As digital technology transforms our world, computer scientists must consider the ethical impact of their work. In her powerful Digifest workshop, Miranda Mowbray illustrated why this is so important. Here, she shows how universities can keep up with the pace of change.
Life was different in the 1970s. In those days, computer scientists didn’t, for the most part, have to make difficult ethical choices as part of their jobs. They were in basements tending to machines, making sure their code worked well. Today, they’re in boardrooms making decisions that may affect democracy.
Because of the rising power, influence and importance of computer systems – which are embedded into pretty much every aspect of modern life - they now have greater capacity for producing good and bad outcomes. Social media, for example, exerts a big influence on the way we work, play, interact, and on our politics. That’s exciting for computer science.
The world has changed
But it means that our teaching has to change. At Bristol University, I've been giving guest lectures on ethics for computer science. Students asked for this topic to be in the curriculum.
The British Computer Society (BCS) has interest in it too. In order to gain BCS accreditation, in fact, computer science degrees are now required to have content on legal, professional, social and ethical issues. But the method of teaching it is important. I have seen ethics courses that just say ‘do this, don’t do that’.
In the workshop I ran at Jisc’s recent Digifest event, I explained that it’s important that students learn to apply ethical reasoning themselves, so that when they come up against a new ethical issue in their professional life, they can analyse it independently.
Part of the problem is that this requires discussions, and computer science students don’t have a reputation for liking discussions. Also, it’s really hard to build efficient computer systems that output the right results and don’t crash all the time.
Learning to do that at university takes at least three years of work, so it’s understandable why traditional computer science degrees just teach the technical skills. But we need to develop ethical reasoning and communication skills too.
'All interesting ethical questions are dilemmas'
I start by telling students that there may not be a single correct answer to a discussion question. If anyone shows a view which is unpopular, I say that’s fantastic; we have a disagreement. I did this at the Digifest workshop too, because even Jisc’s delegates – many of whom are lecturers and educationalists – may be reluctant to openly express an opinion that others may not share.
I try to hold myself back from revealing my own opinions upfront too, so as not to intimate people who think otherwise, and to allow an exploration of different ideas. All interesting ethical questions are dilemmas with arguments on both sides. If you disagree with someone, in order to persuade them of your point of view, it really helps to see where they’re coming from; and if you can do that, if your opinion was actually wrong, you may be able to discover that.
The ability to have respectful, rational discussion with people with whom you disagree is a highly transferable skill. It’s important for life as a citizen and as an employee.
I draw on real-life examples in my class. I saw the video of Christopher Wylie talking about what it was like to be the research director for Cambridge Analytica, and how he now strongly regrets what he did. He thinks it was very unethical. But he was under huge pressure to deliver at the time, and he was only 24!
Computer science graduates can very quickly find themselves in a position where they’re affecting the quality of democracy, and they may have no training and no support in this ethical decision-making position. We talk about these dilemmas in my classes.
For example, in some judicial systems, after a criminal has been convicted, in order to help a judge decide whether they need to be locked up or not, a machine learning algorithm is used to predict whether or not they are likely to reoffend. That machine prediction is more accurate, on average, than predictions made by humans in the justice system. Should we leave the decision entirely to the algorithm? Most people say no. Why?
Looking beyond the law
Ethical considerations need to go a lot further than just asking whether or not an action is legal. Some things are commonly considered unethical but aren’t actually illegal. Laws tend to say whether something is permitted, but ethical analysis can indicate which of two or more options would be better, and so can help improve systems that are already OK.
One of the three main ways ethical philosophers have suggested to tell whether an action is good is to see whether it conforms to rules. But there are two others: look at whether it’s in line with positive values; and consider the likely outcomes for stakeholders. I encourage students to use all three ways of looking at an ethical problem.
Although there are arguments on both sides of interesting ethical questions, that doesn’t mean that it’s all relative and just a matter of opinion. Students need to be able to reach a decision on whether something is good or bad, taking into consideration the arguments on both sides. They should talk about ethical questions with their peers, and with their colleagues when they’ve moved into industry. If there is a potential ethical issue in a company, it’s easier to address it as a group than as a single employee.
Too often, computer scientists feel that discussing ethics isn’t part of their job. Well, it is – and it’s fascinating and important.