Setting the standards for accessible and equitable AI

A GROUNDBREAKING development in Canada offers important guidance for Malaysia as it moves deeper into digital transformation and artificial intelligence innovation.

Canada recently released the world’s first standard on accessible and equitable AI. Called “Accessible and Equitable Artificial Intelligence Systems”, it was designed to help ensure that AI systems are accessible, inclusive and fair from the start, and was released on Dec 3 to coincide with International Day of Persons with Disabilities (PwDs).

Although written for the Canadian context, its principles are directly relevant to Malaysia’s digital ambitions.

The standard emphasises that AI must never become a barrier. It states clearly that individuals must always have the option to interact with a human officer, especially when digital tools are confusing or inaccessible.

This is particularly important in Malaysia where many elderly citizens struggle with automated systems, and low-income users may lack smartphones or stable Internet connection.

The standard also insists that persons with disabilities must be meaningfully involved throughout the AI lifecycle – from design to testing. Without their participation, AI systems risk misinterpreting real-world needs, unintentionally disadvantaging the very communities meant to benefit from digital innovation.

For Malaysia where the majority of websites and apps still fall short of accessibility standards, this is a reminder that accessibility cannot be added as an afterthought; it must be integrated from the start.

The Canadian framework further requires AI systems to be tested rigorously to ensure fairness. This matters significantly for Malaysia because AI is increasingly used in decisions involving credit scoring, insurance claims, medical screening and even job recruitment.

If AI misinterprets the characteristics of PwDs, senior citizens or minority groups, their opportunities can be unfairly restricted.

The standard also stresses continuous monitoring. When AI systems repeatedly make errors affecting vulnerable groups, organisations must correct the issue rather than expect consumers to adapt or endure the inconvenience.

This approach aligns with Malaysia’s broader goals under the National AI Roadmap and MyDIGITAL (Malaysia Digital Economy Blueprint), but adds the crucial dimension of accountability.

If Malaysia does not take proactive steps, older adults may lose access as physical counters disappear and services shift entirely online. PwDs may face new barriers in education, employment or transport because digital systems were not designed with them in mind.

Rural communities may fall further behind as more essential services become AI-driven. Consumers with limited literacy may find themselves unable to access support or resolve disputes.

To prevent this, Malaysia can integrate accessibility and equity into its national AI strategy, including procurement requirements for government agencies.

Ensuring that all AI systems offer a human-help option will protect millions of Malaysians who may struggle with digital platforms. Establishing an AI harm reporting mechanism could help identify patterns of discrimination.

Most importantly, PwDs, senior citizens and vulnerable communities must be meaningfully included in shaping future AI policies.

AI has the power to improve healthcare, strengthen education, enhance safety and drive Malaysia’s economic growth. But these benefits will only be realised if the technology is designed for everyone. Accessibility is a right, fairness is a responsibility, and human choice remains essential in a digital society.

SARAL JAMES MANIAM

Secretary-general

Malaysian Association of Standards Users