Hi everyone, it's Steven Downs here. Welcome back to ethics analytics and the duty of care. Here are looking at my cat, Charlemagne, who has walked himself into a corner along the ledge. This is something that he does quite a bit hard to believe. Hey, sometimes that's how I feel about ethics and analytics in the beauty of care.
You're just following along your ledge. It seems pretty safe and secure, and then you hit the end, you can't back up. You can't go down. Oh, well, so yes, this is. Ethics analytics and the duty of care. We're in module eight and module eight covers. The topic of ethical practices in learning analytics.
And in this section, this is part five of the series of videos on ethical practices. That I'm authoring. As part of the four stair steps in the stairway of ethical practices of talked about that a few times. Now, we started with regulation. Now, today we're dealing with ethical practices just wrapping that up, then we'll be looking at ethics and culture.
And then finally, the ethical ethics as viewed. Try that again, finally toward an approach to ethics as harmony cash flow that when we get there. So, okay, ethical practices part five. What we're gonna be dealing with here today, our basically how to to actually activate the ethics in these ethics frameworks pointed out.
The problem with governance frameworks previously. You know, they're based on organizations agreements, shared presumptions, and naughty. Not actually motivated by ethics, but there is value in the work that is done, and there is value in the work that's done. Particularly with respect to ethics and governance frameworks. And assuming that there is value, you know, is this?
There are criticisms so we need to address those criticisms, but assuming that there is value, we need to talk about how these frameworks are actually activating how you move from something. That's basically raw theory to an actual practice in the field that maybe I could have started this, but I just decided to end with it.
So, five basic points, first two are related, It's from an HBR article. First of all, talking about what not to do, because they love that. And then what I've called, the HBR approach, which I might typify as the approach based on incentives. Second and third are related concepts of knowledge, translation and knowledge mobilization.
And then one that I think is generally where we trend toward ultimately, something like connective action. It these are sort of like minor stair steps in our staircase. That take the part of the big step of ethical practices. Okay, maybe I'm pushing that analogy too far here. So what not to do per read, blackmen in 2020, who's writing in a practical guide to building.
Ethical AI. So, here are the things he says we shouldn't do. First of all, the academic approach, the academic approaches basically, should we do this? Should we do that? Is this right? Is this wrong? And he says no, that's just to Lucy DC. What we should really be asking is how can we do it without making ourselves vulnerable to ethical risks.
Now, I've talked about this risk-based approach already and how the least risky thing is not necessarily. The least ethical thing and vice versa, or the least risky thing is not necessarily the most, ethical thing, and vice versa. Maybe a relation between ethics and risk. But thats a point of view that thinks of ethics as being defined by the role it plays in business outcomes, whether legislatively, or in the bottom line.
And I think most of us, think of ethics as evolving. More than that. The other thing he says we shouldn't do is what's called the on the ground approach beyond the ground approach. Is to actually have people who are working on the software or the AI or data application deal with the ethics themselves, as part of the engineering process.
And we look at that under IT governance frameworks his response and I kind of alluded to this and engineers data scientists and project managers locked the skill knowledge and experience to answer. Ethical questions adequate. So we can't ask the academics who actually study ethics. We can't rely on the engineers data scientists and project managers or product managers because they just don't have the skill.
What about just appealing to the AI principles directly? Well, the difficulty he says is in operationalizing these principles. What exactly does it mean to be for fairness? Okay, that's a fair point to be extent. That is true. Certainly, there's a lot of dispute about what constitutes fairness and we've talked about that.
And we, I think have shown pretty much conclusively in this course. That there's no single definition of ethics that can apply to everyone in all circumstances in that everybody agreed to and that's the main weakness with high level ethics principles. So let's say he has a point here like these kind of almost at a point here and and which is why I included this discussion in the slides.
So what does he suggest instead? Well this is what I'm calling the HBR approach, although really, I should call it the incentives approach. So again, it's it's a checklist. So you got to do a bunch of stuff, identifying infrastructure, create the risk framework change. How you think about ethics?
Well, this is an interesting one change. How you think about ethics by taking cues from the successes and healthcare? We're going to look at knowledge translation and knowledge. Mobilization later on in this video and and not very much takes a lot of takes into account. A lot of what's happened in health care, which is great.
And, you know, we looked in the ethical code section. We looked at a whole range of ethical perspectives and healthcare, but at the same time, somebody in Harvard business school in the United States, talking about health care, as though it's being handled. Ethically is kind of a problem for me because in the United States, as you know, you have to pay on an individual basis for health care.
There's no national health care program, really to speak of and so healthcare is hugely expensive and not available to a large percentage of the populations better now than East TV. Now, Obama care exists which covers a large number of the people who formally had no access to health care but still big problems.
I'm not very ethical approach to health care in my opinion, optimized guidance and tools for product managers all of which falls under the the heading of IT governance. I would say build organizational, awareness, whatever, not means but the key principle here formally and informally incentive eyes employees to play a role in, identifying AI, ethical risks.
Now, probably he means something broader than that, as it stated now, it's basically an ethical risks bounty, right? You find an ethical risk. You get paid. That's probably not the way it should work. Sometimes incentives are negative incentives. So, you would punish employees for not being ethical. Again, probably not what you meant sleepover all.
I think that the idea is to tie compensation and incentives to ethical performance. Not necessarily about ideas and certainly, it's the free market approach, right? And if it were possible, I, I'd kind of be in favor of it, you know? That's like me saying, you know, insofar, you know, to the extent that something I like works.
I like it, but it's a bit empty. Otherwise, you know, there's been a attempts made to do this in other areas. Climate change is an obvious one where the organization of climate change. Treaties, since Kyoto and forward has been to formally informally, incentivize individuals and companies to produce less carbon, you know, pollute less, contribute, less to global warming and the big objection to that was that it's interference in the free market.
Even though it is a free market approach to solving climate change, people didn't want to pay for it. And similarly the sort of objection here that would happen is that people don't want to tie their income, andcentives to ethics. If nothing is going to motivate them, to be ethical in the first place, they're not going to see the value in tying their compensation to ethics and, you know, our arguably people.
Really are incentivized by much. Wider range of things. You know, I think of tying teacher incentives to to the grades produced by students on tests. That's mostly not why teachers got into teaching. And it almost I won't say they don't care what they're incentives, are they don't care what their compensation is because I'm obviously they do, but there were other things in their lives besides compensation.
And this shows up every time you see a teacher pushing, back against tying incentives, to course grades. And the same thing with ethics and the same thing with any sort of exclusively incentive based kind of program. It's really hard to pick out and incentivize the exact motivations on a person has to undertake and action.
And what you get is some people gaming the system to maximize the incentives and other people ignoring the incentives and doing what they were going to do anyways. So I think it's kind of, you know, it's kind of an approach based on a certain ideal about the marketplace. But I even the marketplace doesn't run according to marketplace principles and so I don't see why ethics would and that leads us to knowledge translation which is the the first of the influences that we're going to draw from from the field of health care as it says on the side and I'm quoting from BMC Public Health here.
Knowledge translation was defined at a consensus meeting of the world health organization in 2005 as quote, the synthesis exchange and application of knowledge by relevant stakeholders, to accelerate the benefits of global and local innovation and strengthening health systems and advancing people's health. So we could see how this would apply to ethics to, right?
We could see how the work that has been done in ethics by the academics, by the engineers, by the people who've developed ethical codes of which there are many could be apply by relevant stakeholders to accelerate, the ethical benefits of global local innovation in. Well in our case learning analytics and artificial intelligence makes total sense and we can see the process in the process is one that has been implemented in health care and has produced some beneficial results.
So we'll start with KT1, are know. Let's start with S1. I guess that step one or system, one or whatever. It's hard to know what these things stand for stakeholder one, that's right, publishing in plain language and accessible formats kind of what I'm trying to do with this move.
Although I think I failed that probably in the second or third video, not the last, I'd like to bring it back to that. But okay, so we have our multistakeholder team defining research, questions and methods. So here are requests and methods, and then the research is done with stakeholder engagement stakeholders again, right?
And then our research findings are placed in synthesis placing research, findings in the context of other knowledge and social cultural norms and that's going to feedback into the publications. But more importantly, feeds into the contextualization in education, we usually say localization of knowledge. I mean standards, we might say an application profile and then we apply this knowledge ethically sound application of knowledge upholding.
The priorities of the affected community and ensure, knowledge translation process fits within neighborhood. And then, we look at the impact, which will be fed back into our stakeholder team, and back into our research process. Makes a lot of sense, doesn't it? So of course that our problems and in this article in NCBI there are three assumptions.
Identified that underpin, the knowledge translation metaphor and it's clearly a metaphor oop. My battery need renewed soon, on my, oh, I'm down to 10%. That's not good. Let's see if I can inner hurry, put some battery into the system. Well, that's gonna, but sort of and will that help?
Or will it still keep doing stuff? All right. Still recording. Okay, still recording still doing its thing so hopefully that'll keep working. It's bad. He's done too much audio recording when you've run right out of battery. Okay, so the assumptions also are implicitly here, criticisms. So the first assumption knowledge is seen as unproblematically separable, from the scientists who generated it and the practitioners who may use it.
In other words, it's an objectivist approach to knowledge. You have this knowledge that sits there, as this thing in itself, not related to the people who created it, not really to the people use. It second knowledge and practice can be cleanly separated, both empirically and analytically. You saw the process right?
First, you do the research, then you apply it in practice. Nice, clean separation. And then third practice consists more or less of a series of rational decisions on which scientific research findings can be brought to bear but arguably and we've seen this in the ethics. All of our discussions on ethics the all of the codes, all of the theories.
All of the practices etc. Under determine specific practices in the field when somebody's actually doing a thing of learning analytics and they're wondering is this ethical or not. The everything that the research tells them will not determine and answer and that's kind of a weakness in the knowledge translation metaphor, you know, no matter how much you translate the knowledge.
It's not going to help you in the actual practice. It's interesting. This is interesting to me because this is very much what happened in the philosophy of science generally and in the philosophy of science, you had this idea, a scientific progress was objective and rational and then could be applied to the world and through the writings of people like Thomas, try that again.
Thomas, Coon and WVO, Kwine, and others. We began to see that the distinction between theory and evidence isn't nice and clean. That our theory influencers are evidence in our evidence influences, our theory and the same with with practice right theory and practice. The division is not nice and clean evidence theory practice.
They're all meshed up in this one, big mushy ball and so you can't just say, we'll take the research and translate it that as an aside because it came up earlier today is also a pretty good argument against hatches effect, scales of different educational interventions. So okay, what do we have beyond translation?
Well, we're still in translation, okay, so lost in translation models and processes developed in learning analytics. Research are increasing in sophistication and predictive power true. However, the ability to translate analytic findings, to practice remains problematic. Also true. And in this paper quoting first awareness, that a particular tactic or strategy is general?
Generally beneficial is often insufficient motivation for a student to adopt it. No surprise there. Just look at statistics on people who smoke cigarettes, they know what to kill them. They're still smoking cigarettes second. The labeling of these practices is itself a challenge and we look at labeling, a labeling issues when we are looking at the workflow in the analytics and AI workflow.
So what we get then following from that is a refinement of the approach called knowledge. Mobilization, there's a lot to be said for knowledge mobilization. So, here's and I'm drawing from the drawing from the KMB tool kit here. But also from the social sciences and humanities research, council of Canada discussion, paper on knowledge, mobilization, and number of principles or attributes of knowledge.
Mobilization one making evidence accessible understandable and useful for knowledge users. So, you think about that, that's kind of a difference from knowledge translation because before the evidence became part of the research, you need present the research to the practitioner. What now you presenting, the evidence directly to them. See the difference there, right?
Secondly, meaningful use of evidence and expertise to align research policy and practice in order to improve outcomes. No, it comes. Again is one of these open-ended words that could mean anything depending on who you are. But believe that aside, the idea here is that research policy and practice and or I could say evidence theory and practice and be saying the same thing here in a sense, only need to come together to to improve outcomes.
Third thing, not just about disseminating information. In other words not just about sharing or publishing or one way information flow. It works both ways, we have engagement and user participation and attention to impact and it includes practiced-based evidence from the real world from the expertise of practitioners. So really what you're doing here is you're breaking down.
What might be called? The researcher practitioner barrier and the practitioners become researchers and an important sense. The researchers become practitioners. They become too haves of the same enterprise and they're each engageed in the activities of the others. Not is obviously going to result in an approach based on dialogue, collaboration working together.
And so you get the idea here of socialization, externalization combination, and internalization all as part of the knowledge mobilization process. It's kind of in the diagrams, trying to get at this and kind of a confusing way. A blending of tacit and explicit knowledge. Now, why some of these areas are two ways and so, okay, the task is to ways and the explicit to explicit is two ways but the explicit leads into the task at the task it leads into the explicit what leads into means.
I don't know. Something's diagrams are vague and it's interesting, right? We go intoward the center, I guess, that's the practitioner and out towards the theorist, and perhaps the other way around, it's again, kind of a two-dimensional description here, but that's okay. We know how we could extend that to multiple dimensions if we needed to.
So it's an interesting approach and this kind of approach does respond to the major objections of knowledge translation from a knowledge management or knowledge. Mobilization I think that's a better from a knowledge mobilization perspective. And here I'm drawing from a paper from Oasis. It involves some water called things we know about knowledge.
I think these things that we know are not necessarily things that we know and I think some of them could be considered contentious but nonetheless, it does give us a framework in a perspective. Knowledge is socially constructed and it's used takes multiple forms. Bodies of consistent, evidence are more powerful and effective over time than single studies.
Most practitioners have a range of connections to research sources of, practice and decisions, are hard, to define. With precision knowledge is always mediated in some way through various social. And political processes, knowledge by itself is not enough to change. Practice The scale of impact matters, the relationship between knowledge and use runs in, both directions, and cite the Vickenstein here.
Personal contact and interaction remains the most powerful vehicle for moving evidence into practice interaction. Sure. Personal contact. Not only a matter of producing more knowledge but also of improving, both the desire and capacity for its use and I've left out a few points here. But also dedicated effort matters.
So this is what might be described as a very post-modernist perspective on knowledge. At the same time it does take knowledge properly so-called maybe not as objective in the way that knowledge translation does. But as objective from the point of view of a particular individual, when something is socially constructed, it is a thing that's out there and not a thing that's in here and it's, it's a thing that is going to be deliberate developed collaboratively, through various social, and political practices and is going to have different relations to different things, depending on their perspectives.
But it's not, anything goes, right? It's not subjectivism. It's not a subjective picture of knowledge and that may be part of the problem with knowledge management when it comes to ethics because it's really hard to say. For example, ethics is socially constructed. I mean, sure, ethics is socially constructed to a degree but there isn't a thing ethics out there that is socially constructed.
It's like we went to construct one thing and constructed a couple thousand things. Um, similarly bodies of consistent evidence or more powerful and effective over time than single studies, probably true. The problem is with ethics, we don't get consistent evidence etc, right? There's enough differences here, between ethics and knowledge.
A lot of what's true about knowledge or at least held to be true from a knowledge mobilization perspective doesn't hold true in the case of ethics I would arse also argue. It doesn't hold true from the perspective of knowledge, you know. Like for example the relationship between knowledge and use runs in both directions, I would take a very vicinity and approach here and say knowledge is use.
We demonstrate that we know something by doing it, for example. No that's a bit sloganish but you know I think that we think that there is this artifact out there in the world that constitutes knowledge that we've all built together and there is this artifact in the world and it's something and it's really interesting but it's not knowledge, it's tools mechanisms frameworks, definitions ontologies.
All of these artifacts that we leave behind as we carry ourselves through the world and these artifacts are useful and they are used just like roads and bridges and buildings and cars and whatever else. But all of these things are society, you know? And similarly all of the different artifacts that we use, in the context of creating knowledge, don't themselves constant acknowledge, you know, it's the difference between talking about all of our ethical behaviors in the world and ethics ethics for me, is how I behave not how the world behaves.
And there's a really big difference between the two and this is why applying knowledge mobilization can be a bit difficult. And some of the practices here, you can see already how it's sort of pushes and stretches. What your typical researcher might find comfortable the need to go beyond traditional publishing.
It's a really good example of this. Now, I'm a living walking talking breathing example of the need to go beyond traditional publishing because traditional publishing is a terrible outlet for the sort of work that I do. But for a lot of academics, a traditional publishing is what they need in order to make their career.
So there's a very strong disincentive to mobilize their own knowledge properly. So called secondly researchers need to find an uptake for quotes strategies that excite them I have encountered this right? For example, I spent many years working on personal learning and personal learning environments. It's really hard to find an uptake for that.
Because personal learning environments, do not support the business model of 99% of education technology companies. And in fact, it effectively undermines their business model and significant ways because it makes education more accessible and cheaper and more in the control of students. And therefore, they need less intervention from teachers universities and content providers.
So no uptake, but I think that personal learning environments are in many ways, more ethical, a way to manage learning than say, learning management systems. I could simply be wrong about this, but the fact that the reasons up take, for the strategy that excites me does not mean that it is not ethically, right?
And that's why we we hit this third challenge, the need to put aside and I'm quoting, they need to put aside your personal agenda and to challenge your own disciplinary, or sectoral, approaches to become a part of something bigger, which is great, for the people who want to become a part of something bigger.
Not everybody wants to do it and more, particularly not everyone wants to be a part of that thing that's bigger rather than that thing that's bigger. I mean, I want to be a part of something bigger in the sense that I want to be aligned with all the radicals, and the activists and the rights advocates, and maybe even the anarchists a little bit.
I mean the environmentalists the gay rights people etc. Etc. Um, that's the bigger. I want to be a part of but mobilizing knowledge. Especially the way I do. It doesn't really mesh with the crowd. I love that crowd, but they're not tackies mostly. And so there isn't the techie agenda.
And so what do I do? Do I put aside everything that I'm doing, even though I think it kind of works toward that something bigger. No, I just forge ahead in my own way, doing my own thing, doing it to the best of my ability and maybe it works and maybe it doesn't.
That's not up to me to decide and that's different. And, and that's the thing with knowledge. Mobilization is that, it really tends us toward these group-based collaborative-based projects, where you subsume your own interests, underneath a larger broader, some other kind of interest, it might be a corporate interest, it might be a cultural interest to social interest whatever.
And and these really are basically defined by these four sectors here community government private sector or academic institutions. If you can't find an interest under any of those, you can't do my knowledge mobilization. And to me, I say so much the worst for knowledge. Mobilization.
So not surprisingly instead of Well collective action, which is essentially what knowledge mobilization is. I favor something like connective action. And no, I don't want to miss interpret this, this article here and let's, let's be clear that this article isn't saying exactly what I'm saying, but I'm drawing from the article to make the point so drawing from the article from Priyanka Singh.
It's the idea that forms of the idea of forms of activism that focus on quiet acts of caring rather than an amplification of a message or a platform. So, connective action.
Isn't based on mass it's based on connection. Now in the particular discussion here it talks about using hashtags as a form of connective action but I don't think that's quite it. I don't think hashtag activism is the same as connective action. So I'll leave that aside and let's not bring that part of it into the discussion, but there is the distinction between working as a person in a network touching other people in a caring, kind of quiet way as opposed to getting up on a podium and mobilizing people into action.
So idea, heart and says, it's a movement driven, not by uplift or the struggle for recognition or citizenship, but by the vision of a world that would guarantee to every human being free access to earth, and full enjoyment of the necessities of life. According to individual desires, tastes and inclination, let's kind of what I'm after with connective action.
So the idea of personalizing contentious politics in the personalization of contentious politics was written about by Bennett and Sagerbird a few years ago. Now, it's not that long ago, but the key ideas here are of a platforming clinician as an alternative to standing up to or against something. And I interpret that, as you know, building rather than agitating being the person who actually creates the connections, does the thing that instantiates the value rather than trying to convince someone else to instantiate the value?
It's the idea of producing hope rather than looking for hoping the sky, it's the distinction between a performance of care. You know, performative act versus doing the work of care actually doing the work and it's the idea of communicative labor at the point of organizing rather than the more visible in the more mass-based forms of resistance.
So there's a lot there that speaks to me and where we're going to go with this in general. I'm going to be talking about culture next, but it's not culture, in the sense of mass, it's not culture in the sense of everybody's speaking French, or everybody going to church on Sunday, or everybody enjoying the hockey game, or anything like that.
It's closer. I think to culture in terms of the way I defined community earlier community, as consensus of ways of coming together to decide on how we agree on what is true and society. And in the case of ethics, it would be how we get together to decide on what's right?
And not just generally, you know, that would apply to everyone. But what's right here now in this context, when I'm performing, my individual act of care. How do I decide? What is right? What is the mechanism for that? And I think that that's a different way of seeing the role of all the ethical work that gets done.
And what actually constitutes ethics in practice in real life, it's not really about the research and the organization and the publication of papers and the running of surveys and all of that. It's about the cumulative weight of individual acts of care in a society, which defines what our ethics are, and the first level of talking about that will be to talk about that as a cultural phenomenon.
And then later on, we'll talk about what that means from the perspective of ethics itself. So that's where I'll leave this discussion a little bit hanging but let's probably where we should be because really what we're concluding at the end of this section is that ethical practices? Don't define ethics.
It's like regulation. Maybe tells us what we shouldn't do ethical practices. Give us a framework for implementing what we should do and shouldn't do. But we need this other thing in order to inform the practices, the research, the product, the implementation of analytics, and artificial intelligence. So that's it for ethical practices.
I'll see you again for the culture segment of the module. I'm Stephen Downs. Thanks for hanging in there with me. We're almost through this.
- Course Outline
- Course Newsletter
- Activity Centre
- -1. Getting Ready
- 1. Introduction
- 2. Applications of Learning Analytics
- 3. Ethical Issues in Learning Analytics
- 4. Ethical Codes
- 5. Approaches to Ethics
- 6. The Duty of Care
- 7. The Decisions We Make
- 8. Ethical Practices in Learning Analytics
- Course Events
- Your Feeds
- Submit Feed
- Terms of Service