Start main page content

Can robots decide on right and wrong?

- Gillian Anstey

The '4IR: Philosophical, Ethical, Legal Dimensions' conference explored the Fourth Industrial Revolution (4IR).

4IR is a buzzword frequently devoid of meaning but the conference, presented by the Wits Institute of Data Science and the National e-Science Postgraduate Teaching and Training Platform, provided a platform to delve deeper into the implications of 4IR technologies. 

Do machines have moral agency, that is, can they make decisions based on right and wrong? Yes, in a functional sense they do, argued Fabio Tollon of Stellenbosch University’s Department of Philosophy. 

Should hackers who break into computer systems of their own accord and expose weaknesses be paid for doing so? No, argued Yonnique Goliath, a strategist at Youth Employment Service completing her Masters degree on the subject at Wits. 

Brazil does not have proper legislation to deal with the fake news spread during President Jair Bolsonaro’s 2018 electoral campaign, including political WhatsApp messages masterminded in Spain, lamented Adjunct Professor Nathalie Cadena of Brazil’s Federal University of Juiz de Fora. 

Those were just some of the topics explored by 16 speakers from a total of 32 delegates who attended the 4IR: Philosophical, Ethical, Legal Dimensions conference at the TW Kambule Mathematical Sciences Building on West Campus from 3 – 5 September.

A special 2020 issue of the Springer journal, Philosophy & Technology, will feature selected papers from the conference. It will be guest edited by conference organiser, Helen Robertson of the Wits School of Computer Science and Applied Mathematics, who also presented a paper. 

Professor Rod Alence, Head of the Department of International Relations at Wits, set the tone for how the outlandish becomes commonplace in industrial revolutions when he welcomed the delegates by speaking about watching Terminator 2 in Accra, Ghana in the early 1990s. The film features a cyborg who turns into molten metal but the part that drove the audience wild was when a young boy seemingly walks up to a wall and withdraws cash  - they had never seen an ATM before. 

The movies RoboCop, Dr Strangelove, The Island, Minority Report and Surrogates also popped up during the conference; in the world of 4IR what was sci-fi then, is closer to reality today. 

As Karabo Maiyane, a lecturer at the Department of Philosophy at the University of Pretoria, said in his paper on lethal autonomous weapons: “As far as we know they have not been deployed  -  not that they have not been created". 

Brent Mittelstadt, a Research Fellow and British Academy Postdoctoral Fellow in data ethics at the Oxford Internet Institute at Oxford University, and a member of the UK National Statistician’s Data Ethics Advisory Committee, presented the keynote address via a video link-up. 

Titled: Data Ethics: From Principles to Practice, it showed how data has changed professions such as medicine. Mittelstadt said when he sees a doctor he expects the doctor to be "doing data entry and looking at the screen for most of the time’’. 

He compared developing artificial intelligence with medicine as it is ‘’the most prominent example we have of attempting to go from high level principles down to practical recommendations’’. 

But "will a principled approach to ethics actually work in the context of AI?" He believes not. 

He said medicine promotes health and wellbeing which is what patients also want "so there is this fundamental alignment of common aims" which encourages trust. 

"I don’t think we can say the same for AI,’’ he said, because it is often developed by institutions whose responsibilities are to shareholders, not users. "There is no real equivalent of the ‘patient’ in AI,’’ he said, so ‘’ethical decision-making in AI will often be adversarial rather than cooperative’’. 

He presented a list of global initiatives addressing AI ethics – and scrolled down one screen image after another to reveal a total of 84, as at June 2019. 

”Many countries are producing a national strategy because they see AI both as an opportunity and a risk," Mittelstadt said. 

John Ostrowick, a director at SA’s Department of Justice and Constitutional Development but who spoke in his personal capacity, said the South African government is very aware of 4IR.  

It has established an advisory commission which will create a report on what to do about 4IR and most departments are ‘’working on plans’’. 

He produced stats from international strategy consulting firm Roland Berger which showed the likes of China, Britain and Germany each spending about 2b-euros on 4IR. In contrast, SA has spent only a few million rand, he said. 

"In other words we haven’t exactly taken it seriously enough," said Ostrowick who feels the state is not showing awareness of 4IR threats such as bioengineering.