Artificial intelligence (AI) is already reshaping how councils plan, deliver and maintain services. But what actually is AI and how should local government approach it? Drawing on insights from a recent industry presentation, this article breaks down the fundamentals and what they mean in practice.
AI has quickly become one of the most talked about technologies in local government. Councils are trialling it in customer service, asset management and planning, while vendors are embedding it across platforms and services.
Yet for many, the conversation has moved beyond their understanding.
Speaking at the Municipal Works Australia Conference in Bendigo, Daswin de Silva, Professor of AI and Analytics at La Trobe University, set out to answer a deceptively simple question; “What’s all the fuss about?”
His answer was clear. AI is not just another tool, it is a new layer of technology that is driving a paradigm shift on the scale of the Industrial Revolution.
“Across the board, AI is impacting everything that we do,” de Silva said.
To explain the scale of change, de Silva pointed to the numbers. Millions of users, billions in market value and trillions in projected economic impact.
“Close to ten per cent of the world’s population is using an AI tool,” he said.
Those tools are led by platforms such as ChatGPT or Claude, which have brought generative AI into the mainstream by allowing users to interact with powerful language models through simple, conversational prompts.
The strength of these Large Language Models (LLMs) lies in their versatility, from drafting documents and analysing data to supporting decision making, which has made it a default entry point for both individuals and organisations.
At the same time, companies including NVIDIA are leading on a different front by developing the specialised chips and computing infrastructure that make modern AI possible. These systems require enormous processing power to train and run, and NVIDIA’s graphics processing units have become the global standard for that work.
Together, these software platforms and hardware providers sit at the centre of the current AI ecosystem, shaping both how the technology is used and how quickly it continues to advance.
de Silva explained that the general-purpose nature of this technology is what makes it so universally transformative.
He added that AI could be placed in the same category as the steam engine, electricity or the computer. Technologies that do not just improve a single task but reshape entire economies.
“Where the Industrial Revolution automated physical labour, AI is beginning to automate some aspects of human thinking. It is mental work that is becoming automated,” de Silva said.
This distinction has major implications for councils. Much of local government work sits squarely in this category, from planning and compliance to customer engagement and reporting.
“It calls for reinvention and also resilience in the face of significant and disruptive change,” he said.
That disruption is already visible. de Silva pointed to recent job losses in software engineering as an early example of AI’s impact.
“AI is writing the code,” he said. “It can produce this at scale, and 24/7.”
At the same time, he emphasised that not all work will be affected equally. AI-generated code must still be evaluated by humans and tested across diverse real-world performance benchmarks.
AI is still confined to the digital world. This absence of real-world physical experience means trades and hands-on roles are less exposed, but even within those jobs, some tasks like customer engagement, image processing and workflow documentation, can be automated.
“Any job can be unpacked into a range of different constituent tasks, and some of those tasks can be automated,” he said.
Beyond the chatbot
For many people, AI is still synonymous with chatbots. Typing a question into a system and receiving a response. de Silva argued that this is a narrow view.
“It might be challenging to think of AI in this way, because usually it defaults to a tool,” he said. “It is not like simply knowing how to use a tool, like Excel.”
Instead, he encouraged organisations to think more broadly.
“Organisations must start to think of AI as part of systems, processes and the workforce, a digital species or digital companion,” he said.
This idea reflects how AI is evolving. Early systems responded to prompts. Newer systems are beginning to act more independently, completing tasks, suggesting next steps and coordinating multiple actions.
“It’s like having your own little army of digital companions,” de Silva said.
The shift from tools to assistants and now to agents is central to understanding AI’s potential. It moves the technology from something that supports work to something that actively participates in it.
Assistants typically refer to systems that respond to prompts and help users complete defined tasks, such as drafting text. Agents go a step further, operating with a greater degree of autonomy by breaking down goals into steps, making decisions about how to complete those steps, and interacting with other systems or tools to carry out tasks.
More importantly, those systems are beginning to anticipate user needs.
“When modern models have completed a prompt for the user, they will have already considered the next two or three steps that you might have to do,” he said.
For councils, this opens the door to more proactive and integrated uses of AI, particularly in areas involving complex workflows or decision making.
To make sense of how AI fits into existing systems, de Silva described it as a layer built on top of data.
“First came the physical systems, such as infrastructure and machinery. Then came the data layer, with sensors and digital platforms providing visibility and insight,” de Silva said.
“On top of that sits the AI layer. This layer is where decisions are made. It makes the decision partially or entirely on behalf of the human users.”
In practice, this can mean identifying when an asset is likely to fail, optimising how resources are allocated or streamlining processes that once required multiple steps and stakeholders.
“It can streamline or expedite what used to take several people and a lot of back and forth,” he said.
For local government, this has direct relevance across both office and field-based work.
In office environments, AI is already proving effective in tasks such as drafting, planning and reporting.
“A lot of the boring documentation type work, AI has perfected this,” de Silva said.
He pointed to examples such as automated meeting summaries and action tracking, which can reduce administrative burden and keep projects moving.
“We use that in the university,” he said. “It’s highly effective.”
In field work, the opportunities are more complex but equally significant.
“AI is not built for the physical environment,” he said. “AI can take data from devices like cameras or LiDAR, but the results of these kinds of detectors can be affected by something as small as the difference between it being sunny or cloudy.”
Despite the challenges of tuning AI to a changing environment, de Silva said that applications such as defect detection, predictive maintenance and safety monitoring are advancing quickly.
While the long-term potential of AI is transformative, de Silva emphasised that organisations should start with practical steps.
“It starts with having an AI-first mindset on productivity,” he said.
“This means looking at existing processes and asking where AI can improve output or quality. Given ‘X’ input, what’s the output, and is there an opportunity to explore how AI can improve the quality of that work?”
From there, organisations can move to innovation, doing things differently and identifying new opportunities. For councils, this might mean new ways of engaging with communities, delivering services or managing assets.
However, de Silva cautioned that realising these benefits requires deliberate effort.
“Consider where your teams spend most of their time. ‘In what areas is productivity being lost?’ These can be the best starting points for AI adoption,” he said.
Alongside its benefits, AI introduces new risks that councils must manage. One of the most immediate is accuracy.
“There can be what we call ‘hallucinations’,” de Silva said.
He referred to an incident in 2025, when Deloitte Australia delivered a flawed welfare compliance report to the Department of Employment and Workplace Relations (DEWR) that – having used generative AI – fabricated citations, legal quotes and academic references.
These false details are referred to as AI hallucinations and arise from the way LLMs can generate responses by predicting patterns of language rather than retrieving verified facts from a database, or applying real-world knowledge and experience as human workers would do.
“That’s one great example of the importance of keeping a human in the loop,” de Silva said.
Other risks include data sovereignty and reliance on overseas providers, which raise questions about control, privacy and alignment with local values.
“There can be certain omissions of information that align with ethics or belief systems,” he said.
Globally, different approaches are emerging. The European Union has introduced comprehensive regulation, while Australia is taking a more measured approach.
Building capability in councils
For councils looking to get started, de Silva stressed the importance of strategy.
“It’s always good to have a strategy, an idea of where you’re headed,” he said.
“This begins with defining purpose and identifying use cases, but also requires attention to governance, data readiness and workforce capability. Data readiness is crucial.”
de Silva noted that accessing and preparing data can often be a significant barrier.
He also highlighted the need for new skills.
“Prompting AI is a lot of back-and-forth work,” he said.
This creates opportunities to engage with universities and emerging talent, particularly students who are already experimenting with AI in their studies.
“The future is AI powered,” he said.
“For councils, that future is not abstract. It is already beginning to take shape in the systems they use, the services they deliver and the expectations of their communities.
“The challenge is not simply adopting new tools but understanding the shift underway and responding with intention.
“In that sense, the question is no longer what AI is, it is how local government chooses to engage with it.”
