Artificial Intelligence
Bookmark

Public servants key to unlock value of AI for the government

By Apolitical

Global network and platform for public service, Apolitical, has published an insights report on GenAI use in governments with case studies from nine countries, including Singapore.

Apolitical's new insights report features in-depth case studies across Singapore, South Korea, Australia and other countries on how these governments are progressing with GenAI adoption. Image: Apolitical

What it is:  

Apolitical, a global platform and network for public service, has published an insights report on the government’s use of generative artificial intelligence (AI), comprising of five briefs. 

 

The report assessed 311 government GenAI projects from nine countries, Australia, Denmark, Finland, Germany, the Netherlands, Singapore, South Korea, UK and the US. 

 

The first two briefs focused on key insights to unlock GenAI adoption in government, while the next three highlighted the three types of GenAI application being pursued.  

 

The report features valuable case studies from the featured countries.

 

We recommend viewing this report on a desktop for easier readability. 


Key findings: 

  • The impact gap: Lack of GenAI solutions in scalable areas where there are high volumes of routine citizen interactions. Most initiatives focus on internal operations to create and manage data and knowledge. 
     

  • Three levels of AI use in the government: Level 1 was using commercially available tools like ChatGPT and Gemini; Level 2 was building customised AI solutions using proprietary data; and Level 3 has seen the creation of platforms that allow public servants to build their own AI solutions (similar to what AI providers have been doing). 
     

  • Two main challenges for governments: A lack of training and skills (first level of adoption has been using commercially available tools like ChatGPT and Gemini, and next level is building internal solutions which required public servants to manage through prompting, reviewing and editing outputs), as well as not designing AI for scale (rarely embedded into organisational workflows and direct citizen interactions). 


Three takeaways for government agency leaders: 

  1. Pursue fast, iterative and responsible experimentation in big opportunity areas, and co-design with public servants. 
     

  2. Encourage an open-minded approach to developing solutions, as AI might not be what’s needed. 
     

  3. Engage both technical and business units to understand how AI will affect existing processes and services.