AI has the potential to deliver real value in business, creating unprecedented efficiencies across countless processes. But is South Africa’s industry ready?
In the 1964 children’s novel by British author Roald Dahl Charlie and the Chocolate Factory, Charlie’s dad, Mr Bucket, loses his job when the factory where he works is mechanised. Once responsible for screwing the lids onto tubes of toothpaste, he gets laid off because a robot is able to perform his job more cheaply and efficiently.
This trope of man being replaced by machine often comes up in discussions about artificial intelligence (AI) and its potential impact on our world and workforce. But the fear around how emerging technologies will challenge our existing culture and change our lives fails to acknowledge the many benefits that accompany the growth of innovations such as AI. And these gains span all industries, not only computing and IT.
In 2021, Forrester Research predicted that more and more companies will implement AI for no other reason than because they have to do so if they want to compete in an AI-driven future. 2020 gave business leaders the impetus, born out of necessity, to embrace technologies that make them future-proof and digital-ready, the research firm notes, predicting that the AI software market will grow to $37-billion globally by 2025.
“I would go as far as to say that AI is at the heart of Industry 4.0 and/or the fourth industrial revolution,” says Spiwe Chireka, telecoms, media and technology industry expert. “It is one of those technological advancements that adds the ‘oomph’ to other technology advancements.”
New technology, new ways of working
AI has expanded the bounds of possibility, especially when it comes to data, explains Ndabenhle Ngulube, one of the cofounders of Pineapple, a local insurtech start-up: “Once upon a time, something like time-to-purchase data was seen as futile and meaningless. But today, this kind of data is the de facto input for thousands of AI models around the world.” Why? Because some customers will make a purchase on their first visit to your website, while others will only purchase a product after their second, third or fourth visit. If you understand how long it takes customers to buy something, you have a better understanding of their behaviour and you’re better equipped to tailor your marketing campaigns to offer a more personalised experience.
The insurance industry has historically been guilty of sitting on massive amounts of data, with no application for it. Not only is artificial intelligence changing this narrative, it’s also making the process of taking out insurance faster and easier, Ngulube says. For example, the Pineapple app allows a user to get cover by simply snapping an image of the item they want to insure. The app’s computer vision AI recognises what the image is and then places that image in an appropriate category for pricing purposes before providing the customer with an estimated premium.
Today, every business owner needs to understand how AI can be leveraged across their operations and how it is relevant to their progress and sustainability, says Volente Morais, executive manager at Koogan Plastics, a South African plastics and packaging manufacturer. What do these applications and technologies look like for a business like mine, where production is heavily reliant on people, she asks? Well, they’ve incorporated a smart robot into their labelling assembly line. This AI timed system is trained to respond to the exact location of the product so that it can mould labels directly onto the product instead of sticking on an adhesive label. “The value gained here is efficiency, process improvement and reduction in waste costs. And our increased output ultimately improves our revenue,” she states, noting that innovations like this are game changers in sectors where everyone is selling the same product. But with the introduction of new technologies comes an evolution of the workforce and growth in the sizable skills gaps that already exist.
According to Professor Duncan Coulter, deputy head of the Academy of Computer Science and Software Engineering at The University of Johannesburg (UJ), as intelligent systems become more ubiquitous, there is a great need to educate non-specialists in both the use and limitations of these emerging technologies. For those looking to up their AI expertise, UJ has added an AI specialisation option to their existing BSc honours in Computer Science and Informatics. They have also introduced an undergraduate BSc (Computer Science) with an AI specialisation, which includes an AI-focused project module in the final year. In addition to this, the Academy of Computer Science and Software Engineering offers a short learning programme focused on the creation of intelligent systems using modern techniques and technologies.
“If you talk about artificial intelligence as the pursuit of creating genuine, generalised intelligence, we are quite far from realising that goal. If you consider artificial intelligence as being our capacity to embed pre-existing intelligent behaviours into artefacts, then we have made considerable progress,” concludes Coulter. “But even fairly simple tools can be used for great good and for great harm; just as a hammer could serve as a tool for construction or a weapon of destruction. For this reason, there is increasing work that needs to be done to maximise the gains that can be made in terms of the tasks these systems are good at, while minimising the harm that can be done through these very same tasks.” — Joanne Carew
AI and legal liability
Back in 2018, a self-driving Volvo knocked down and killed a pedestrian in Arizona. This was the first pedestrian fatality involving an autonomous car. Who was responsible? AI law refers to the application of the law on AI. Like any branch of law, it has to evolve in order to deal with new challenges and new situations such as this one, notes Patrick Bracher, a director at Norton Rose Fulbright. The increased use of AI demands that existing legal principles be adapted to account for new ways of life and new ways of doing business, and modern precedents will have to be established to determine who is responsible should death, injury or damage be caused by the failure of an AI system.
AI gone wrong
Part science, part art, we need to remember that Artificial Intelligence is only as smart as the data it’s fed, which is why AI systems sometimes get things wrong. Back in 2016, a Microsoft Twitter chatbot spewed out misogynistic and racist remarks after learning these sentiments from users. Around the same time, an Amazon recruitment tool came under fire for disadvantaging female candidates because it was based on historical hiring decisions, which traditionally favoured men over women. This is why diversity is so important — diversity in the people developing the AI systems and in the data being used to power them.