Artificial Intelligence (AI) and automation will continue to play an important role in government, and when implemented appropriately, have huge potential to help our citizens and communities.
We spoke to Jessica Ho, Director of Investment and Assurance within Digital Strategy, Investment and Assurance (DSIA), to hear her insights on the AI Assurance Framework. She also shares how NSW Government balances the benefits of AI investments while managing potential risks.
The important role of the AI Assurance Framework
At the heart of the NSW Government’s approach to AI is the AI Assurance Framework, a tool designed to evaluate AI projects. “The AI Assurance Framework is a set of risk questions which allow projects to self-assess themselves to help determine if they need additional assistance from the AI Review Committee,” says Jessica.
The risk questions in the AI Assurance Framework mirror the AI Ethics Policy. By completing the self-assessment, project teams can assess whether their proposed solutions involve risk in how the AI system is used, including whether it has an operational impact or if it is not transparent or not explainable.
“We all acknowledge that AI has risk in government and in any other industry. If the government is going to invest in a technology solution that involves AI, we need to consider how the government should treat that particular investment piece.” Jessica Ho, Director of Investment and Assurance.
Managing the NSW Government’s investments in AI
While the AI Assurance Framework plays a crucial role in managing AI opportunities and risks, Jessica emphasises her team’s keen interest is in AI investments. Since 2019, the Digital Restart Fund (DRF) has allowed for significant growth in AI projects and investments.
To help manage AI investments, Jessica and her team are integrating the new AI risk factors into the Investment Framework (Gateway Assurance). This will help ensure that the government understands the opportunities and risks of AI investment.
Striking a balance between opportunity and risk
There’s always a delicate balance between harnessing AI’s benefits and mitigating potential risks. Jessica noted that the technology itself is not the risky part, the challenge lies in the solutioning and implementation of AI projects. “Once the government understands the investment and riskiness of the solution proposed, then they can make a decision about whether they want to accept the risk, defer the risk or mitigate what they have seen.”
Jessica also emphasises that AI and automation has been around for a long time, but has recently become a public focus as generative AI has become more prominent. “Predictive AI has been around for a long time. The NSW Government has been using AI since 1999 - it’s not new in government,” says Jessica. “We want to regulate AI by taking the balanced approach of understanding the investment risks and what we’re putting out there for our citizens.”
The enormous potential of AI to help our communities
It’s clear that AI, when implemented appropriately, has enormous potential to help citizens. “There’s so much AI can do to improve regulations, healthcare, transport, reduce costs, enhance sustainability and improve efficiency,” says Jessica.
Real-world examples of AI projects that are funded by the Digital Restart Fund and have been through the assurance process are Asset AI and Smart Irrigation Management for Parks and Cool Towns (SIMPaCT). Asset AI uses artificial intelligence (AI) to revolutionise road maintenance and operations to make roads safer for our communities. The project will help councils and the NSW government streamline the way they maintain roads by using near-real-time data on the condition of roads.
SIMPaCT is an innovative new approach to using and conserving water in the urban environment. It tackles three of the biggest challenges facing our communities this century: urban heat, water scarcity and the need to provide quality green public space. The SIMPaCT system ‘learns’ and optimises the watering regime to create a cool, green environment for the park and surrounding residences. Information on the coolest and warmest parts of the park is available to the community on a mobile app.
Ethical considerations and the AI Assurance Framework
Making sure AI solutions adhere to ethical principles is extremely important to the NSW Government. “We use the AI Assurance Framework to make sure that a particular solution meets the AI Ethics Policy,” says Jessica. “We’ll look at it from an investment risk perspective, so both sides have been considered, and then we continue to monitor the project until its end.”
Human oversight is always included as part of the process when automated decision-making is used in NSW Government. This helps to safeguard against any errors or misuse. “If you’re using automated decision-making, you still need a human to sign off on your final decision. For example, if we’re using a mobile camera detection system, there’s a person that has sign off on that file to determine whether someone is driving holding a mobile phone or if they’re holding something else. That level of protection still exists in NSW Government.”
By balancing opportunity and risk through the AI Assurance Framework and AI Ethics Policy, the NSW Government is ensuring that AI investments are implemented safely and to the benefit of the citizens of NSW.
If you’re interested in finding out more about the AI Assurance Framework, you can read more here.