The benefits of predictive analytics based on Artificial Intelligence and machine learning is that it makes the rendering of social services to vulnerable people so much more effective. Photo: African News Agency (ANA)

CAPE TOWN – Depending on the particular research consulted, it is calculated that around 50% of all current jobs will be automated in the coming years as a result of the Fourth Industrial Revolution (4IR). The implication of this is that a broad range of jobs will become obsolete.

The disappearance of certain well-known jobs will initially be in fields requiring a low level of skills, but will thereafter start affecting higher-skilled and even professional people. Any worker, who is paid a high salary, just because they have merely learned known knowledge or the code of decision-making, is in danger. Wherever any decision-making formula is logically known, Artificial Intelligence (AI) will soon be able to surpass human beings in performing decision-making responsibilities. 

Interestingly enough, it seems that even a high “human touch” field such as social work is not left unaltered by the 4IR. Prof Eric Rice, a social work professor and co-founder of the University of California Center for AI in Society (CAIS), and engineering professor Milind Tambe developed predictive models for public health interventions using AI. 

Typically, they use algorithms to analyse social networks of homeless youth who have registered with them based on the principles of social networking theory. The aim is to replace social workers’ subjectivity of who could take leadership and become influencers. The algorithm thus selected a group of influencers or peer educators to spread information about HIV to their peers.

Many of these young homeless persons does not struck one as an exemplary leader, such as Jacob, a twenty year old skateboarder that was usually high when he visited the centre. Despite Rice’s misgivings about Jacob, the algorithm picked him as a peer educator. Throughout the project Jacob proved to be a crucial connector to the homeless youth.

The success of the system is that machine learning capability of the AI has the ability to bypass human assumptions and biases. The algorithm selection proved to be a huge success since the AI assisted group has much higher HIV testing rates and a higher use of condoms.

Due the staggering homeless figures in Los Angeles, Rice’s team is working with the LA Homeless Services Authority to deploy AI-infused vulnerability assessments that match the homeless people with the best-suited housing and social interventions. This not only expedites the process of placement, but also alleviates the task of the grossly overloaded social workers. The housing algorithm is akin to the Uber algorithm that successfully matches Uber drivers with ride-seekers while optimising the route and travelling time.

The economic situation in many countries, including South Africa, has put homelessness in the public eye as people increasingly started to spread to open spaces, unused land or buildings and highway underpasses.  Except for a lack of political will and resources to address the problem, we still use out-dated housing assignment systems that are prone to manipulation and most often lead to unhappiness and dissatisfaction with the allocation of housing.

Perhaps it is time that South Africa follows the example of professor Eric Rice of Los Angeles by using computer science together with in our current efforts to combat this complex problem. It may indeed assist social workers to solve the growing problem of homelessness that they were unable to solve by themselves due to limited resources.

In the UK some local councils are experimenting with AI systems to select vulnerable families and children in need of intervention, thereby using machines to fulfil some of the functions previously done by social workers. This typically entails predictive analytics, which is based on the analysis of large quantities of data to detect certain trends and patterns in the data. These trends are then used to identify more families who might be at risk.

What makes predictive analytics so powerful is the capability of machine learning or the ability of the intelligent computer to learn by itself from every new case. AI therefore makes rendering of social services to vulnerable people so much more effective by understanding exactly which services or interventions are successful, thus enabling local government to redesign services to focus on the most impactful and cost effective. 

AI and predictive analytics further assists with determining which children and families are the most likely to need help, enabling social workers to spend their limited time in the most effective way. But perhaps the greatest value of predictive analytics in social work is that it helps to detect certain needs at a much earlier stage. 

The use of AI in social work situations is growing and is making a difference in society. But will intelligent machines replace social workers? AI could surely help caring professionals with the huge overload of work, but it is doubtful if AI would totally replace them in the immediate future. 

However, the reality is that intelligent machines are already better and much faster than human beings in their judgement of families that need help. Machines have a much better capability of making sense from a confusing mass of information than human beings and therefore may in the future replace social workers.

Professor Louis C H Fourie is a futurist and technology strategist. [email protected]

BUSINESS REPORT