This article is part one in a two part series.
As the world continues to become increasingly intertwined with technology, artificial intelligence (AI) presents some of the greatest opportunities and challenges for us to face. For the nearly 30% of people living with disabilities, those impacts will be felt much more acutely. And while many debate the efficacy and ethics of AI (for good reason), it is accessibility that's far too often left out.
Image description: A person wearing jeans and a white tshirt sits on stairs typing on a laptop. Their left hand is prosthetic.
Accessibility is the process of designing products, services, and spaces that can be used and enjoyed by everyone, regardless of ability. It involves actively lifting barriers through innovative solutions tailored to meet the unique needs of every user.
Artificial Intelligence has the power to revolutionize the future of accessibility and disability inclusion. But, it requires measured awareness and willingness to face challenges head-on.
Endless Potential
In many ways, AI already impacts disabled people. There are several technologies dedicated to making life easier, including communication tools for people living with vision, hearing, and speech limitations, and mobility barriers.
Voiceitt, an accessible speech recognition company, uses AI and machine learning to recognize and adapt to non-standard speech patterns. The technology works by using input speech data to learn from individuals' unique speech patterns and translate it into a form that's clear and easily understood. It's a game-changer for people with Cerebral Palsy, Parkinson's disease, and other conditions that can make clear speech difficult.
For the nearly 2.2 billion people with vision-related disabilities, AI-powered imaging tools now can assist by converting visual data into other formats. There are several options on the market, giving users an opportunity to choose what best meets their needs. OCR.best and Image2TxT, for instance, are designed to convert visual cues into text and audio-based alternatives. Models like ChatGPT-4 and Claude 2 now have plugins able to decipher complex image data And several AI tools have the ability to alter image contrast and resolution in real time.
AI is also helping disabled people regain independence. Smart home technology makes everyday home tasks possible through simple voice commands. AI-powered wheelchairs enabled with audio cues and autonomous movement capabilities allow users to navigate the world with less support. AI is even being used in prosthetics to enhance finer movement capabilities and user autonomy.
And it's just the beginning. If humans harness the full power of AI, we have the potential to revolutionize daily life for millions of people for the better. Artificial Intelligence is not without its challenges, however. Especially in terms of accessibility.
Challenges of AI Technologies
Cost
Artificial Intelligence technology is expensive, and is likely to remain so for the foreseeable future. Beyond that, the costs associated with personalizing tools for specific barriers and limitations are steep. It also brings with it additional time and resources which are often left to consumers to shoulder through fees. Disabled people often face more expenses than nondisabled people just to meet basic needs. That added to the fact that the vast majority of disabled people already live in poverty, access to many of the targeted AI tools is incredibly limited.
Knowledge Gaps
There may be a growing number of product and web developers committed to positively impacting disabled lives. Unfortunately, very few have deep, direct understanding of accessibility best practices. In fact, only 2% of digital products meet basic digital accessibility requirements. And merely 4% of businesses have any initiative directly related to disability at all. And training is limited.
For the field of AI to adequately benefit the lives of disabled people, it requires teams of highly experienced designers and developers with the necessary accessibility skills to meet a variety of unique needs. Currently, there aren't enough of these professionals to support the ambitious goals of AI for disabled users.
Bias
Related to poor understanding of accessibility are deep-seated misconceptions and biases inherent to AI's current development path. In the short time since Artificial Intelligence exploded on the market, multiple inquiries have found discriminatory biases related to its use. Facial recognition software has been the focus of widespread criticism after it was found to be racially biased against dark-skinned people. And a study at NYU found that AI-powered language learning associated disability with negativity and violence, among other ableist stereotypes.
The origin of AI-based bias is multilayered. Systemic influences, human perceptions, and algorithmic processes work together to create harmful impacts on everyday people. AI is a product of the people and systems that direct it. And unless we properly address the way the influence AI, accessible and inclusive uses will be impossible.
The Way Forward
Artificial Intelligence is here to stay. And with it a number of benefits and challenges. These challenges don't have to spell the end of Artificial Intelligence ability to benefit consumers of all abilities though. With deliberate and mindful action, the future of AI can be a profound tool for good.
In part two, I'll cover the future vision of Artificial Intelligence in the lives of disabled people. I'll also discuss potential solutions to the challenges facing AI and provide recommendations for the continued advancement of accessibility as AI grows.