency. For example, pattern identification, language translation, and image recognition. Google's AlphaFold AI model has already had such a profound impact that its developers have been awarded the Nobel Prize in Chemistry for it.
Our scientists have used narrow AI to make fishing more sustainable, monitor livestock health, estimate the yield potential of wine grapes, and create an award-winning automated chest x-ray report generator. These achievements highlight how leveraging AI for complex scientific challenges can help us solve intricate challenges faster.
Recent improvements and increased access to generative AI technology have the potential to enable researchers to use AI in entirely new ways. However, there is currently no software as a service that provides scientists with state-of-the-art AI tools, empowering them to leverage AI at scale without needing to be experts or engineers.
Science Digital is changing that.
The solution: AI for scientists, by scientists
As a world leading science agency, CSIRO is uniquely positioned to design a generative AI tool that supports and empowers scientists.
To seize this opportunity, Science Digital will build a multi-agent AI platform capable of accelerating the scientific research and discovery process. It will be built on a deep understanding of the challenges faced by researchers in various fields.
To build and deploy this tool as a service to our scientists, Science Digital has gathered a multidisciplinary team of researchers, engineers, designers, product specialists, and responsible AI experts from across the organisation.
The team is engaging in a human-centered co-design and development process to ensure the platform is accessible and valuable for users. By focusing on meaningful and quantifiable validation, Science Digital aims to create a tool that scientists can trust to be responsible, accurate, transparent, and effective in delivering value to their work.
Making it real: Use cases
To ensure the platform delivers meaningful and measurable value, the Science Digital team is:
- clarifying the problems users want to solve
- assessing how AI can most effectively help
- building a system that integrates effortlessly into practical workflows.
Experts from across CSIRO will collaborate with Science Digital to develop a platform prototype and apply it to several high-impact scientific studies. This process helps to validate and enhance the platform’s functionality, capabilities, and results.
Genome annotation
Although strides have been made in genome sequencing and assembly, functional gene annotation discovery and inference remains particularly time-consuming and complex for scientists (e.g. biologists), hindering progress of gene function discovery.
This challenge is especially pertinent in non-model species where very little data exists and experimental design/inference is non-standardised/not democratised, as well as for gene families with no or very little experimental characterisation.
To overcome this, a prototype of the platform will be applied to accelerate high resolution functional annotation or gene discovery.
It will use data and computational methods to accurately predict gene and protein function. This use case will train the platform to be useful for researchers in materials science, life sciences, and environmental sciences. It also addresses our sustainability and healthy future challenges.
This case study provides a real-world example of the transformative potential that Dr Stefan Harrer (Science Digital Director), Dr Rahul Rane (Senior Research Scientist), and Dr Robert Speight (Principal Research Scientist) explore in their Lancet Commentary.
'Generative AI agents are transforming biology research: high resolution functional genome annotation for multiscale understanding of life' explores the game-changing integration of AI into biology. By harnessing generative AI and advanced agentic tools, biologists can better interpret genomes, predict biological functions, and design novel systems.
The paper emphasises the potential of these technologies to revolutionise drug discovery, crop development, and sustainable biomanufacturing.
Digitisation of national collection and autonomous labs
Biological collections provide a unique record of species diversity, distribution and biological interactions across space and time.
These collections are vast warehouses of novel genes, molecules and biochemical processes that can be explored to identify the next generation of foods, materials and medicines needed to meet the demands of a rapidly growing human population.
However, the labour-intensive process of taxonomic discovery and documentation remains essentially unchanged and at current rates, will take hundreds of years to unpack the rich and potentially 'game-changing' biological information within the world's collections – information that is sorely needed right now.
This use case aims to automate the digitisation workflow of National Collection's insect collection.
It will do this with AI-powered robotics technology, imaging, data extraction, and data management and analysis in physical environments.
Science Digital aims for this use case to empower researchers to leverage the platform as a tool for designing and executing autonomous lab workflows through assistive AI-driven robots and other physical systems.
This shift from digitisation to autonomous labs will refine robotic accuracy in the lower-risk environment of the national collection, paving the way for operations in more complex labs, domains and settings.
This vision of AI-driven biology was the focus of a paper published by Dr Stefan Harrer, Director of Science Digital, Dr Rahul Rane, CSIRO Senior Research Consultant in Applied Genomics, and Dr Robert Speight, Director of CSIRO's Advanced Engineering Biology Future Science Platform.
'Generative AI agents are transforming biology research: high resolution functional genome annotation for multiscale understanding of life' explores the game-changing integration of AI into biology.
By harnessing generative AI and advanced agentic tools, biologists can better interpret genomes, predict biological functions, and design novel systems. The paper emphasises the potential of these technologies to revolutionise drug discovery, crop development, and sustainable biomanufacturing.
Our Google partnership: a collaboration that enables transformation
To inform, inspire, and catalyse the use of AI in the Australian science community.
As a part of our five-year partnership with Google, Science Digital has entered a partnership with Google that utilises Google's capability to accelerate and transform Australian innovation.
Through the collaboration, Google and Science Digital will develop and showcase practical AI solutions to real-world challenges, including Science Digital's use cases Genomic annotation, and Digitisation of National Collections and Autonomous Labs.
As Science Digital's inaugural strategic partner, Google will play a critical role in raising awareness of Science Digital's initiatives globally, offer technical and ecosystem resources, and encourage scientists to harness the capabilities of AI.
Artificial intelligence (AI) has the power to revolutionise the way we approach scientific research, offering new opportunities for discovery and innovation. By leveraging AI, scientists could access tools that may help unlock some of our most complex and persistent scientific problems.
This is where Science Digital comes in. This program brings together digital science experts from CSIRO's Data61, National Collections,, Agriculture & Food, Health & Biosecurity, Environment, Manufacturing, and Early Career Research (CERC) Fellowship program.
Together we are building a platform and community of practice that aims to become the one-stop-shop for our scientists wanting to harness data and AI to unlock new discoveries in their field.
Let's take a closer look at how it all comes together.
The Challenge: meeting our AI potential
Narrow AI, trained for specific tasks, has been used by scientists for years to great effect. It's one of the most prevalent forms of AI and performs targeted functions with high proficiency. For example, pattern identification, language translation, and image recognition. Google's AlphaFold AI model has already had such a profound impact that its developers have been awarded the Nobel Prize in Chemistry for it.
Our scientists have used narrow AI to make fishing more sustainable, monitor livestock health, estimate the yield potential of wine grapes, and create an award-winning automated chest x-ray report generator. These achievements highlight how leveraging AI for complex scientific challenges can help us solve intricate challenges faster.
Recent improvements and increased access to generative AI technology have the potential to enable researchers to use AI in entirely new ways. However, there is currently no software as a service that provides scientists with state-of-the-art AI tools, empowering them to leverage AI at scale without needing to be experts or engineers.
Science Digital is changing that.
The solution: AI for scientists, by scientists
As a world leading science agency, CSIRO is uniquely positioned to design a generative AI tool that supports and empowers scientists.
To seize this opportunity, Science Digital will build a multi-agent AI platform capable of accelerating the scientific research and discovery process. It will be built on a deep understanding of the challenges faced by researchers in various fields.
To build and deploy this tool as a service to our scientists, Science Digital has gathered a multidisciplinary team of researchers, engineers, designers, product specialists, and responsible AI experts from across the organisation.
The team is engaging in a human-centered co-design and development process to ensure the platform is accessible and valuable for users. By focusing on meaningful and quantifiable validation, Science Digital aims to create a tool that scientists can trust to be responsible, accurate, transparent, and effective in delivering value to their work.
Making it real: Use cases
To ensure the platform delivers meaningful and measurable value, the Science Digital team is:
- clarifying the problems users want to solve
- assessing how AI can most effectively help
- building a system that integrates effortlessly into practical workflows.
Experts from across CSIRO will collaborate with Science Digital to develop a platform prototype and apply it to several high-impact scientific studies. This process helps to validate and enhance the platform’s functionality, capabilities, and results.
Genome annotation
Although strides have been made in genome sequencing and assembly, functional gene annotation discovery and inference remains particularly time-consuming and complex for scientists (e.g. biologists), hindering progress of gene function discovery.
This challenge is especially pertinent in non-model species where very little data exists and experimental design/inference is non-standardised/not democratised, as well as for gene families with no or very little experimental characterisation.
To overcome this, a prototype of the platform will be applied to accelerate high resolution functional annotation or gene discovery.
It will use data and computational methods to accurately predict gene and protein function. This use case will train the platform to be useful for researchers in materials science, life sciences, and environmental sciences. It also addresses our sustainability and healthy future challenges.
This case study provides a real-world example of the transformative potential that Dr Stefan Harrer (Science Digital Director), Dr Rahul Rane (Senior Research Scientist), and Dr Robert Speight (Principal Research Scientist) explore in their Lancet Commentary.
'Generative AI agents are transforming biology research: high resolution functional genome annotation for multiscale understanding of life' explores the game-changing integration of AI into biology. By harnessing generative AI and advanced agentic tools, biologists can better interpret genomes, predict biological functions, and design novel systems.
The paper emphasises the potential of these technologies to revolutionise drug discovery, crop development, and sustainable biomanufacturing.
Digitisation of national collection and autonomous labs
Biological collections provide a unique record of species diversity, distribution and biological interactions across space and time.
These collections are vast warehouses of novel genes, molecules and biochemical processes that can be explored to identify the next generation of foods, materials and medicines needed to meet the demands of a rapidly growing human population.
However, the labour-intensive process of taxonomic discovery and documentation remains essentially unchanged and at current rates, will take hundreds of years to unpack the rich and potentially 'game-changing' biological information within the world's collections – information that is sorely needed right now.
This use case aims to automate the digitisation workflow of National Collection's insect collection.
It will do this with AI-powered robotics technology, imaging, data extraction, and data management and analysis in physical environments.
Science Digital aims for this use case to empower researchers to leverage the platform as a tool for designing and executing autonomous lab workflows through assistive AI-driven robots and other physical systems.
This shift from digitisation to autonomous labs will refine robotic accuracy in the lower-risk environment of the national collection, paving the way for operations in more complex labs, domains and settings.
This vision of AI-driven biology was the focus of a paper published by Dr Stefan Harrer, Director of Science Digital, Dr Rahul Rane, CSIRO Senior Research Consultant in Applied Genomics, and Dr Robert Speight, Director of CSIRO's Advanced Engineering Biology Future Science Platform.
'Generative AI agents are transforming biology research: high resolution functional genome annotation for multiscale understanding of life' explores the game-changing integration of AI into biology.
By harnessing generative AI and advanced agentic tools, biologists can better interpret genomes, predict biological functions, and design novel systems. The paper emphasises the potential of these technologies to revolutionise drug discovery, crop development, and sustainable biomanufacturing.
Our Google partnership: a collaboration that enables transformation
To inform, inspire, and catalyse the use of AI in the Australian science community.
As a part of our five-year partnership with Google, Science Digital has entered a partnership with Google that utilises Google's capability to accelerate and transform Australian innovation.
Through the collaboration, Google and Science Digital will develop and showcase practical AI solutions to real-world challenges, including Science Digital's use cases Genomic annotation, and Digitisation of National Collections and Autonomous Labs.
As Science Digital's inaugural strategic partner, Google will play a critical role in raising awareness of Science Digital's initiatives globally, offer technical and ecosystem resources, and encourage scientists to harness the capabilities of AI.