Thirteen years ago, a brutal assault thrust my family into a world not designed for persons with disabilities (PWDs). A transformative moment for both us and a beloved family member forced us to look for affordable and accessible assistive technologies frantically. With the world rapidly beginning to design and harness artificial intelligence for all areas of our professional and personal lives, it is critical we look at how emerging technologies can empower persons with disabilities or risk them creating even more exclusion and marginalization.
Hundreds of civic, business, and disability rights leaders met this week at the 2024 Inclusive Africa Conference organized by InABLE to discuss the risks and opportunities of artificial intelligence for disability inclusion. While the topic has been discussed in Europe and America, it has yet to significantly inform national policymaking in Africa and Kenya.
The development of national artificial intelligence policy under the Artificial Intelligence Taskforce, which was set up by the Information, Communications and Technology Taskforce, and this week’s ICT Ministry National Strategy development workshop offer an opportunity for Kenya to provide best practice law and policy in this area.
Dignity, access, and affirmative action for PWDs are embedded in our constitution (Article 21). Recent legal changes have shifted state obligations from a “special needs” approach to an integrated model that seeks to redesign all public services and spaces and make them more universally accessible and inclusive. This model is in line with progress being achieved elsewhere in the world. Vitu kwaground is different as any person with a disability who tries to cross streets, access e-citizen, attend schools or other services will tell you.
The emergence of new technologies like generative artificial intelligence offers tremendous risks and opportunities for building disability-inclusive physical and online spaces and resources for Kenya’s 4.5 million PWDs and the rest of the population.
While still largely a highly technical conversation, the risks are already showing up for millions across the world. The dangers of algorithms and language models that are disability exclusive and exclude the specific needs of those that are visually, hearing, mentally or physically different from most of society. Uncaptioned images, abusive or disrespectful or inaccurate voice-to-text responses, and the under-representation of the existence of persons with disabilities are some of the challenges inABLE conference participants spoke about this week.
AI is built on datasets. If that data has societal anti- or PWD-neutral biases, artificial intelligence will perpetuate that bias tenfold for us all. Suppose the technology is not designed from a perspective of universal accessibility and is unable to capture speech impediments or persons with certain facial or physical features. In that case, it will lock out and violate the rights of PWDs. AI must also be designed within the principle of privacy by design and the right of us all to be forgotten by erasing our personal data from the internet.
There are also huge and unimaginable opportunities. In the area of assistive technologies, the world can now design better wheelchairs, internet browsers and smarter devices that understand and can simultaneously translate sign language, our facial expressions, for instance. The opportunity for AI-power systems to identify and programme smarter for PWDs and live more independently by controlling their environment has never been this great. New applications are already offering PWDs more effective ways to analyse and adapt to the environment around them, suggest more accessible routes, translate speech to text in real-time or caption videos at a fraction of the cost previously.
Once again, this generation has an opportunity to leap-frog decades of disability exclusion, exploitation, and the violation of rights by designing emerging technologies. We will need to work collaboratively across businesses, public benefit organisations, and governments to do this comprehensively.
As co-panellist Googler Chris Patnoe so aptly put it this week, we all need to lean in but ensure that we “operate at the speed of trust” to ensure that all, and especially persons with disabilities, are able to confidently seize and design the opportunities before us.
Irũngũ Houghton is Amnesty International Kenya’s Executive Director and writes in his personal capacity. Email: [email protected]