Full results: cross-agency survey of use cases for artificial intelligence (AI)
Survey process
The Government Chief Digital Officer (GCDO) sent out the survey in and received the following results:
- 37 out of the 50 agencies surveyed responded with at least one AI use case
- 108 total use cases from the 37 participating agencies.
Current and planned government use of AI
The majority of AI use cases identified were focused on internal agency use for improving productivity and efficiency. Examples of use cases for agency productivity and cost efficiency are as follows.
- Quality and efficiency: chatbots to help people more quickly find appropriate guidance and resources, document drafting, meeting summaries, and using search.
- Digital detection and analysis: image analysis and detection for large image datasets, as well as retinal scanning, detection, and diagnosis.
- Automated data summary and analysis: automated sentiment analysis, automated text, data analysis, factors modelling, disputes, submissions, engagement, staff sentiment summaries, and fraud detection.
- Automated transcription: tools that help summarise meeting notes, phone and video calls, hearings and notes.
A smaller set of use cases were focused on improved customer experience. Examples of use cases exploring improved customer experience are as follows.
- Assisted web search: chatbots to help customers find the information they’re looking for more quickly and easily, chat-based search on government websites that aim to support improved customer experience.
- Contact-centre assistance: chatbots to help staff navigate knowledge-sets, to support staff in helping customers, real-time knowledge assist, sentiment tracking, and transcription.
- Customer self-service: automated and predictive forms, to help customers log requests to the right place, help customers to independently resolve issues through automated tools or processes.
Benefits of AI
Out of the 108 use cases reported, surveyed agencies indicate that AI shows potential to:
- boost productivity
- automate routine tasks
- process work more efficiently.
Surveyed agencies responded that AI also helps with self-service options and personalisation of government services, leading to an improved customer experience.
Risks of AI
The survey responses highlighted several risks and barriers to safe and transparent adoption of AI within agencies.
Detailed description of graph
Ordered by the most to least number of responses, agencies listed the AI risks and barriers facing them as:
- skills, capabilities — 22
- policies, guidance — 21
- privacy — 21
- security — 21
- technical barriers — 19
- cost, funding — 18
- understanding, support — 17.
Some of these risks were also identified in the survey, such as:
- skills
- privacy
- security.
cross-agency AI survey: highlights
Next steps for the GCDO and public service
Agencies identified several areas that could benefit from further guidance. The GCDO will cover many of these areas in future updates to the AI guidance for the public service.
Interim Generative AI guidance for the public service
There is some growing maturity with AI among some agencies. Work is under way to share lessons and system insights from these projects with other agencies considering AI in their contexts.
More information and data
Last updated
