Tackling AI complexity and trust in the Australian public sector

Australia is facing growing productivity challenges, with pressure to improve efficiency in the public sector. Productivity plays an important role in driving long-term economic growth, wages, inflation and interest rates — all key factors as Australians navigate a cost-of-living crisis and a challenged economy.

Amid these challenges, artificial intelligence (AI) has been driving some much-needed optimism. The recent workplace AI trial by the Australian Public Sector (APS) revealed that a significant majority of managers (64%) observed improved efficiency and quality within their teams. This promising development highlights the transformative potential of AI in the public sector.

However, research by Avanade shows that where most sectors expect up to a four-times return on every $1 spent on AI copilots and agents, government organisations typically expect only up to a two-times return within 12 months. This is primarily owing to a myriad of challenges that government organisations need to tackle: from building trust and managing ethical concerns to data security and skill development.

A game changer for public sector productivity

In a sector where a huge amount of time and energy is spent on mundane and repetitive tasks, AI’s ability to process vast amounts of data and automate routine tasks has been a game changer. By leveraging AI, government agencies can dramatically streamline operations, reduce administrative burdens and enhance decision-making processes. For instance, AI-powered chatbots can handle a multitude of citizen inquiries — freeing up human resources for more complex tasks as a result. Similarly, chatbots used for internal operations can also be incredibly useful in quickly locating information from thousands of legal, regulatory or public documents.

The APS trial is a testament to AI’s potential, with 69% of survey respondents agreeing that the generative AI tool on trial improved the speed at which they could complete tasks and 40% of respondents reporting reallocating their time for mentoring/culture building, strategic planning, engaging with stakeholders and product enhancement.

Organisational challenges in AI adoption

Despite its productivity potential, AI adoption in the public sector is not without its challenges. One of the primary concerns is data security: public sector organisations manage sensitive information, and any breach can have severe consequences. To mitigate these risks, robust data protection measures and compliance with regulations are essential.

Skills development is another significant barrier. AI’s impact depends on the ability of employees to harness its capabilities effectively. Investing in training and development programs is crucial to equip employees with the necessary skills to leverage AI well. This was clearly demonstrated in the APS trial, which shows that 75% of the AI trial participants who received three or more forms of training were confident in their ability to use the tool — 28 percentage points higher than those who received only one form of training.

Change management also poses a challenge. Integrating AI into existing workflows requires a cultural shift. Resistance to change can hinder AI adoption, making it imperative to foster a culture of innovation and continuous improvement.

Addressing trust and ethical concerns

Building trust in AI is paramount for its successful implementation, especially in the public sector, where trust in government is fundamental to maintaining public confidence. Our research shows that 34% of public service organisations believe that building trust and addressing ethical concerns are the most daunting obstacles they face in realising AI’s full potential.

Algorithm bias is at the top of the list when it comes to ethical concerns. If left unchecked, bias can lead to unfair outcomes, undermining public trust not only in AI but also in the institutions that use it. In fact, when asked about legal considerations relating to AI adoption, 36% of public agencies told us they were extremely concerned about the amplification of biases, leading to discrimination.

Overcoming challenges and delivering value faster

To overcome these challenges and deliver value more rapidly, public sector organisations can adopt several strategies. Firstly, they can consider working in partnership with private sector entities and academic institutions to gain access to cutting-edge AI technologies and expertise. The right partnership can be the stepping stone for agencies to accelerate AI adoption and innovation.

Secondly, mitigating risks and building confidence can be achieved by adopting a phased implementation strategy. Starting with pilot projects, such as the APS Copilot trial, empowers organisations to test AI applications, gather valuable insights and make informed adjustments before scaling across the organisation.

Thirdly, effective AI adoption hinges on robust governance. Public sector organisations must invest in comprehensive data and AI governance frameworks that encompass clear policies, auditing processes and risk management strategies. Prioritising data quality, security and privacy lays a strong foundation for AI initiatives.

Ethical considerations must also be at the forefront, with guidelines and frameworks that address bias, fairness, transparency and accountability. It’s also important to note that frameworks can’t function in silos. They needed to be accompanied by human action. Organisations that provide ethical AI practice training and education to employees help foster a culture of responsibility and awareness.

Lastly, developing a culture of innovation is crucial. Encouraging experimentation and rewarding innovative ideas can help drive AI adoption and integration. This is where leadership will also play a pivotal role in championing AI initiatives and inspiring a forward-thinking mindset within the organisation. Public sector organisations should also prioritise continuous learning and development to bridge the skills gap and keep pace with AI advancements.

Top image credit: iStock.com/BlackJack3D