How does government enable and/or use industry as a multiplier for AI projects?
By Adam Reid, Chief Executive Officer, Department for Industry, Innovation and Science - Government of South Australia.
This article is an extract from Artificial intelligence: your questions answered, a report published in partnership with the Australian Strategic Policy Institute (ASPI).
Traditional models of innovation are typically based on new ideas and inventions being exploited by new businesses that can disrupt markets and gain competitive advantage over incumbent enterprises. We’ve seen many examples of businesses, including technology businesses, that have lost their once-dominant position due to emerging technologies and competition from more innovative and nimble businesses that are taking risks and investing in disruptive technologies to gain market share. The challenge remains to convince established businesses to realise the value of investing in innovation or risk losing their competitive advantage to more agile enterprises with a greater appetite for risk, and reward.
Australian business investment in R&D and innovation is largely heading in the wrong direction. Business expenditure on R&D sits at under 1% as a share of GDP, down from a historical high of 1.37% in 2008–09. Some of the world’s most advanced economies are recording business expenditure on R&D of over 2%, nearing 4% of GDP in the case of Israel.
Although the share of innovation-active businesses has risen in the past decade in Australia, relatively few of those enterprises are producing innovations that are new to Australia and the world, and our share of business expenditure on innovation mirrors our performance on R&D compared with the rest of the world. The performance of South Australia’s business sector is broadly consistent with those national trends.
Clearly, a continuation of this situation is going to have a negative impact on our current high standards of living as competitive advantage and the rewards from investing in innovation shift to our competitors overseas. The challenge facing Australian governments is obvious, although the solutions are often complex—especially given the significant shift in culture required of industry and often government itself.
In the past few years, the South Australian government has been working to encourage greater business engagement in R&D and innovation by facilitating collaborative partnerships between the research sector and industry. Our ‘EXCITE’ Science Strategy aims to connect ideas and expertise with small and large businesses through initiatives such as the Innovation Challenge on Augmenting Ability in partnership with MTP Connect, which will bring together businesses, researchers and innovators to develop cutting-edge products in the health, ageing and disability sectors.
Our external innovation and translation intermediaries will work to bridge the gaps in information and access that can hamper innovation and will drive collaboration between businesses and researchers within the state’s innovation districts. The government has committed up to $3 million to appoint our first intermediary organisation for Adelaide BioMed City, and further intermediaries are planned for our innovation districts at Tonsley in Adelaide’s south and Lot Fourteen in central Adelaide.
EXCITE’s Frontier Technology Centres program will work with small and medium-sized enterprises (SMEs) and business leaders to harness the capability of emerging technologies. As a precursor to that program, the state government made strategic investments to grow leading-edge capabilities in key strategic technologies, such as photonics and advanced sensing, satellite technologies, cybersecurity, and AI.
While AI has been part of our daily lives since at least the 1980s, the surge in awareness—driven by the growing digitalisation of business and processes, increases in computing power and access to large datasets for training AI algorithms—is driving major new investment in AI technologies, including in South Australia.
The state has seen the benefits of long-term investment in AI, stretching back to the government’s initial investment in the Australian Centre for Visual Technologies (ACVT) as early as 2007, in collaboration with the (then) Defence Science and Technology Organisation and other industry partners. That relatively modest investment provided the impetus for ACVT to forge partnerships beyond defence and security and to work with a range of diverse businesses, such as the Victorian sports statistics company Champion Data, medical image analysis company LBT Innovations, simulation company Sydac (now part of the global Sogeclair Group) and digital media company Monkeystack.
This approach to industry engagement by ACVT continued to grow and evolve through the Australian Institute for Machine Learning (AIML), which launched in 2018 with a significant investment of $7.1 million from the state government. Based at Adelaide’s latest innovation district, Lot Fourteen, AIML has committed to renewed engagement with industry and government, which has seen the institute grow in size and capability to become the largest university-based machine-learning research group in Australia and one of the global top five sites for computer vision research output.
The calibre of AIML’s research has attracted the attention of some of the world’s largest multinational corporations, such as Amazon, Accenture, Deloitte, Microsoft Azure, MTX, Google Cloud and Nokia 5G, all of which have established a presence in Adelaide to collaborate with researchers at AIML.
Importantly, the government’s investment in AIML is supporting the institute to engage with SMEs that are looking to automate and harness AI to gain competitive advantage:
- Adelaide visual-effects studio Rising Sun Pictures has produced special effects for some of Hollywood’s biggest blockbuster films and worked with AIML to develop novel VFX tools that have enabled Rising Sun to deliver timely and superior results to global production studios.
- Geospatial tech company Aerometrex used AIML’s deep-learning capabilities to develop highly accurate 3D maps that can reveal shadows cast by buildings at different times of the day, helping planners and developers to gauge the impact on surrounding areas.
- Regulatory technology business Neo Analytics worked with AIML to apply machine-learning models to improve its regulatory compliance monitoring software to deal with significant amounts of data necessary for financial institutions to carry out their business.
As well as supporting SMEs to explore the benefits of AI, government has a role to play as a regulator and service provider by exploring uses of AI to improve accuracy and efficiency across a range of functions.
The state’s Department of Primary Industries and Regions worked with AIML to apply machine learning to satellite imagery to assist land condition monitoring across South Australian pastoral leases. AI is also being applied in the resources sector as the state’s Department for Energy and Mining works with AIML researchers to apply deep learning to South Australia’s vast collection of geological survey data. The work has resulted in a model to predict rock outcrops across the state, which will benefit mineral exploration and help South Australia retain its position as a top 10 destination in the world for mining investment attractiveness.
Clearly, the government’s early investment in frontier AI technology is starting to bear fruit as the state sees a nascent community of more than 50 AI and AI-enabled businesses, ranging from start-ups to global multinationals, tapping into world-class research and extensive datasets and applying analytical tools to get the edge in competitive markets.
Our ongoing challenge is to convey the benefits of innovation, including the potential benefits of AI and automation, to the broader business sector. By demystifying the technology and facilitating networks across business, academia and government, South Australia is building a strong ecosystem that will drive employment and exports and attract talent to work on the next generation of disruptive technologies.
This article is an extract from Artificial intelligence: your questions answered, a report published in partnership with the Australian Strategic Policy Institute (ASPI).