The future of Generative AI in policy work

Many bold claims have been made about Generative Artificial Intelligence (GenAI) and its capacity to improve productivity and generate efficiencies within the workplace.

In recent years, there have been significant developments in digital technologies. Increased processing power and decreased prices for many technologies make them more accessible and provide more potential for use in a wider range of contexts. The field of Artificial Intelligence (AI) has made significant impact on several areas of public services, such as through chatbots and virtual assistants, data analysis, visualisation, and text recognition and processing.

According to a recent Microsoft survey, 24% of leaders in the private sector have already deployed GenAI across their organisations. Many are considering laying off staff and replacing them with GenAI systems.

Researchers at UNSW Canberra and Australian National University have undertaken a national study examining the current and anticipated use of GenAI in policy work. The research by Dr Jade Hart (UNSW Sydney), Professor Helen Dickinson (UNSW Sydney), Kate Henne (ANU), Dr Vanessa McDermott (UNSW Sydney), Dr Shibaab Rahman (UNSW Canberra) and Associate Professor James Connor (UNSW Canberra) explores attitudes within the Australian bureaucracy to using this technology.

Between November 2024 to January 2025, senior public servants from 22 state, territory and federal government agencies participated in research to understand how GenAI is used in policy work, including perspectives of risks and enablers for adoption. The research sought to understand how senior public servants perceived the strengths and weaknesses of these applications, the potential of GenAI for policy work, and what needs to happen to make GenAI safe to use in policy work.

Read the full report and learn more about the research on the project website.

Read their article published in The Conversation.

Watch the seminar recording of The future of Generative AI in policy work in Australia, presented in partnership with the AI CoLab Alliance here.