SA needs to prepare as rise of the machines will result in fall in jobs
• Some experts are arguing that not enough is being done to prepare the country for the disruption the fourth industrial revolution will inevitably bring
Entrepreneur Ryan Falkenberg should be smiling. His artificial intelligence (AI) platform has helped a local bank to increase sales of financial products 52% and an oil company’s call centre to resolve nearly all suppliers’ technical issues.
But Falkenberg, who runs Cape-based AI developers Clevva, is worried. “I don’t think there’s an acknowledgement of the speed at which human redundancy is going to happen.”
Much of AI involves machines learning how to respond to certain things without being explicitly programmed to do so.
Falkenberg says the full effect of the technology will probably only begin to show in the next generation after it becomes more integrated with big data and the Internet of Things.
At the bank, his AI platform helps guide sales consultants through its list of hundreds of financial products. Consultants barely need have any knowledge of the products.
Falkenberg says this doesn’t mean consultants are no longer needed, but turns them into super agents able to focus more on client relationships.
However, he says jobs involving predictable tasks and which add little value will soon vanish. This could result in the loss of as many as two-thirds of jobs in SA, a study in 2016 by Citi and the Oxford Martin School at the University of Oxford estimated.
University of Johannesburg deputy vice-chancellor of research Tshilidzi Marwala believes SA is ill-prepared to tackle the challenge of AI.
He says not enough is being spent on studying the effects of the technology. The National Research Foundation (NRF) should, for example, be incentivising researchers to carry out more research in the field.
NRF deputy CE Dorsamy Pillay confirms that the institution has no specific incentives for AI researchers. This is despite the ICT RDI Roadmap, adopted by the Cabinet in 2013, identifying the need to increase investment in AI through a collaborative ecosystem approach.
“I don’t see enough hunger for AI as I see elsewhere,” says Marwala, who has written 10 books on the subject. In 2016, one of his books was translated into Mandarin.
He says SA needs a plan on how to get companies to adopt AI and tackle the effect it will have on society. Pillay says the Chinese government announced earlier in 2017 that it would develop an AI strategy and the White House released a report in December 2016 on the effect AI will have on society.
China produces about 40% of research papers on AI and is investing in the field in a bid to find solutions that tackle the country’s rising labour costs.
Marwala says AI is increasingly being used to make already automated processes “more intelligent” by, for example, being able to correct faults on the factory floor.
While jobs that are taskbased will increasingly be automated, he believes there will still be a demand for jobs that require human interaction — while medical diagnoses can be performed by machines, a person will be required to brief a patient on the results.
Martin Sanne, Council for Scientific and Industrial Research (CSIR) executive director of materials science and manufacturing, says the efficiencies introduced by AI and robotics will make it more costeffective for multinational companies to manufacture in SA.
AI will also allow for production to be better managed and for the maintenance of machines to be more efficient. Intelligent machines will also allow artisans to work on many different products, improving their skills.
Sanne believes SA is not behind other countries in preparing for the deployment of AI and other new technologies. “We understand the trends, but we need to do more,” he says.
Centre for Artificial Intelligence research chairman Tommie Meyer believes the government appreciates the benefits of AI, but is not holding enough discussions on how it will need to reskill employees whose jobs are made redundant.
“In the next 10 to 15 years, there is the potential for a major disruption in jobs,” he says, adding that the state will also need to provide financial aid to people who lose their jobs.
The centre, based at the CSIR with five universities involved, has 30 researchers examining areas such as machine learning, speech recognition, natural language understanding and knowledge representation (getting machines to reason as humans do).
Meyer says this won’t be the first time the world is shaken by change of this kind — the industrial revolution and emergence of computers presented similar challenges. But now there is an opportunity to be prepared for the changes.
He believes the only way to deal with the uncertainty around AI is to get rid of the perception that the machines will develop powers similar to those seen in a Terminator movie and destroy the world.
But Centre for Cyber Security director Basie von Solms believes anything is possible — including machines that turn on humans.
“The more you get those machines to learn by themselves, the less control you have over what they are doing,” he says.
Von Solms says recent cyber attacks involved criminals who used devices such as cameras connected to the internet as bots to launch co-ordinated attacks. Criminals might soon be able to bring down stock exchanges, as many use AI in brokering trades, or attack critical infrastructure such as power grids.
To fend off these increasing risks, Solms says the government and business will have to craft regulations to specify what limits to place on AI and machine learning.
He points to the Asilomar principles signed in January by 2,000 leading minds including physicist Stephen Hawking and entrepreneur Elon Musk and that deal with various ethical issues around AI.
But for now, the biggest concern over AI is likely to remain the effect it will have on jobs.
Yet Johan Fourie, associate professor in economics at Stellenbosch University, believes while AI will make some types of jobs redundant, it won’t lead to the mass unemployment many envision.
He says after the adoption of ATMs in the 1980s, many feared bank tellers would lose their jobs. But employment rose as ATMs freed up resources to open more branches.
“This is because humans are innovative and creative and will find new things to do with these technologies,” Fourie says.