MENA 4.0: People with Disabilities

Digital technologies, due to its capacity of molding content depending on the data inputs, can play a catalytic role in designing systematic alternatives to include marginalized societal groups, such as people that face physical impairments to attend a job training course. People-centered technologies are springing up, bringing forth novel frontiers expected to embed appropriate and sustainable digital solutions that maximize training and the development of both the private sector and the financial system, thus representing an opportunity to enhance levels of skilled labor and diversify professional networks.

However, if technological developments enabling transformations do not reach these individuals, supposedly innovative business models, planning measures, strategies, and policies, will be incomplete; society would be merely enduring systematic inequalities and explicitly missing the opportunity of reaping the benefits of creating an entire web of diversified personnel.

The branch of Artificial Intelligence is currently enabling myriad applications, bridging the access of often excluded individuals to participate actively in society. Accordingly, the AI Mentorship Software is a solution expanding educational inputs through resources capable of adapting according to the users' needs. This AI-based software contains algorithms that learn from the users' patterns, allowing the system to become extremely personalized to the point of being able to predict future learning struggles while guiding learning experiences to avoid users missing out on any content.

Currently, a prominent use case is the QTrobot developed by LuxAI, an expressive social robot that offers educational inputs that encourage users with special needs to engage in learning environments. This technological tool provides a user-friendly system, capable of retaining the attention of people with learning disorders, adapting the content in real-time depending on users' facial expressions and body gestures.

Accordingly, QTrobot could also support high-risk areas where professors and trainers often face considerable hurdles accessing training institutions physically, especially in rural and risk areas. This technology could give the possibility for kids, youth and adults alike, or even the elderly, and people with limited mobility, to approach educational systems and autonomously accompany the different styles of education provided worldwide.

On Technical and Vocational Education Training (TVET), QTrobot could get specialized in teaching vocational content, such as executing practical tasks systematically, as well as assisting in career guidance, enabling users to improve social interaction skills for employment opportunities based on their current social behavior. This could greatly contribute to business development, as ideas would be first simulated with robots and later taken to real world situations after proper assessment and tests. In financial system development, this technology offers a wealth of additional options to people with disabilities, including training while using digital banking tools, but also providing appropriate instruction in handling their finances independently. A considerable challenge is that the technology remains costly, and access to the tool would still be reserved to a few selected target groups.

Another solution is the Mtabe App, a conversational AI tool that provides e-learning inputs for learners without access to the Internet, textbooks, or smartphones. This solution can suggest and develop customized individual learning programs, according to the specific needs of each user. By supporting individuals in social and educational contexts, as well as personal interests, this AI-powered program would run like a cross-media chatbot featuring both conversational and curated content adapted according to training methods embedded into its system.

This tool could also provide individualized answers for theoretical TVET questions and assist in daily work routines by sharing specific text guidance in work situations, thus working as a personal assistant. In business development, the tool could train people with disabilities in entrepreneurship theory, but also support real life issues, such as handling banking independently, which in turn could empower them to further participate in the labor market and assist their entry in financial systems.

On an institutional level, AI Mentorship Software could help evaluate students as well as the institution itself. These analyses would occur based on the quality of their curriculum and training materials while taking advantage of the potential for artificial intelligence to create unique learning pathways for individual lessons in massive open online courses (MOOCs), combined through blended and online learning.

Beyond training, it could function as personal counselors for people with learning difficulties or special needs who are searching for jobs and those who desire to improve their skills. By cross-referencing data from users, it could deeply understand each individual's best skills, connect them with more suitable companies, and provide opportunities for employment and courses all across the globe.

Impact on People with Disabilities

According to participants, this technology could help people with disabilities to participate more actively in the public sphere. By providing individualized learning content and considering different inputs depending on the learner’s capacity skills, AI Mentorship Software could enable greater learning outcomes in all areas of life. In terms of sector change, this technology could help drive people with disabilities to have more independence in the job, creating bank accounts and handling their finances, as well as starting their businesses on their own. On the same note, Technical and Vocational Education and Training (TVET) could be benefited too, as data harnessed from users would let trainers and employers understand better the needs of people with disabilities, thus allocating them to the roles they would most suitably develop.

However, even if this technology seems to deem very positive results, participants shared some concerns regarding its application. On the one hand, it would allow users to feel more autonomous and open up greater opportunities in both the inclusion of people with disabilities in financial systems as well as in business development, but, on the other, it may take away agency from its users, possibly increasing dependency on the tool and/or on the people who control these tools.

On a more nebulous note, dehumanizing training, education, and capacity building through the replacement of human agents with AI software could possibly expose people with disabilities to a filter bubble, thus limiting these users to accessing other types of inputs, for example, peer-to-peer relationships and human affection. Also, if learning content remains generalized and not addressed to individual needs —especially people with disabilities that have more specific demands to be considered, which oftentimes cannot be covered by standard learning content and procedures— individualized learning could continue to be costly because it would still require the presence of teachers and trainers.

We all know at this point that emerging technologies have inherent dangers, and mapping them is as important as detecting the opportunities within the application of these technological tools. The table below displays the participants' rating related to some potential risks when applying AI Mentorship Software into the lives of people with disabilities.

Similar to the risks associated with Micro-learning Platforms, participants raised concerns regarding the possible negative impacts produced by the misuse of the data contained in AI Mentorship Software. In order to circumvent biases, discrimination, unauthorized surveillance, or attacks to factual information, it is necessary to enable coordination and knowledge transfer about the risks of data sharing between the actors employing and using this technology. Regulatory bodies and law enforcement would also play a pivotal role in defining usability standards. This could possibly act as a mitigation measure against offensive actions that may affect the integrity of the societal groups making use of the technology.

Future Perspectives

One downside raised by participants is that humans are prone to biases, even unknowingly. As AI tools infiltrate society, we have realized that algorithms are not always perfect: algorithmic bias has already been detected many times. Although AI tools are made of a continuous form of statistical discrimination by its very nature, the kind of bias primarily addressed is the ‘minority’ variety (for example, words and sentences that have appeared fewer times in data sets), thus placing dominant groups at a systematic advantage and minority groups at a systematic disadvantage.

Participants suggested that a possible solution to this matter was to make sure the data used in AI systems come from reliable sources, as well as to set continuous monitoring to rigorously check the AI Mentorship Software for algorithmic preferences or biases. In addition, as the repositories of AI tools gather immense amounts of sensitive data they could be susceptible to attacks. Together with an AI Mentorship Software, robust defense systems should be used to ensure security and privacy of the data stored.

Also, participants questioned about the ownership over collected data and how said data could be used. One solution to this matter is to work with and consult people with disabilities before promoting the tool. By doing this, target groups would understand better how digital tools work and consequently participate in the implementation of technological tools. Likewise, it is imperative to run a complete assessment of the use cases implementing this tool (e.g., analysis of each individual or target group) to decide on when and where this technology makes sense precisely.