The ethics of artificial intelligence

In this issue of Farmer’s Weekly, our focus is on intensive agriculture and the technology available to help farmers produce more with less water and fewer inputs and on smaller areas of land.

As the global population continues to grow at a rapid pace, so too will the demand for food, which means farmers will find themselves under increasing pressure to meet this demand.

Objectively speaking, of course, the best way to meet said demand is through increased production and, in turn, the best way to boost production is through the use of technology and machine learning. In our feature on page 44, we take a look at the latest irrigation technology available to farmers. While they differ in terms of application and use, they all have one thing in common: artificial intelligence (AI).

While drip irrigation and centre pivots have been around for some time, the recent advancements in these technologies call for the use of soil sensors, among other features, to measure soil moisture, humidity and temperature. These measurements are then analysed by computers to advise producers on the best time to irrigate and how much to irrigate. These technological advances have certainly helped farmers to reduce inputs and optimise efficiencies and yields, but we rarely pay much attention to the price we pay for them.

A few weeks ago (see FW, 24 February), I wrote about machines taking over menial labour jobs, such as Chippy, the chip-frying robot in California, US. However, AI is advancing at such a rate that even professionals are being put out of work.

For now, I’m not going to concentrate on the labour component of the rise of the machines, but rather on the ethical implications thereof. In a recent conversation with my cousin, a medical doctor in Canada, he said that in that country, self-driving vehicles were becoming more and more of a reality. However, he mentioned a theoretical case study in which a self-driving car had to swerve to avoid hitting either a deer or an oncoming car. In most cases, many people would rather hit the deer in an attempt to safeguard the lives of the people in the other vehicle. Could we expect the same from an autonomous vehicle? Can a computer be programmed to recognise a human life over that of an animal? Who takes responsibility for the outcome?

The latter question can also be asked of the AI used in any other sector including agriculture. While the agriculture sector does not work directly with the health and well-being of people, farmers and agro-processors are indirectly responsible for the health of those who consume what they produce.

Should AI become the driving force behind decisions made in agribusiness, who will ultimately be responsible for what is put on our plates?

Currently, farmers still make farming decisions, in collaboration with analysis provided by AI. My concern, however, is that we are treading a fine line, and I think it may not be entirely clear where that line is. I believe that, at least for the time being, we are better off using technology in food production. But this may not always be the case, as sentience and automatic decision-making become more prevalent in AI.