Accessibility and GenAI
Accessibility means designing things to work for disabled people. The New Zealand Government has legal and ethical obligations to create accessible information and services, both for the public and for public servants.
This guidance is aligned with the following OECD AI principles:
Prevent harm to disabled people
Disabled people can be left out of the design of products and services, and the data used to train GenAI. Without disabled people’s input, GenAI might ignore their needs and suggest irrelevant or incorrect solutions. Privacy issues can also arise, for example, if a person’s disability status is revealed without their consent.
Involving disabled people in the creation of GenAI can help to prevent discrimination and harm.
Enable benefits for disabled people
GenAI can help disabled people participate in society on an equal basis with others.
For example, GenAI can improve communication for people with speech or hearing impairments and make content easier to read for those with learning or reading disabilities. This can increase their independence and reduce their use of other service channels.
To realise these benefits, GenAI systems must be accessible and usable by disabled people.
Accessibility in AI requires a 2-pronged approach
1. Include disabled people in all stages of the AI lifecycle
- Engage disabled people as stakeholders from conception of the GenAI solution to its deployment.
- Develop a vision statement and policy that prioritise inclusion and accessibility.
A Blueprint for Equity and Inclusion in Artificial Intelligence — World Economic Forum
AI & Disability Inclusion Toolkit Overview — The Partnership on Employment & Accessible Technology
2. Ensure equitable treatment of disabled people by GenAI
- Consider disabled people’s experiences in training data, models and outputs to remove bias.
- Build GenAI to mandated accessibility standards and test it with disabled people.
- Throughout all stages of development and use, continuously monitor that GenAI is delivering safe, trustworthy and non-discriminatory interactions.
Accessible and Equitable Artificial Intelligence Systems: Technical Guide — Government of Canada
AI Accessibility Scenario
Related guidance
- Governance
- Transparency and Explainability
- Accountability and Responsibility
- Privacy
- Bias, Discrimination and Fairness
- Skills and Capability
Utility links and page information
Last updated