Voices

Can AI make us more rational?

2019-01-06 11:25

Growing up as a young boy in the land of the Tshivhasa in Limpopo, I knew a middle-aged man, John, who had a wife named Sarah. The couple had a son, Peter, who must have been about 20 years old.

One morning, John and Sarah had a verbal fight. Later that day, there was a mighty thunderstorm. Lightning struck their compound and killed Sarah. A few days later, in broad daylight, Peter hacked his father to death with a machete. Word had it that Peter had made a causal link between his parents’ fight and the lightning striking Sarah. In fact, while he was killing his father, Peter accused John of being responsible for creating the lightning.

Peter was arrested and sentenced to a long prison term.

For the purposes of this article, I have used fictitious names, but Sarah, John and Peter represent people I knew personally. At the time, I was just 10 years old and unfamiliar with the principles of electromagnetism that could help one to understand the source of a lightning strike. The person who came up with the theory of electromagnetism was British scientist James Clerk Maxwell, who studied at Cambridge University.

Now I know that John was an innocent man. The idea of making false causal links based on some hidden, often supernatural force is called superstition. Superstition is irrational and can be very dangerous.

Years later, I went to Cambridge University to study for a doctorate in artificial intelligence (AI). The house I lived in overlooked the city’s famous park, Parker’s Piece, and the two were separated by a road called Park Parade. Every year, there was a bonfire at Parker’s Piece. We were told that this bonfire had been taking place for more than 300 years. Initially, it was held to burn alleged witches. Many poor, innocent souls were burnt at the stake in Cambridge for practising “witchcraft”.

In fact, I found out that a not insignificant number of Europeans went to the Americas to escape being burnt to death because they were deemed to be witches. Unfortunately, the witchcraft label followed them across the sea. Burning people for any reason is immoral; burning people because it is believed they are witches is irrational and superstitious.

In the land where Maxwell lived, there were superstitions. I suspect that, somewhere along the line, scientific thinking permeated the society and the burning of witches stopped. Like I said before, superstition is dangerous.

China is a very old society with a sophisticated civilisation. Yet, until the 20th century, women’s feet were bound so that they could not grow. The results were severely deformed feet, regular infections and walking disabilities. Whatever the reason for women to be subjected to this cruelty, it was irrational. When the Republic of China was formed, this practice was banned.

When China’s paramount leader, Deng Xiaoping, was modernising the country, one of the principles he adopted was that China should be ruled by people with scientific education. This is the reason the incumbent, President Xi Jinping, is a chemical engineer. His predecessor, Hu Jintao, is a civil engineer – and Hu’s predecessor, Jiang Zemin, is an electrical engineer. So embedded is scientific thinking in the minds of Chinese leaders that former president Hu’s theory, which is written into China’s Constitution, is called the Scientific Outlook on Development.

I suspect that, somewhere along the line, scientific thinking permeated the society and the burning of witches stopped. Like I said before, superstition is dangerous.

What Deng was doing was moving the Chinese people from superstitious beliefs to scientific thinking. I was inspired by Deng – so much so that, when I became vice-chancellor of the University of Johannesburg, I identified, as part of my mission, to contribute to the creation of a society that makes decisions based on evidence using scientific principles. In essence, my goal is to move our people from superstitious beliefs to scientific thinking.

At the core of superstition is irrationality. Rational thinking is the use of information and logic to arrive efficiently at a conclusion.

Economists use the bombastic phrase of “maximising utility” to explain rationality. But we are now in the era called the fourth industrial revolution, when technology is becoming intelligent, and people and machines are becoming a single being.

The fourth industrial revolution is driven by AI and is changing the world of work, human identity, politics and society. Today, people cannot be separated from their technological gadgets. A modern instrument of torture is to separate a person from their smartphone. After a few hours, they will experience withdrawal symptoms and their brain activity will show similar workings to people who are trying to kick a drug habit.

Today’s computer gadgets are so powerful that they can measure our temperature, heart rate and so on. They gather so much information from us that entire businesses are being built around the use of data that these technological gadgets gather.

Take Discovery Health for instance – it has an app that monitors our behaviour, including the distance we walk every day, our daily heart rate and other activities, and uses these to predict how healthy we will be in the future. This information is important for the actuarial pricing of health insurance.

The technologies that these phones have can help us make rational decisions.

Daniel Kahneman and Richard Thaler, each a winner of the Nobel prize for economics, observed how human beings on their own are fairly irrational when making decisions. They further illustrated how such irrational thinking creates inefficient markets, which are bad for fair trade. This irrationality among people seems to be universal, whether it is in Tshivhasa or Cambridge or Beijing.

With the advent of AI, can this irrationality be minimised? In our research on whether AI machines are rational, we observed that, just like human beings, machines have limited (bounded) rationality. What we also observed is that AI machines are more rational than human beings.

For this reason, AI machines are now deployed in hospitals to help doctors diagnose diseases. They are also currently working alongside pilots to drive aeroplanes. AI machines are being used in our banks to assign loans, price assets and manage risks. They are used by the military in combat drones to maximise enemy casualties.

Given the fact that AI machines are more rational than humans, what is to be done to increase the aggregate rationality in our society?

Firstly, we should invest in AI technologies in our schools, communities and industries. Secondly, we should study the science of human-machine interactions in all facets of our society.

Thirdly, we should embed universal scientific thinking in our preschools, as well as in primary, secondary and tertiary education. Fourthly, we should mobilise our society – be it in the political, civil or industry spheres – to adopt technology as an important motivating force to solve our problems.

Finally, we should make a concerted effort in all our spaces to move our people towards adopting scientific thinking by integrating technology into our decisionmaking.

Marwala is vice-chancellor and principal of the University of Johannesburg and the author of Artificial Intelligence Techniques for Rational Decision Making

Read News24’s Comments Policy

24.com publishes all comments posted on articles provided that they adhere to our Comments Policy. Should you wish to report a comment for editorial review, please do so by clicking the 'Report Comment' button to the right of each comment.

Comment on this story
0 comments
Add your comment
Comment 0 characters remaining

January 20 2019