Optimising the value of AI in the public sector

Australia has great potential to leverage machine learning and AI technologies to accelerate digital transformation and boost its competitiveness on the global stage.

The trajectory of generative AI is poised to pivot from conceptualisation to tangible application, as industries across the board ramp up adoption this year. GenAI capabilities are maturing, particularly within enterprise realms such as content creation, software development and data scrutiny.

In this climate of technological advancement, organisations spanning various sectors, including the public sector, are harnessing GenAI to foster creativity and enhance productivity. In parallel, the Australian Government has initiated an AI policy committee¹, integrating perspectives from both the public and private sectors, to oversee the deployment of AI technologies.

Australia has great potential to leverage GenAI, machine learning and other AI technologies to accelerate its digital transformation and boost its competitiveness on the global stage. Recent research² shows AI has the potential to add between $1 trillion and $4 trillion to Australia’s economy in the next decade, significantly boosting the country’s current annual GDP of about $1.5 trillion. GenAI will also reshape the future of work, leading to an estimated 1.3 million job transition in Australia³ by 2030.

AI initiatives in the public sector in Australia

Across Australia and New Zealand’s public sector, AI is already being used to streamline administrative processes and improve the quality of public service. For example, the Queensland Government uses the CSIRO Patient Admission Prediction Tool (PAPT) to forecast the demand for hospital resources and cut patient waiting times at 27 major hospitals across the state. Recent initiatives such as predictive policing by the NSW Police Force, utilising AI, further demonstrate how AI can change and improve the service within the public sector.

There is also great potential for AI to revolutionise traditional industries such as agriculture, farming and manufacturing by optimising production, distribution and consumption. For example, the eGrazor collar system for cattle by CSIRO in collaboration with the NSW Department of Primary Industries utilises sensors and machine learning to monitor the grazing behaviour of cattle, enabling the prediction of pasture intake based on their movements.

The risks and limitations hindering adoption

Implementing AI solutions at scale is not without its challenges or risks. Operational challenges include varying stakeholder interests; concerns about governance, data privacy and security; high implementation costs; existing government frameworks; and ethical risks. Gartner reported⁴ that more sophisticated cyber attacks and growing regulatory obligations are keeping Australian organisations on their toes this year, spending more than $7.3bn — an 11.5% increase — on security and risk management from 2023. However, AI models are only as good as the data they are fed. Therefore, trusting AI begins with first trusting your dataset. Agencies must have robust data governance frameworks in place to ensure their data assets are trusted and protected.

As the environment for AI development in Australia is still maturing, there may be resistance or lack of support for AI initiatives. Recent research by Cloudera shows that the main challenges for APAC in regard to deploying AI include: scalability and deployment (38%), change management and organisational culture (36%) and data quality and availability (35%).

On this, Australia has recently developed a new Cyber Security Strategy, which — while having a high focus on the cybersecurity landscape — also offers insight into the changes for the digital landscape and data management by 2030.

Furthermore, the shortage of skilled professionals who can develop and implement AI technologies, and the existing ICT infrastructure, may not be sufficient to support the advanced requirements of AI technologies.

There are also ethical risks associated with AI⁵, such as the potential for bias in AI algorithms. These challenges require complex, effective and efficient strategic regulations to fulfil the vision of AI development, such as regulatory frameworks that clearly set out the requirements and implementation of AI technology.

Harnessing the value of AI to serve citizens

As government agencies continue to look to AI to transform operations and citizen services, here are some suggestions on how agencies can optimise the value of AI today while balancing risks and limitations:

1. Start small: Limit access and capabilities initially. Start with narrow, low-risk use cases. Slowly expand capabilities as benefits are proven and risks addressed.
2. Improve dataset quality: Ensure you can trust your data by using only diverse, high-quality training data that represents different demographics and viewpoints. Make sure to audit data regularly.
3. Develop mitigation strategies: Have plans to address issues like harmful content generation, data abuse and algorithmic bias. Disable models if serious problems occur.
4. Identify operational problems AI can solve: Identify and prioritise potential use cases by their potential value to the organisation, potential impact and feasibility.
5. Establish clear AI ethics principles and policies: Form an ethics review board to oversee AI projects and ensure they align with ethical values. Update policies as needed when new challenges emerge.
6. Implement rigorous testing: Thoroughly test generative AI models for errors, bias and safety issues before deployment. Continuously monitor models post-launch.
7. Increase AI model explainability: Employ techniques like LIME to better understand model behaviour. Make key decisions interpretable.
8. Collaborate across sectors: Partner with academia, industry and civil society to develop best practices. Learn from each other’s experiences.
9. Enhance AI expertise within the government: Hire technical talent, and provide training on AI ethics, governance and risk mitigation.
10. Communicate transparently with the public: Share progress updates and involve citizens in AI policymaking. Build public trust through education on AI.

The transformative power of AI technologies, particularly generative AI, has the potential to significantly boost economies and revolutionise industries. As seen in Australia, AI is already being leveraged to streamline administrative processes, improve public services and optimise traditional industries. However, the implementation of AI solutions at scale within the public sector is not without its challenges. These include operational hurdles, data governance, privacy and security concerns, high implementation costs and ethical risks.

To optimise the value of AI while balancing these risks and limitations, it is crucial to consider the building blocks above to harness the full potential of AI to transform operations and citizen services, ultimately driving economic growth and societal advancement.

<sup>1. Department of Industry, Science and Resources 2024, New expert group will help guide the future of safe and responsible AI in Australia, Australian Government, <<https://www.industry.gov.au/news/new-expert-group-will-help-guide-future-safe-and-responsible-ai-australia>>
2. Technology Council of Australia 2023, Generative AI could contribute $115 billion annually to Australia’s economy by 2030, <<https://techcouncil.com.au/newsroom/generative-ai-could-contribute-115-billion-annually-to-australias-economy-by-2030/>>
3. Mitchell A and Bruce D 2024, Proceed with caution: Three questions for Australian governments to answer as they consider gen AI, McKinsey & Company, <<https://www.mckinsey.com/au/our-insights/australia-and-new-zealand-perspectives/proceed-with-caution-three-questions-for-australian-governments-to-answer-as-they-consider-gen-ai>>
4. Tan A 2024, Australia’s cyber security spending to grow 11.5% this year, TechTarget, <<https://www.computerweekly.com/news/366574452/Australias-cyber-security-spending-to-grow-115-this-year>>
5. Australian Human Rights Commission 2023, Australia needs to be a world leader in ethical AI, <<https://humanrights.gov.au/about/news/opinions/australia-needs-be-world-leader-ethical-ai>></sup>

*Keir Garrett is Regional VP ANZ at Cloudera. Keir brings more than 20 years of management, strategic consulting and digital transformation experience to Cloudera, and has successfully developed lines of business in global markets and across multiple industries, both directly with customers and in collaboration with the partner ecosystem. Most recently, Keir spent two years at Crayon in the role of Chief Executive Officer and prior to this was Head of Software, Cloud, Advisory, Professional & Managed Services Sales at Datacom. She has also held several senior positions at Microsoft, SAP and Infor.

Top image credit: iStock.com/cherdchai chawienghong