How AI could transform the way public servants work

The Mandarin and Liquid’s recent survey, ‘Let’s get real about AI’, reveals public servants are ready to work with AI tools. Of the 600 survey respondents, 84% said they are personally ready to use AI at work. 

So what’s holding them back? 

At The Mandarin’s Building a Better Public Service conference in Brisbane last week, Nathan Bines, Queensland government’s executive director of data and AI, and Stefan Hajkowicz, chief research consultant at CSIRO, discussed the barriers and opportunities of using AI in the public sector.

“There’s a lot of desire, but people [are] looking for the structure and the support to know how to use AI,” said Bines, who believes the state government is ensuring public servants understand the risks associated with AI use.

“We’ve all lent in and built risk assessment frameworks that allow people to assess the risks with AI,” he said. “We’ve just got to use them and take a risk-based approach.”

Essential reading for Australia’s public service.

Stay ahead of policy shifts, leadership moves, and the big ideas shaping the public service. Sign up to The Mandarin’s free newsletters.

By continuing, you agree to our Terms & Conditions and Privacy Policy.

Bines said it’s time to ask: “How do we enable people? How do we give them the playbook?”

Hajkowicz said waiting for AI to be ‘safe’ isn’t an option. Like any new technology, there is bound to be trial and error.

Public servants are “eager to use it”, he said. “They’re wanting to do it more. I think there is this urgency to start.

“Australian governments are looking at using these models. They rightfully had concerns about data, privacy, confidentiality, safety and behaviour of these models, as they’re managed offshore.”

Bines warned governments often hold onto mistakes as “an excuse not to innovate”.

“But we’ve learned a lot. There’s been more than enough reports and inquiries…we know what not to do,” Bines said. 

Despite caution, there is an appetite for action. In The Mandarin/Liquid survey, 58% of respondents said they are willing to lead AI projects despite risks. But among frontline staff, four in five report feeling either unsupported or unprepared. 

“I think the risks of not doing it are greater. The risks of not creating value for your agency are greater,” Hajkowicz said.

“This is how we will contain budgets, make things efficient, deliver better services and policies. So yeah, the risk of not doing this is pretty big. I think the risk aversion culture in the public sector generally is more prevalent here than the AI technology.”

Hajkowicz said there is no other way forward for the public sector, especially as AI tools continue to get smarter.

The survey shows that four of the top five referenced use cases for AI were operational, including things like reducing paperwork, improving workflow and alleviating day-to-day frustrations.

Both speakers agreed that AI’s potential is huge, but it must be applied wisely.

“Doing things that you couldn’t do before” is the point of AI tools, Bines said. But AI can only go so far. “Every problem has an AI solution in people’s minds, because it’s the first time they really get exposed to [this] technology.”

Bines said different issues need different tools, and we need to build on existing processes.

“Leveraging the platform we already had to focus on this distinct business problem. And that’s really where the future is. It’s about applying AI to distinct problems,” he said.

According to our survey, few public servants think in terms of large-scale transformation, but Hajkowicz thinks we need to be open to AI’s potential.

“The journey we’re on is about [the] transformation of how we do things,” Hajkowicz said. “Let’s use all of the tech that we’ve got today, data science, AI, digital technologies, tracing technologies in that sector, and let’s rebuild.”

He said AI shouldn’t be a band-aid solution but a way to transform productivity.

“Smart AI won’t help a dumb organisation get better. It’ll just do dumb stuff faster,” he said.

Looking ahead, Bines said AI is critical to modernising ageing systems.

“The ability to start building the new world, leveraging AI technology while the old systems are running. In some cases, it’s going to be our only choice. We can’t just recreate another monolithic system,” Bines said.

“The benefit there is we’ve got the opportunity to rethink how those systems deliver the services that they do. I think there’s a great opportunity for generative AI to bring the capability to test and experiment and do things in new ways.”

Artificial intelligence can be beneficial for everyone, not just rich tech companies, Bines said. “I think there’s a big opportunity to bridge that [capability] gap”, so the “benefits of this technology aren’t concentrated at the top”.

Still, many public servants don’t believe their workplaces are prepared to integrate AI. Almost 40% of survey respondents are personally ready to use AI but lack organisational backing.

“The onus is on departments and others to encourage and foster that,” Bines said.

Hajkowicz and Bine say the benefits will be worth it. “This is a possibility to really transform how we work and what we do,” Hajkowicz said.