AI for all: Inclusive design principles for the use of AI in the public sector

The use of AI in the public sector has the potential to radically change the way government services are designed and delivered. It offers local governments opportunities to boost efficiency, effectively manage legacy content, and can support better access to services for citizens.  

But with this potential also comes the risk of creating additional barriers for users with access needs, and those already at risk of digital exclusion. So how can organisations in the public sector make sure they are implementing AI solutions in an inclusive way?  
 
When including AI technologies as part of digital products or services, organisations should adhere to some key principles to meet the needs of all citizens, including those with disabilities. 

How can AI support the delivery of inclusive products and services?

Over the last year we've been exploring how AI is used to support the delivery of inclusive services. Some of this has been within our own project work and how we collaborate within the team. We've been experimenting, using AI tools like Copilot to support a more inclusive digital workplace and foster collaborative, meaningfully productive ways of working. 

We’ve been reviewing and auditing other public facing products and services to see what we can learn from how they are embedding AI to benefit their users. Our Accessibility Team Lead Danny Lancaster has recently published a blog post about some of the tools we’ve found. We’ve also explored the role AI can play in improving the user experience for people interacting with contact centres, and those who work in them too, something that could have broad applications in the public sector. 

As part of that review process, we identified a number of attributes across these products and services that helped ensure that AI was being used in a thoughtful, inclusive, and user centred manner.

These are Nexer’s principles for inclusive design for AI: 

Involve real users. Many AI technologies are still in their emergent state. This means user expectations about the technology, it’s potential, and its challenges still need to be understood. By including users with disabilities and access needs throughout the design and development of your products and services you can ensure that you are meeting their expectations, and using technology that helps them achieve their goals. This could be done through traditional user research methods such as interviews and usability studies, but also consider using more collaborative approaches such as community panels, co-design and co-production sessions.  

Focus on user needs and their context. AI provides the most value when it's integrated as part of a properly researched and designed user journey.  Take the time to understand what the users of your product or service are trying to do, and the situations and environments they are in when they are doing them. This includes understanding the user's physical and cognitive abilities, as well as their cultural and social backgrounds.   

Remove barriers. Use AI to enhance your products and services and remove the barriers that prevent users from achieving their goals. Explore the gaps in existing service provisions and design interactions that leverage AI to support users seamlessly through their full journey   

Provide control: Ensure that users remain in control. Make it clear when and where AI is being used and provide controls for customising AI settings or turning off AI enhancements entirely. 

Design with transparency: Clearly identify AI-generated content that may require human validation and educate users on the limitations of generative AI.  

Maintain quality and accuracy. Many existing AI models are built using datasets that lack high-quality examples of inclusive and accessible design, and care should be taken to refine any outputs to ensure they are free from bias and meeting user needs.  

Ethical and responsible use: Adhere to ethical guidelines and policies to guide the development and deployment of products and services that utilise AI. This includes considering the impact on individual lives, broader society, and the environment.  

Support different interaction preferences: Users will all have different interaction preferences, such as keyboard, switches, voice, and device types (Desktops, Laptops, Tablets, and Smartphones). Design products and services that recognise and adapt to these different preferences. 

Compliance with accessibility standards: Follow established accessibility standards, such as the Web Content Accessibility Guidelines (WCAG) to ensure that products and services are accessible to the widest possible audience.  

We believe this list of principles will evolve over time, but it has already sparked richer conversations within our own design and development teams and has also contributed to some interesting dialogues with clients and public sector teams. 

The journey towards inclusive AI in the public sector is ongoing and requires continuous learning and adaptation. We invite you to join us in this exploration. If you have any thoughts or ideas about how these principles might develop, or how they might be useful to your own teams, we would love to hear from you. Together, we can ensure that AI serves as a tool for inclusion, making public services accessible to all. 

Some useful resources:

Microsoft Inclusive Design Kit -  https://inclusive.microsoft.design/ 

Empowering Accessibility with AI: 5 Real-World Use Cases  - https://www.nexerdigital.com/news-and-thoughts/empowering-accessibility-with-ai 

Generative UI and Outcome-Oriented Design - https://www.nngroup.com/articles/generative-ui 

Designing Generative AI to Work for People with Disabilities - https://hbr.org/2023/08/designing-generative-ai-to-work-for-people-with-disabilities 

Be My Eyes: AI and a New Era for Disability Services - https://www.youtube.com/watch?v=zXa6h7dZDZg