Approaches to Ethics

By Stephen Downes
Nov 08, 2021
Transcript of Approaches to Ethics

Unedited transcript from Google Recorder

Hi everyone, and welcome once again to ethics analytics and the duty of care. I'll just get this power point set up here, which I forgot to do before the start of this. So I'll do that now. Sorry about that. I thought it was all set and ready to go, but of course I wasn't because that's what happens when you're doing stuff, live, someone sometimes.

And when you're doing it by yourself, and you don't always think of these things. And of course, we're gonna have a train go by two just to make life fun. There we go. Approaches to ethics. I might trim the first few seconds off this video, if so, welcome to ethics analytics and the duty of care, the beginning of module five, which is approaches to ethics.

I'm Stephen Downes. And I'm leading you through this course today. Now, what are you going to start off talking about when we talk about approaches to ethics? Is the question of the basis for statements about the ethics of learning analytics? And AI, we make lots of statements. We say things like people's rights to privacy should be respected, or we need to balance the risks and benefits.

Etc. Etc. We've gone through a whole number of these already. The different ethical principles, the different ethical values. The issues that are been raised, excuse me. We've covered a lot of this stuff already in this module, we're going to look at the deep and rich history of ethics to see if we can draw out.

What the basis of these ethical statements is why we make them. What justifies them? So let's look at some of the language that's actually used in some of the papers that talk about ethics and analytics in AI, here's one from K, corn and up in Heinz, 10 years old but it's still relevant.

It says to satisfy expectations of the born digital slash born. Social generations. There is a likely requirement to take on ethical considerations, which may run contrary to the sensibilities of previous generations. Especially in respect of the trying to trade off between privacy and service. Excuse me, one moment. Nothing like getting a frog in your throat just as you're doing a live video.

So my take on their take is really that why we're doing it to satisfy the expectation of born digital and born, social generations that doesn't really seem right to me. I don't really see that as being a reason to start looking into the ethics or even to change our positions on the ethics.

It's not about satisfying expectations of a generation. How about this drew 2016 saying? People think that it would be irresponsible not to use data science in certain cases and that we should not lag behind other countries again. What sort of reasoning is this? Do we draft our principles about AI in the ethics of AI based on what people think or whether or not we like behind other countries you know this comparison between other countries is used to justify a lot of things whether we should develop a certain industry or adopt a certain policy.

The the idea that we're lagging behind suggests a moral superiority to the position. Those countries who are ahead are taking. But what does it mean to be ahead? Well, look at the Brookings institute which writes basically whoever leads in artificial intelligence in 2030 will rule the world until 2100.

So the idea here is what is right? Is ruling the world. Really? I'm not so sure about that. What about this? A voltiers candidate might have said, were faced with the imperative to seek out the best of all possible worlds. This is one of those worlds where we do the trade-offs, the risks the balance is, and all of that.

We have this requirement to ask is this the best of all possible, worlds and candid goes through a world, where there's all kinds of violence and injury and just generally bad behavior and horrible conditions. And yet still comes to say this is the best of all possible worlds. This is as good as it gets.

And the reason why we have to put up with this bad stuff is so we can have this. Good stuff was on a reason for ethics, is that a reason for artificial intelligence? I think these three approaches are kind of superficial. They're kind of superficial not in the sense that they aren't good arguments.

Oh, I don't think they are a good argument but they're kind of superficial, in the sense that they don't actually address ethical questions at all. They're all fancy ways of changing the subject. You know, we go back expectations. That's a different subject. Not liking behind the others. That's a different subject.

Best of all possible. Roles that too. Is a different subject. But what about what is, right? And what is wrong? These are the questions that ethics gets at and these are the questions that we want and ethics of learning analytics and artificial intelligence and education to get at not.

Whether it's popular or whether it rules the world, whether it's right or wrong. But how do we determine that? Well, ethics has been a topic of interest. Well forever human interest in, ethics goes back, at least 3,000 years. I was going to put 3,500 years and it wouldn't have mattered.

We go back to the epic of Gilgamesh. We've go back to the Iliad of Homer, the Icelandic Edis, and we see an ethics with a set of values that see, strong leaders of small tribes and we see that that code throughout history. We see things like the Sumerian farmers element and the Egyptians instructions which both advised farmers to leave some green for poor leaners.

These are both examples from Wikipedia so we don't see a difference sense of ethics. This study of what is the right way to live? What is the right way to be? What are the appropriate actions to take in different circumstances in different conditions? Probably is as old as human society itself.

And in fact, we might argue that the very possibility of a society requires ethics or another way of putting it. Ethics the results of ethics is the creation of human society. Once we started thinking about, what does it mean to do right or wrong to each other, that's when we began to get along with each other.

Ethics has a long history and association with religious beliefs. I was going to list a whole bunch of things I could have gone and looked at Sharia code from the Islamic tradition. And then the law codes that follow from that or I could have looked at say, the 10 commandments or the guidance of Jesus, in the Christian Bible could refer to the analytics of Confucius, or the guidance that allowed Sue makes in the downaging.

There is a long history of ethics. Telling us what the right way to live is. At the same time this association is weaning. I say we mean you know loose sense not as a precise mathematical calculation but you know, we see in the chart here, on the right hand side, whether belief in God is essential to morality and we see in some societies such as say, the United States Egypt, Indonesia El, Salvador and Ghana, people believe it is.

I yet another societies Canada. France israel Australia. People believe it's not necessary to believe in God. To be moral, there is a tradition of writing in this school. I referenced Miguel Goulan's summary of Greg Epstein's book, good without God and it follows in a long tradition of such books.

Kai Neilson who taught at the University of Calgary. While I study there, has written ethics without God in jail Mackie as well known for his book, ethics inventing, right and wrong. So, there is a sense in which ethics is associated with religion, but arguably, it's not a necessary connection.

At least from the perspective of this inquiry. At this point, in time, in Western traditions and I emphasize Western ethics begins with as a philosophy. Begins within the Greek tradition and histories of ethics, such as it added encyclopedia.com, which take a very Western perspective. Talk about it. Beginning with Plato now, arguably or not Plato with Socrates arguably, there was discussion about that before.

And again I mentioned and a previous discussion the law codes of so on which predates any of these, but Socrates asks, what is justice? What is right? And we get answers, like justice is the rule, the stronger or man is the measure of all things. These are the sorts of answers that the softest might give and Socrates, you know, and basically devastating analysis shows the difficulty inside simple views.

It's so it's a lot more difficult than that. This rule of the stronger can be unjust, man, might not be the measure of all things. How would we know Plato comes up with what we might call the ideal form of the good, it's like triangles. We have the idea of form of the triangle.

We have the ideal form of the good and just like the triangle we can know it by thinking about it and just like the triangle, it's how we measure all other objects of knowledge Aristotle is well known as being less ideally minded than Plato. He comes up with a theory.

He comes up with a complex approach to ethics complex approaches to ethics, but we might summarize them if we had to and we do as something like the exercise of natural human facilities in accordance with virtue. So these are taken from that article on history of ethics from encyclopedia.com, but you'll see similar statements expressed and pretty much every history of ethics as defined from the Greek tradition since the enlightenment, arguably before.

But certainly not long after ethics has been associated with reason, the enlightened, the enlightenment and humanism. And a lot of the philosophy that we see associated with people, like Descartes and Voltaire and Pascal with this famous wager, remove or move things like ethics from the domain of the, Well, I won't say supernatural, that's the wrong word.

But from the domain of the ethereal, to the domain of humans, to be something that we as humans fallible, as we are, can know and cannot prehat. And this idea that ethics has been associated. With reason, has been with us. Since is probably the dominant approach to ethics today.

Yes, we still see appeals to character and human nature. Yes. We still see appeals to religion and even ancient values. But all of these even today are generally couched within the form of the argument, the form of the rational inquiry. Perhaps looking for evidence, certainly posing things like thought experiments.

So that the reflective person came somehow think of and come to grasp at least for themselves and idea of what ethics must be. There's a strong relation in this rationalist tradition between belief and ethics. And in fact, the one feeds into the other, we have an ethics, or a moral obligation to believe responsibly.

And we have ethics that are founded on the basis of responsible belief. You see how they feed back into each other, it was hard to express that without making it a tight little circle but the argument is there, right? For example, false beliefs about physical or social facts like say the side effects of vaccination, may be listened to poor habits of action or poor practices of belief formation, like say listening to fake news.

May turn us into careless credulous believers, and here's something that's, especially relevant for today. Elbow, what was written in 1877. We have a moral responsibility, says Clifford Clifford. Not to pollute the well of collective knowledge. Now that can't be a direct quote, but it does come from a rebes article.

Believing with out, evidence is always morally wrong. There's different ways of believing in a, the try not again, there are different ways, belief and ethics can interoperate, and I've depicted two here on the left, the way reasoning, including ethical reasoning about say beliefs and desires influences actions on the right.

However we have what is normally called something like rationalization, where our actions are actually caused by our instincts, our norms and our habit. But when we think about it, we are able to infer to what our beliefs are desires were or perhaps, are this form of reasoning is known as abduction or inference to the best explanation.

And you can see that operating here and although it's typical to say something like, you know, it's only a rationalization after the fact it doesn't mean anything. It might be that in fact, rationalization is rational if we're not trying to determine what our ethical beliefs are, but rather trying to explain what they are.

Mars and Metcalf looking at the question, ask whether ethics itself is to big a word that maybe the concept of ethics just brings in too many things. I mean, we've already seen that it brings in religion and history and culture and calculation. That's an awful lot there. And then today, we have a much more complex, ethical picture.

Looking at things like compliance, legal risk, corporate values, and moral justice, and more all sometimes, describe musics and awesome. Sometimes working in parallel. As they say, sometimes coming into tension with each other and they may have a point. Although and the discussion at noon, thought about this talking with Mark and thinking, you know, I don't think ethics is too big a word.

I think there is a concept. It may be a loose concept. It may be a poorly defined folk concept, but there is a concept of ethics. There is a sense, maybe? Right and wrong aren't exactly the right words. Maybe rational is the right word, maybe compliance is the right word, maybe more, always the right word.

Maybe justice is the right word, but something like that, that we can point to and we can describe and we can refer to, with respect to our conduct in education teaching, and learning with respect to artificial intelligence and analytics there is something there there. I don't think breaking it all.

Breaking it apart and hiding. Off sections that we won't talk about is really the answer and effect. If we are to come up with an answer, I think that we need a concept of ethics that is going to at the very least explain if not give us insight and how to form things like compliance legorous, corporate values and moral justice.

Among others, they mass in that calf. Look at the adjective ethical and they look at the different senses of what we mean by ethical. For example, we might have ethical outcomes, they write that this might consist of a cloud service company declining a contract with abuse of within abusive government, or agency, or equalizing error, rates across protected classes, in an automated hiring system.

The outcome in other words, is the result of the process. It's the non-discrimination that we see in our hiring. It's the non-compliance with unethical behaviors. We see in some governments, it is presumably as well. The pursuit of a positively defined, good such as the betterment of society, the provision of more rights and more capacities for people, however, you're going to define it.

Here we have a picture here that talks about organizational and citizenship behavior in versus unethical, pro-organizational behavior. Let's, that's pretty good indication of what we mean by moral or ethical outcomes. But of course, ethics is more than it comes. There's ethical processes and you know, we can compare ethics in this sense to something like scientific method where what's important isn't, what comes out the other end now?

Presumeably in science. We want, what comes out the other end to be, you know, true, accurate, productive effect of or whatever the different scientific values are, but what makes it? So is the process that when in to producing that output and in science we call that the scientific method.

Now we could talk for a long time about exactly what we mean by the scientific method. But it is generally agreed that there is one and it is generally agreed that the raise one because it produces good outcomes. I mean after all we're living in an age of technical and and and other wonders technical and scientific wonders we're able to develop a vaccine for covid in a year.

I mean that's virtually miraculous but it's a consequence of scientific method. Similarly in the domain of ethics, a rigorously ethical practice process can look like some kind of messy, even unnecessary process but the idea here is you follow the process and something good, comes out the other end. What does that process mean?

Well there are many descriptions of ethical process and that's part of the problem. This one here talks about knowing the facts looking for the right people knowing the applicable laws being accountable towards stakeholders noting the core values of the company and being objective in decision. Making. Now that would seem like a very odd statement of an ethical process to other people.

And so it's not simply having a process that makes something ethical, but it's having an ethical process, which means something like a process. I guess that would produce desirable. Ethical outcomes. Another way to look at this and we talk quite a bit about this and the previous module is ethical values.

Now, here we have a description of values that describe states humans or any other being desires, such as beauty, justice wealth and I had our expressed as ethical principles little diagram, there shows a bunch of them right determination respect integrity quality culture, morality trust discipline character energetic industry. Well, there's too many to name.

I came up with a reasonably long list of values, but arguably, there are many more and then you have to take these values, instantiate them as principles capable of concrete action, and then figure out what to do when the principles conflict with each other. As they inevitably, will another way of looking at ethics is to think, in terms of ethical requirements, they check box system that we seem to have adopted as the norm for so many of our institutional and technological processes.

For example, here, we have an article from nature telling us ethical requirements or requirements for AI systems. Derived from ethical principles or ethical codes or norms. So we have unique value risk benefit consent, traceability privacy and these are feed into an institutional or ethic review board, approval process, thick click, click the same effects, that's that what we mean by ethical.

I don't think it is. I don't think Amy of these four accounts covers what we mean by ethics, maybe another way of looking at it is to look at the difference approaches to ethics. We're going to do a lot of that in this module, this year is a very quick sketch and I'm not actually going to divide the approaches to ethics in exactly this way.

But I wanted to highlight this from the last quiz because it does actually represent a fairly contemporary perspective on what the different approaches to ethics are. And the fiber approaches are as follows the utilitarian approach, which is a consequentialist-based approach based on something, like, creating the most good, for the most people are different approaches of interpreting, a consequentialist position on ethics.

And we'll look at those in some of the next talks. Another approach is the right space to approach. And and we've seen quite a number of the ethical codes and statements on ethical principles. Adopt and explicitly writes based approach to ethical issues in analytics and AI in learning and teaching.

And there's a lot to be said for that approach. Certainly, we would think that it's unethical to violate somebody's rights. We'll have to look at the how we identify, what goes rights are, what constitutes a violation of a right? And whether the domain of ethics is exhausted by a discussion of rights.

And of course, somebody will always say, well, if you're going to talk about, right, you have to talk about responsibilities. Another approach is the fairness or justice approach this derives in current philosophy, at least from John Rawls book, a theory of justice which defines justice as fairness and many ethical principles that we've looked at so far are based on this concept of fairness fairness itself is represented as an ethical value.

And we talk to about that, there's a history of this sort of approach and I'm going to kind of I'm going to characterize that in this module under a more general heading of a contractarian or a social contract based approach. And so we'll be looking not just at roles but also at people like hobs lot and Russo.

And I think about the different bases for and the different ways we can go about creating social contracts. Additionally, there's an approach based on the common good once again makes us think of Russo. I bet it also makes us think of Mal and it makes us think of communitarian-based ethics where whats good.

And what's right is based as much on what is good and right for everybody as on what is good and right for oneself, and then finally, as mentioned earlier, there's the virtue approach and I will talk about that virtue, based approaches may have begun with our Aristotle. Although certainly, I think people have talked about that before Aristotle, and it's enjoying a modern renaissance if you will, and there's quite a bit of discussion of character-based approaches to teaching and learning.

And these character-based approaches to teaching learning, I think will inevitably sleep into a discussion of the ethics of analytics and AI learning. So, like I say, I'm going to recast these five approaches. I want to bring in a more historical perspective, and I want to break out a bit from the fairly classes classic Western approach, to ethics that.

These five approaches represent. Finally, we're going to look in this module at what is called meta ethics or as my old professor, John Baker would say matter of ethics and I would say there's no air in meta ethics. And here, we're looking for, you know, we've got all of these ethical theories and ethical approaches here.

We're looking for the basis for ethical reasoning. How do we choose among them? How do we make decisions among them? How do we? If we want balance them off or how do we assign priority to one approach over another, if this gets down to some of the really fundamental questions in both the historical and contemporary discussion of ethics questions, like does might make right?

You and you know our first inclination is to answer. No it doesn't. I'm yeah, if we look at say international diplomacy, it's certainly seems that it does and so we need to take that question. Seriously, or how about the question of whether ethics describes duties? Can a fairly abstract?

And it is abstract concept such as ethics and moral reasoning, create and impose duties on people, can I say to you and be justified in saying, you have a duty to give money to the poor, or you have a duty to rescue a drowning dog, or you have a duty to refrain from murdering people.

Our ethics based on rights, we mentioned that before in the different ethical approaches. But here we flip around the question. Suppose we have rights does not give us a theory of ethics suppose rights are foundational on the other hand. Suppose rights are not foundational. How then to write and ethics relate to each other, drew ethics require agency.

There's an old principle bought. Implies can that is to say if it is a more obligation to do something than necessarily, you need to be able to do it and usually this is expressed in the converse. If you are not able to do something, then you do not have a moral obligation to do it.

Well that seems to me sense but what do we mean? When we say not able to do something? Because we hear this kind of argument alone in contemporary society. We cannot stop using oil through power our society. Therefore we have no ethical obligation to stop using oil or to move toward renewable energy sources for example.

And then finally, the big one that gets everybody relative and relative ism versus universality. If there is an ethical principle doesn't apply to everyone all the time. If there is a system of ethics, maybe a consequentialist system, maybe a right-space system, does it apply to everyone all the time?

After all the United Nations didn't just call it the declaration of human rights, they called it, the universal declaration of human rights. Conversely, what if ethics is different from one environment to another and our discussion so far? And looking at the ethical codes and particular we've seen that we can think of ethics as being very context, bound bound to specific profession bound to a specific set of objectives or outcomes that we're trying to achieve.

So maybe ethics is relative and if it's relative by profession maybe it's relative by culture. We saw that difference societies view, the role of God in ethics differently. Does that mean? Then that different societies can have consistently different sex of ethics, or does it mean that some side societies are basically unethical and other societies are ethical.

And if so, how do you judge, which one's these are tough questions but these are the sorts of questions that we need to wrestle with. And the expressed this express this kind of badly in the initial discussion that we had at new today where I said something like, the reason for this course, is that people having learned about ethics, haven't learned about analytics and haven't learned about education.

But what I mean by that isn't that people didn't haven't been scored by someone like me, who will tell them about all these things? What I mean by that is and people haven't asked the questions. A lot of the discussion of ethics in artificial intelligence and analytics, simply assumes say that privacy is a right.

Must be respected, but when we push that and we must push that, what is the basis for such a statement after all privacy protects criminals, as well as the innocent? And, you know, it seems like well, of course we should just balance this but what makes a consequentialist approach a technical approach of balancing it the right approach.

I mean you wouldn't balance killing and not killing. Would you or would you sometimes it seems that our society would. So those are the kinds of questions that need to be asked and they need to be asked and we need to think about some of the answers that are possible before coming up with pronouncements on what the ethics of analytics and AI in learning and teaching looks like, If you haven't asked the question, you're not in a position to provide the answer.

I think that's obvious, but but maybe it's not. So are study of ethical approaches in this module, isn't about learning the different, ethical approaches, I could care less, whether people know the different ethical approaches. It's just that the different ethical approaches give us possible answers to some of these questions.

And at some point in the process of reasoning about AI analytics and ethics, we should consider these possible answers and we don't have to remember them all that would be silly, but we should consider them the way we might consider, you know, what route to take to New York City, or whether to have chicken or beef for dinner.

We're not gonna memorize all the options, but, you know, if we don't consider the options and we may spend the lifetime eating nothing, but chicken and never try and beef. And that would be sad unless you're a vegetarian, which face that would be good thing. Yeah, you know what I mean?

So that's what we're up to, in this module, looking at the answers to some of these questions, both in terms of the ethical approaches. And in terms of meta ethics that people have come up with over the years and where it will lead. I think is for us to think about how to approach ethical reasoning.

Generally, I think may change our approach. Ethical reasoning is certainly in my case, it did. And we'll talk about that toward the end of this module, where we will, ask basically, what is the end of ethics. What are we up to here? And does this discussion ever end? So, with that, I'm Stephen Downes and I welcome you to module five and the continuation of ethics analytics and the duty of care.

Force:yes