6 minute read
What South Africans think about AI, robotics & automation
By Daniel Mpala
THE SIENNA (Stakeholder-informed ethics for new technologies with high socio-economic and human rights impact) project seeks to develop ethical protocols and codes for human genomics, human enhancement, as well as AI and robotics.
The report, which was published in October and can be accessed here, is a quantitative investigation of public views and awareness of artificial intelligence (AI), robotics, human enhancement and genomics.It is one of seven studies undertaken through the project which seeks to study the ethical, legal and social issues (ELSI) of the three technologies. The report’s aim was to provide a snapshot of opinions in 2019 on intelligent machines and their impact on society.
The study was carried out by research firm Kantar which surveyed 11 000 respondents — 1000 each from France, Germany, Greece, Netherlands, Poland Spain, Sweden, Brazil, South Africa, South Korea, and the US — through telephonic interviews between March and April 2019. Brazil, South Africa, South Korea and the USA were included in the study to provide comparative insights.
The report’s authors pointed out that relative to the other 10 countries in the study, South Africa stood out as a country with very polarised views on many measures. The authors attributed this to “extremely wide inequalities in society”.
“There were several measures where South Africans were both among the most positive and the most negative. This is evident in the proportions of South Africans who: support vs oppose the idea of robots resembling humans in everyday life and as romantic partners; feel that intelligent machines will lead to more autonomy vs less autonomy; and feel that their country will be completely different vs not different at all. It is also notable that in all countries apart from South Africa, a majority considered that more widespread automation would lead to greater inequalities. In South Africa, a sizeable minority (21%, higher than all other countries) thought it would instead lead to reduced inequalities within their county,” the report stated.
Need for education on AI, robots
The findings from the report seem to suggest that there’s need for education across the board on robots and AI, and the capabilities of these technologies and their impact on society. The report’s authors state that although most of the survey participants across the 11 countries were familiar with AI and robotic applications, “relatively small proportions” felt they were well informed.
Across most of the 11 countries, the report found that respondents not only anticipated wide-scale changes over the next two decades in the development of capabilities of intelligent machines and their effects of societies, but also understood that the technologies brought risks.
In the majority of countries, the study found that people were “more negative than positive” about the prospect of robots taking on more human features, and the widespread use of intelligent machines which the respondents felt would widen inequalities and result in people having less control.
“Despite this, people were on balance more positive than negative about the overall impact of intelligent machines in society. There were wide variations by country, with South Korea and Brazil most positive, and South Africa most polarised. No single country stood as especially negative, though France, Greece, Spain and Germany were all more negative than average on two or more measures,” the report’s authors added.
On average, across all the countries in the study, half of the people surveyed said they’d heard or read “a lot” or “ a fair amount” about robots (51%) and AI (50%). The highest proportion of respondents who’d heard or read “ a fair amount” about either robotic or AI technologies was highest in Germany and South Korea ( range between 66% and 72%) and lowest in South Africa, Spain and Poland ( range between 32% and 41%).
“South Africa stood apart from all other countries in terms of the proportion who lacked any awareness of these technologies. Based on the all-country average, almost all respondents had at least heard of robots (96%) and AI (93%). However, a substantial minority of respondents in South Africa had never heard of these technologies at all (24% had never heard of robots, this while 23% had never heard of AI),” says the report.
The report’s authors also pointed out that for both AI and robotics across all the surveyed countries, self-reported awareness was higher among the more educated as well as men.
When asked questions around the perceived impact of intelligent machines on society’s, half of the South Africans who participated in the study thought the more widespread use of intelligent machines in the workplace would lead to greater levels of inequality between the rich and the poor. In addition, South Africans were also twice as likely as the other countries to think that the use of intelligent machines in the workplace would reduce inequalities (21%, compared to an all-country average of 11%). However, South Africans and Germans (at 62% and 66%, respectively) are more likely than the average (55%) to think that the increased use of intelligent machines would lead to less control.
It’s worth noting that the report’s authors caution against the over-interpretation of the study’s findings outside of the research context and add that the results aren’t meant to answer policy questions.
Here are some findings that stood out for us:
• South Africa had the highest proportion of people who were not familiar with AI and robots. The proportion of respondents who’d heard or read hardly anything or nothing at all about robots was much higher in South Africa and Brazil when compared with other countries. 1 in 3 respondents in South Africa (37%) said they’d heard or read anything or nothing at all about robots
• In South Africa a quarter (24%) had never heard of robots which is substantially higher than any country in the study (in all other countries this proportion was less than 5%)
• South Africa and Poland (both at 36%) were least likely to know at least a fair amount about AI
• Respondents in South Africa and Brazil were most likely to say they knew hardly nothing at all about robots (both at 36%), while respondents in South Africa were markedly more likely than all other countries to have no knowledge at all about AI, 23% of respondents in South Africa had never heard of AI, while this figure was between 1% and 12% in all other countries
• With a country average of 29% strongly agreeing or tending to agree that robots should resemble human beings, South Africans at (33%) had higher than average levels of acceptability towards the concept of robots looking like humans
• South Africa was the most polarised in terms of opinion on the acceptability of robots looking and behaving like humans in a workplace or public settings (33% agreeing, and 58% disagreeing)
• 67% of South Africans think intelligent machines will be able to communicate as well as humans
• When asked about the impact of intelligent machines on inequality between the rich and poor, 50% believed there would be greater inequality, however compared with other countries a much higher proportion of South Africans (21%) considered that a proliferation of intelligent machines in the workplace would instead result in reduced inequalities
• Respondents in South Africa and the USA were most likely to think that more widespread use of intelligent machines would lead to their respective countries being completely different (54% and 44% for South Africa and USA, respectively)
• In a reflection of South Africans polarised views on intelligent machines, 14% of South Africans were of the opinion that intelligent machines would not bring about any change in their country
• 26% of South African respondents gave a strongly negative score on the impact of intelligent machines in their country, while 31% gave a strongly positive score ai
