Accessibility considerations for AI copilots
AI copilots are systems that provide guidance, feedback, or assistance to users in completing tasks. Such as, writing, coding, or designing. They're useful tools that can enhance the productivity and creativity of users, especially people with disabilities who might experience accessibility barriers.
For example, people with dyslexia might struggle with reading long documents. By prompting Microsoft Copilot to use generative AI to summarize information in a lengthy Word document, individuals with dyslexia can enhance their comprehension, accuracy, and productivity.
AI copilots must be designed and developed in ways that respect the diversity and preferences of users and that ensure the system's accessibility, usability, and quality.
Accessibility considerations for AI copilots include:
Keyboard interaction: AI copilots should allow users to navigate and interact with the system by using the keyboard and follow a logical and consistent order of focus. This approach can benefit users who have difficulties using the mouse or other pointing devices, such as people with mobility disabilities or low vision.
Keyboard interaction should support common keyboard shortcuts and commands. It must also provide visual indicators of the current focus and selection.
Prompt suggestions: AI copilots should provide simple and clear prompt suggestions that a global audience can understand. Prompt suggestions should avoid complex or specialized terms. This approach can benefit users with limited literacy or language skills, such as people with cognitive disabilities or non-native speakers.
Prompt suggestions should also be relevant, appropriate, and helpful for a user's task and context. They should avoid misleading, confusing, or offensive content.
Sounds that convey important information: AI copilots can provide sounds to indicate the availability, completion, or error of a task. Sounds can also convey the presence of new or updated prompts. These sounds should be distinct, recognizable, and consistent. They should provide options to adjust the volume, frequency, and duration.
Cognitive overload: AI copilots should avoid overwhelming users with too much information or too many options. They need to provide a user-friendly and intuitive interface. This approach can benefit people who have difficulties processing or remembering large amounts of information.
Developers can reduce cognitive overload and help enable a consistent experience across all applications by following these best practices:
- Provide clear and concise instructions.
- Group and organize related items.
- Allow people to control the pace and sequence of the interaction.
Multiple types of input and output: AI has the potential to allow people to interact with the system through various modes of input and output, such as voice or keyboard input, to increase usability and accessibility. This approach can benefit people with speech disabilities or mobility disabilities who have preferences or limitations in using certain modalities. Multiple types of input and output can also enhance the flexibility and adaptability of the system and enable natural, conversational interaction.
How do AI copilot innovations affect users?
This video describes how copilot innovations in Windows and Microsoft 365 are opening new AI-powered accessibility experiences for people with disabilities.