Google ‘AI’ and ‘jobs’ and you’ll find no shortage of headlines to make you nervous. So we asked two people who work with artificial intelligence every day whether we have reason to worry.
It’s up to us
‘Whenever you automate anything, at the very least, jobs will change, and that makes it necessary to upskill people. The argument in the robotics space is that you then have to put people’s skills to better use,’ says Doctor Benjamin Rosman, principal researcher in the Mobile Intelligent Autonomous Systems group at the Council for Scientific and Industrial Research (CSIR).
Healthcare is one area where AI (artificial intelligence) is progressing in leaps and bounds, possibly because having the latest technology is essential when it’s a matter of life and death. The great irony is that the smarter we get in our quest to ‘run with the machines’, as an Accenture report on the future of South Africa’s workforce puts it, the smarter will be the machines we build to make our lives easier. But this doesn’t mean that we have to worry about an apocalyptic uprising by terminators, says Rosman.
‘The truth is, we already interact daily with powerful and fast-learning artificial intelligence.’
‘The fear we really have is what we call a value-alignment problem. If you build something that is supercompetent, and you don’t give it detailed enough instructions, it’s going to do something you didn’t intend, because you did not specify properly what you want it to do.’
The bright side, he says, is that robots, if we build and programme them correctly, have the power to make us not only smarter but also happier humans. ‘Everybody has things about their jobs they hate that in many cases are also bad for them. Taking away some of the worst parts of people’s jobs and giving them to robots to do sounds like something we should be aspiring to as a society.’
We already live with AI
The truth is, we already interact daily with powerful and fast-learning artificial intelligence in the form of online chatbots that lead us through financial-service applications, neural network optimisers that unlock our phones by looking at our faces and uncannily accurate algorithms that recommend movies for us to stream on Netflix.
For professor Brian Armstrong, who heads up the Wits Business School/Telkom Chair in Digital Business, the epiphany came while he was checking in at Gatwick Airport in the UK for a family trip. From baggage handling to boarding passes, the system is entirely automated. ‘There isn’t a human being in the loop whatsoever,’ he says. ‘What’s interesting is that it works really well, until something goes wrong.’
In this case, it was a group of travellers who either misunderstood or misused the process, leading to a breakdown and a frantic search for a human supervisor. ‘Human beings are emotional and subjective,’ says Armstrong. ‘Machines, whether hardware or software machines, are not. The problem is that automation often struggles with exceptions.’
Given South Africa’s high level of unemployment and our social priorities, it makes sense for human attendants to do certain jobs – like filling our cars with petrol – and human assistants to smile us through the check-in desk at the airport.
‘At the same time, it can be argued that should we fail to adopt global practices with regard to automation,’ says Armstrong, ‘we will be compromising the competitiveness of our industries. If we compromise the competitiveness of our industries, they will decline and eventually vanish, which will have an even greater impact.’