Tech in Gov 2026: Designing digital government for better citizen outcomes
Fireside chat — 4 August 2026, 11:30am
Alana Lundy, Chief Information Officer and Chief Information Security Officer, Department of Social Services, in conversation with Julian Bajkowski, Senior Journalist, The Mandarin.
One of the standout sessions on Day 1 of Tech in Gov 2026 was a fireside chat on designing digital government around better citizen outcomes. The conversation covered equity of access, the deliberate rollout of AI, and trust.
From service delivery to policy
The session started with an intro to Alana Lundy, who has been at the Department of Social Services (DSS) for just over two years. Her current role also marks her first time in a policy department, moving from service delivery. She touched on the ‘great divide’ that sometimes exists between the policy side and the technology side of an organisation, noting that the secretary, Michael Lye, has been very good at bringing the policy people and technology people together.
Reaching people who can't engage digitally
A key theme at the start of the session was around digital equity and how AI is starting to shape access. While AI can be used to help people "converse in digital speak", in some cases AI is actually contributing to inequity. For example, Lundy mentioned that people who use AI are more likely to get a grant (such as NDIS funding). So how do we level the playing field?
Lundy framed this as the technologist's challenge: to make the scenario more seamless — the forms, myGov, ID verification and so on. She also raised the opportunity of smaller language models: training them, giving them boundaries, and thinking about what that could mean for access to grants and similar services.
A deliberate, planned approach to AI
DSS is rolling out AI, but it's a slow and deliberate approach, done in a planned way. Lundy cited grants as a real opportunity to make the process secure and faster.
Lundy was clear that the department needs to be transparent and ethical, and that this means some people in the department can't use AI at all. They're being very deliberate about when AI can be used and when it isn't appropriate. Lundy said that staff should use as many tools as they need to get their job done, but the governance needs to sit at the corporate level, so that individual APS staff aren't the ones making those calls.
She also made the point that government has been doing some form of AI for at least 20 years — automation and advanced analytics — sitting on proprietary systems. So some of that knowledge may already exist in the organisation; the question is what it looks like now.
Trust, data and government
Lundy talked about the cultural shift that’s happening no matter where you work: people are thinking differently about their data and where they put their information.
For government, that comes back to trust. Citizens need to trust DSS and Services Australia. Trust is very quickly eroded, particularly in instances involving data breaches. In terms of AI, Lundy focused on the need to work out how to do it safely and securely.
Safety by design
Julian Bajkowski asked what safety by design looked like. He noted that trust has been an issue and referenced the Royal Commission into the Robodebt , while also highlighting that the robodebt issues had been human errors rather than issues with automation. He then asked what lessons had come out of that.
Lundy emphasised keeping a human in the loop on the AI journey. She was candid that the lessons learned were a bit paralysing for a while: nobody wanted to make any decisions. She described it as almost like a grief period, reflecting on what had been done wrong and how to improve.
She talked about a committed workforce: people who were there when it happened, or who joined since, because it happened, and who are driving policy because of their core values. The lesson learned is to reflect and not make the same mistakes again. There's been some growth from that — people know the history and want to do better. She's seen this at DSS, and the same at Services Australia: it's part of their past, but they need to make better decisions, and to think about the impacts on humans when looking at technology. Applying this to AI, the lesson is to tread carefully with AI, and keep a human in the loop.
People, skills and what the conversation is really about
Bajkowski found it fascinating that some people had joined the organisation specifically to do this kind of work, and asked whether people are coming into the APS looking at the human application of technology.
Lundy agreed. She described AI as a tsunami — people need to decide for themselves what they're going to buy into. She's also seeing a change in the way people interact with technology.
The technology isn't the conversation; the conversation is about what we're delivering and who we're delivering it for.
Asked how much easier it is now for people from a non-technical background to gain technical skills, Lundy said it varies from person to person. People have far more choice now, and in government the focus is on what the department does, with the technology coming later.