Content-type: text/html Downes.ca ~ Stephen's Web ~ Three Frameworks for Data Literacy - Transcript

Stephen Downes

Knowledge, Learning, Community
Three Frameworks for Data Literacy - Transcript


Unedited audio transcript from Google Recorder


All right. Hi everyone. I'm stephen downes, purveyor of bad videos and recordings. Uh, it is recording video right now and i'm off the frame. Um, so if you have any comments later on, By all means. Please be aware that you're being recorded. And the internet will. Use what you say against you in some way, shape, or form, because it's the internet.


So i work for the national research council of canada, And, But Canada's national research agency, they somehow gave me a job. And, I've been with them since 2001. So i somehow kept that job. I specialize in the area of online, learning new instructional media. I've been involved in things like massive open online courses.


Open educational resources, personal learning environments. And i have my daily newsletter called OLK. Uh, you can scan the qr code on the slide. That will take you to my presentation page. Where you can download. Uh, these slides or a pdf version of the slides. As well as link to the paper that the top is based on.


As well. As after the talk, the audio recording, Which is being trans or not trans. Well transcribed. So the actual transcription of this talk plus some bad video. Um which i hope i was hoping to stream it live but at the last minute it said you have to set up your live streaming and it's first time it's done that in several years.


So That's working with tech. This talk is titled three frameworks for data literacy.


All right, i have a clicker. Ahead the opportunity to practice with the clicker earlier but i declined it. Now, i regret that decision. So data literacy. Is and i'm just pulling a definition out of the air here. Uh, we'll discard it In a little bit. But basically the ability to collect manage evaluate and apply data, all in a critical manner, And it includes various skills to discover access data.


To evaluate data quality. To interpret the results of the analysis and so on. It's a relatively new field. Started more or less around 2010. Sort of went away. And then came back over the last few years with a vengeance. And we can pretty much imagine why it came back in the last few years, with a vengeance.


It's kind of interesting to me because it covers The different topics that we've talked about today. Going everywhere from Analyzing. The content validity of say surveys. To looking at. Different ways of teaching people skills like conferencing to even thinking about and creating artworks. And so, all of these will be captured in some way.


In this talk. I'll try to feed them in as we go through.


Oops, either way. So, What i do in this talk is i present data literacy in terms of three frameworks. The first framework consists of what is data literacy itself. What is it actually trying to do? Boy, that's really small. Sorry. I have a very short attention, spam. I don't know how to make it bigger because the plus sign is gone.


Don't know, i can't see it. Turn around. Yeah, it's you know, talking about teaching people how to go to conferences, how to do conferences. The first thing they teach you is don't be like that guy. Who stands there and gives his talk like this. Right. Is not true and that's why they have a presentation view.


That, but literally this, here's my slide that i'm showing here. It is this big My eyes are 64 years old. They don't work like that anymore keyboardology. I think that's required. Yeah, it's incredible.


So anyhow. Um, something in the competency model, obviously no, sorry. Oh, it's not this one. Sorry. Sorry. I just come back to the To getting their sneak preview here. Well. Yeah, the notes section is huge, right? But of course, i have no notes and even if i did i wouldn't be able to read now the good better.


Oh, that's way better. Thank you. My pleasure. I can actually read it. Good. All right. Yeah. Yeah, don't get 64 year old eyes, unless you absolutely have to Hey, but you will absolutely have to i hope All right, so three frameworks. First of all talking about the competency models that define data literacy, second the assessment of these competencies and third methods for the development of these competencies in an organization.


Now, Very high level presentation because we had what eight pages to produce our results. So, unfortunately, we skip over a bunch of things in this presentation. Although i will do some follow-up papers looking into some of the details and i'll flag those when i hit the All right, literacy.


Again, we can do volumes books on what literacy is. Um, But i think we can say that learning a literacy is more than just learning the components. Of illiteracy. It's like learning a language, is more than just learning the words, right? It's when we use them how to use them, what do you dreams are etc.


And there's an element of Being literate that comes out of. That learning right literate. Literacy isn't just knowledge that you have. It's something that you become as a result of that knowledge. What is then that you become while you embody shall we say a set of skills or competencies typically thought to define the literacy in question?


Um, these are reflected in the assessment of that literacy and in turn the teaching of that literacy is based on. That assessment. The actual study of data literacy, again it's a new field, it's pretty limited, he was even more limited when i started working on this project a few years have gone by and there is a lot more stuff that's come out in the last year.


Which isn't really reflected in this paper. So, The reviewer said, i need to say more about my methodology. I'm not one of those people who does the You know. How did that go again? Uh, Research questions trying to hypothesis. Yeah, even even the peer review thing and anyhow Um but i did try to be his methodological as i could because this study was being done.


For the assistant, deputy minister. Of an information branch for the department of national defense in canada. So they did kind of want a method. So we did a formal literature review in conjunction with the national science. Library, contents and NRC's own information management service We also did a wider review using the same parameters for google scholar.


Now, either of those were really comprehensive i find especially in rapidly moving topics, That a lot of the best and newest material can be found in what they call gray literature. Pre-prints blog posts, social media etc, and i did depend the lab a lot on that to find sources.


If i simply depended on the official published sources, i would not have had nearly the material to conduct this review. And that's something that's why i ask that question. I'm looking at you but you're not only gave that talk


I mean, how Um, So anyhow, phone about 150 results. In those 150 results. I found. 20 papers that could reasonably be said to define a theory of data literacy. I found three major evaluation frameworks that i looked at there were other. Evaluation frameworks based on those three major frameworks.


Solar was, you know, maybe six or seven all together but really three core ones. And then, A number of highly specific data, literacy development models. Since then, he told me about your course, i also found some Materials on working with data from Carlton. Um, carlton college. I've been stuck it for Carlton university.


I just saw that the other day, there's a bit of a bunch of new stuff come out recently. Um, but it's all highly specific. It's all very Local to a particular environment in a particular course. So, These are the themes that emerge. Virtually everyone depicted data literacy as a set of skills or competencies, which isn't really that surprising given The top or you know, the literature around literacy generally.


A lot of them talked about the idea of deriving. Information or meaningful information from data. Uh, there was a lot of talk about the Data lifestyle or the data workflow. Um, It was the Complexity of skills for different roles. And, What i found, interesting data literacy can be defined both on an individual and they use corporate, but i think organizational more generally can be thought of as an organizational or a group capacity.


And that's not really usually how we think of literacy is it. Usually, we think of literacy as an individual individual set of competencies But in this case, literacy was depicted as and evaluated us. A corporate or an organizational competency. So, let's look at the competency model. So competencies set of basic knowledge and skills for other characteristics.


Having enabled people to work efficiently, etc, Um, We draw on. Well, established concept that includes knowledge, skills, abilities and other characteristics. And, The, the concept of competency also includes. Importantly, some way of measuring for it or evaluating for it. So it's a two part thing. Right. Here's here's what you know.


And here's how we know that, you know what, you know, except it's not what, you know, it's what you can do. It's how it's what dramatitude are. It's it's well etc as we will see. Here's the analysis. Your slides are a little bit out of sync from your one behind.


Oh, oh, okay, i get you. Oh, how silly is that? We're following them, right? I'm really sorry about this one. It's in the paper. Um, but those are the 20 studies across the top. And, These are the competencies. Now, what i did in the study is I kind of had to wordsmith a bit because Among these 20 papers.


Nobody used the same set of definitions terms etc. We have 20 completely distinct depictions of not only data literacy generally, but these competencies and in particular, So, there is a fuzziness there, which is my fuzzy content validity, is a good idea. Um, I'll read them quickly for you just for the fun of it.


Awareness, dispositions strategy, or culture. Uh, Client inquiry, discovery. Ethics gathering duration, communities requirements, valuation, evaluation, assessment, and formed decision. Governance. Standard description or metadata. Conversion or interoperability management preservation cleaning systems and tools, policy quality security manipulations, statistics and reasoning. Critical thinking analysis interpretations that I had to pause for breath there.


Marveling architecture data science and machine learning. Uh, Visualization storytelling pretend presenting data verbally which is what i'm doing now. Change. Using or presenting with. Identifying problems and data generation. Each of these, all each of these breaks down into subcategories like data generation. Creative automated, etc. You can see why they contacted us.


Asking for an analysis of data literacy. Because every paper they read on the subject said something different. To find the differently. Where's? Where's the pattern here? Completed right. You know, the more we analyze it. The last we see a pattern Um, and that's that's the state of the literature.


And that's why i'm doing was paper. So, Oh yeah. First step, a kind of broke it down into Five major models. I borrowed from Uh, Wolf. And i also borrowed from shield. To to name some of these but nobody names all five. That's unique to this paper. These are rough categorizations just as a hubric hubrick.


Lubric rubric. No. Whatever, there's a word there that isn't rubric. But i can't and never mind. Don't present when you're old. All right. So here are the five models. First of all, gave us stewardship model. It talks about the idea of literacy from the context of being a steward of data.


Supporting analytics and decision making Um, You know, there's this includes the collection of approaches from From the data literacy project which is something that came out of the housing and nova Scotia. Um, It's the idea of Keeping data. Us a data custodian, if you will. And that's the type of literacy.


Of course, it's a type of literacy. The information literacy model. This is a very common and widely used. I saw it in a bunch of studies. And it borrows heavily from information literacy even digital literacy a bit. Um, of the problem, of course, is that the domain of information.


Literacy is even more fragmented than the domain of data literacy. So it doesn't help a lot. But you know it's about getting information from the from data theories of information information, management principles, Uh knowledge and the flow of information to borrow a title from fred gretzky and and related topics.


There's one that may be more familiar to this group signs and research data literacy model, it does include things like validity. Uh, different forms of validity. In emphasizes aspects of data related, to computer science mathematics statistics, Forms of statistical representation. The ability to analyze interpret and evaluate statistical information.


Uh, you see it representative for example. Um, In the alberta bureau of Statistics among others. Another model. I only saw it in a single source, but it was a singular model. So it became number five, the social engagement model. This is thinking of data literacy from the perspective of social interaction and especially online, social interaction.


It kind of includes things like data, journalism. Um, it kind of includes things like Seeing patterns and social media, like twitter. Um i refuse to use the other name not that i use it anymore. Anyways, you should be using Master on. Stuff like that. So, Also. The major a major theme that evolves and is discussed in.


The literature is the idea of a data workflow. In another bit of work that i did. On ethics and analytics. We'll see, we've been about that in a little bit. Um, There's a whole machine learning. And artificial intelligence workflow. And part of that is this data workflow and includes everything from the framing of the problem, or context of use the data set itself, defining it getting it, cleaning it.


And so on, The application of the model to some new situation and then testing evaluation presentation, etc. There's, Kind of a common workflow. Um, i wouldn't say it's strictly formalized, notice that it's different from the workflow, we saw presented a bit earlier. About presenting a scientific paper. And then finally, the individual and group competencies.


This is something people have a lot of trouble wrapping their heads around because they hadn't drawn that distinction before even though a number of commercial companies were presenting studies there were selling studies to companies saying evaluate your corporate data literature data literacy skills. And the people who were receiving, this were getting things like Evaluate your individual dental literacy skills.


And there are blending the two together. So they're the same thing, but of course they're not. How does you know and you're thinking about how does it differ? Well, with an individual for a, for example, They might have knowledge of how to use databases how to use. Excel maybe how to use cloud storage to hold your data on the corporate or organizational side?


They would be things like data management, principles, staffing and resources for managing a data pool or data lake. Um you know, a corporate Taylor, retention policy adherence to gdpr stuff like that same skill. But different from an individual to a corporation. So, We have a mass. Francliff. And how do we even assess?


That mess. This is where the model comes with. Why would i get it?


I'd like to say that's the first time that's ever happened to me. All right. Alright, i'm still behind. Yeah. This is so much fun, the slide that i can read really well. He's not the slide that's up there. I've been having fun with this the whole time though. Okay, so I mean, it was a, this is this is one of this is an example of Uh, accurate but but not relevant.


We got the accuracy. I now can read this, slide, it's the wrong slide. All right. So, my proposal is this single factor. Measurements of data literacy. Will not measure the literacy. Uh, they're completely insufficient to account for the variability in the set of data literacy competencies. And the varying degree to, which each competency is required in different jobs.


I haven't shown the second part. That would need to be shown, but i hope i've shown the first part. The second part really kind of comes out as we go along. Recordingly. A role defined data. Literacy model is proposed. And this model, Basically, illustrates the calculation of data literacy.


In this way. So, And how did we come to this idea? Well, first of all, we looked at what was out there for data literacy assessments. The three big assessments, the OECD. Program for international assessment of adult competencies, has a specific data, literacy section. There's also, Endorsed by the american statistical.


Association the guidelines. For assessment and instruction and statistics education, kind of narrowly focused on just one of the models, but still a very influential a widely used low. Actress in group and in particular, david wells has an excellent analysis of data literacy in general. And the assessment of digital literacy in particular, Looked at more.


Um but these were the big ones. The datability is one for example was drawn mostly from OECD and there were others that we looked at So, Thinking of the assessment modeling. Most of these assessed for a list of competencies. But as he saw, we have this completely unstructured list of competencies, no consistency.


At all. So Kind of created a model and unfortunately, this is all you're going to see of it. Based on a modified version of Bloom's. Taxonomy why bloom? Because everyone knows blue. Everybody understands blue Is just a taxonomy. He's not the way the world is. So it's fine, right?


Only we modified it in a couple of ways. Uh, we use this. We most people mostly. When people look at bloom's taxonomy, they look at the cognitive set. There are, of course, three sets, cognitive, cycle, motor and effective. Um, and then these are the for the individual values. We can talk about knowledge skills or competencies gratitudes.


No, i know. I'm talking about skills and competencies big in general, and we have skills and competencies here. That's what happens. When nobody uses terminology consistently, we have skills and competencies containing a subdivision called skills and competence and sorry. And then on the organizational side, instead of the knowledge, it's organizationally defined, right?


Instead of skills accompetencies we have capacities, Instead of the effect of domain being attitudes, we have practices Kind of pulled it over the air. I admit But pulled out of the air based on what those surveys were actually measuring for when they did. You know an hour organization has a standardized practice for data retention.


It's one of the questions, right? And then i likerid scale. Presumably assessed for content validity, although maybe not with a fuzzy definite model. Um, So, Kind of pulled over the air but not randomly pulled over there. Just as in a side. Talking about patterns earlier. Finding patterns isn't a matter of analysis and synthesis.


Uh, not to my mind. Anyway patterns finding patterns is a case of recognition to gestalt thing. You look at it, you see the pattern and then you rationalize it later by going through a process of analysis and synthesis. But the actual finding of the pattern is just a recognition thing.


Based on prior experience, we could talk about that in life because i will, All right. I didn't want to ask that as a question, but i thought about it. Yeah. Oh, you're wonderful. Thank you better feel. That's awesome. Cool. Now, i don't have to be like a first-year conference presenter.


Could be worse. I was doing the talk from the national research council. Um and it's Canada. So it's bilingual. So i'm expected to translate it as i go do the english i can do to french. No problem, i'm not great in French, but i'm not bad. So i thought, okay, i'll put all of my slides in french on the monitor and then the two language here.


So i'll be able to see the french in front of me. Well, the way they set it up, i was on a stage and there's a little computer waiting down there. That was my monitor. And there's no way. Always investigate the setup first. Okay. So, I know in follow-up papers not in this paper.


Unfortunately there's a very detailed breakdown of all the confidence as within this table. But listen, this is how it came out, okay? Um, over here. We have our competencies. And then here, We looked at the different descriptions. For all the different jobs or roles in the canadian armed forces.


There's a lot of And the idea is that. We tested. These drop descriptions. I don't want to say simple word count because that's inaccurate. But for The relevance of the data. Now, let me be clear. We did not actually do that test. So i can't report to you. Here is the statistically validated thing against all of these job descriptions.


It was we presented the conceptual model only. And with the proposal that they shouldn't test it, That'll be clear about that. And you want to pretend that God resolves made? I didn't get rid of that. Not i didn't actually get Intellectual honesty or i didn't do the work. Okay.


But, here's Not what should but what would come out of such an analysis? There's one of these spider web charts. They have all of the competencies around the outside. And then what i call a competency profile. On the inside. And basically, it's the degree to, which Inch job. What our job description more accurately.


Reflects a need for an individual competency. Because, When you get down to it, Data literacy amounts to something different for every single job that you're doing the answer, are similarities. Yeah, there's overlaps. Yeah, the concept of data literacy is a family resemblance kind of thing. But family, resemblance means that there is not something that they all and only.


They have right. There's an overlap, we recognize it if we see it, but there's no way we could have our analyze too. Get it out of what we're seeing. Hope that makes sense.


The same process. Virtually exactly the same process we suggested. But again did not test. Can you use to create actual competency profiles for each individual? He's gonna imagine how much the military loved this. But what's not to love? Right. Take a look. Uh, say their test results or even better since the military watches, everything actual communications generated by the person in question, Subject to ethics and privacy regulations.


Of course. As an aside. This is how we will do assessment generally in the future, not just of data literacy, but everything, why would we Do the whole formalized exam. Kind of thing. Anymore. When we have computers that can look at the sum total of everything you produce. And produce the competency profile.


And where that competency profile can be map mapped to a role profile. Where the role profile can in the first instance, be derived. From the job descriptions. But in the second instance, also not conducted but recommended through the Examination of the actual practices, undertaken by people who are already experts in whatever domain.


We're talking about, Think about it. We look at what they're doing. We create the profile. That profile defines the role. We taking individual. Look at what they're doing. That profile defines their role. We compare them. We can see whether a person is suited to the role.


So given over how do we go about developing they have a literacy in an organization. This was the least studied thing that we saw. There were a few specific proposals that were a few models. Um, but nothing like You know, a disciplined-wide consensus of any sort. Um, So, Gave the literacy basically seems to fall within two extremes.


And this is an important point, i think, i think, On the one hand. And we saw this in the models. It's you know one of the many information and communications competencies it's part of as much larger program. Um, you know, maybe journalism, maybe information science. Uh, etc. Or. On the other hand, for people who are more into science, technology engineering math.


It's the first step in the development of higher level. Competency searches data, engineering data art is hacked, information management, etc, very technical kind of position. Either way. Were envisioning data literacy as one part of something very large, very complex. And when we're teaching data literacy, we have to be teaching it in terms of what they're going to be doing with it later.


So, there's no such thing on this view, as a simple data literacy program. And that's probably not true. I don't think it's true. Um, i think we can think of data literacy kind of apart from that context. Providing we think of it as this raw based sort of thing.


No, i know that sounded a lot. Like saying we can think of it outside that context, as long as we think of it inside that context. Which sounds ridiculous. But essentially, that's what i'm saying. We think of it. Not so much as content or knowledge to be used. So, not in that context of the lighter discipline.


But rather as a part of other processes and strategies employed to achieve real objectives of Brooklyn within that context. So it's not nonsense and it only sounds like nuts. So, Looked at a bunch of data, literacy road maps to see how other people were approaching the subject gave a information.


Literacy project was one quantum. Found series of foundational steps. Big wells which i already mentioned from accrison. Excellent program and then Gartner has a three-phase methodology. They all did it that way, not thinking of it as. Part of this overall program, the thinking of data literacy more in isolation thinking of thinking of it as a process or a method.


Some more initiatives, getting literacy project, no longer exists. But they did very good work. Um and then there's the edge of cause data literacy institute, which really is just getting going on. Phones, a bunch of what could be called teaching and learning methods. And it's very similar to, i would put the The simulated scientific conference right in the middle of this list.


Because i think that's we're probably where it would be the most appropriate a data storming simulations k-spaced. Using real world data data decision, making etc. Can you wrap up? I think you're probably will have some questions for you. Yeah. You you wouldn't believe how close i am to my last slide.


So i developed a data literacy mook, which was a failure. Nobody joined. Um, but i also developed a mook in Ethics analytics and the duty of care. Um, which was much more successful but i used the same principles. And the idea was give people data have them work with the data, have them draw their own connections and their own inferences.


And so, In the subject of ethics analytics, and the duty of care. We have all of these domains, we have relations. Those lines are for representative purposes only. They're not actual links. But there is a graph of actual links in the course. Um, And so they're actually working with and producing.


Models and interpretations of the data. It was, it was a very useful exercise. I learned a ton and got a great publication out of it. Um, student activities in this environment? Classifying identifying identifying relations. Identifying. Data subject. Assessing the resulting data model looking at the threads. If you follow the lines from thing to thing, to think you have a story or a paper or whatever.


Participation. Encourage both individual and organizational data competencies, at least, i think. Again, i did not evaluate them for, did they actually achieve these competencies? Because i don't like giving students test. Um, But i did find it where it requires working with others. In order to develop. Not only the individual capacities and skills, but also social capacities and skills.


So, last slide almost. Sorry about that. Uh, There's no single definition of data literacy, but it's not yet taught that way. Just kind of weird. Uh, so we recommend developing and piloting non-hierarchical cooperative learning communities such as a MOOC in order to foster data literacy. But keep in mind, all of this is conceptual.


Except for the actual course which was built and delivered. This hasn't been tested. It's a framework to be used in future tests. Epilogue. This is a picasa. It is now my picasso. And i'm going to take it and i'm going to print it and i'm going to hang it on my wall, i'll have it the castle, but it's actually in the functional city hall.


And, The important point here is that data isn't just Rows and columns of text data is everything. All that stuff. And the thing is, i do a lot of photography. You saw a few examples of it in this slides. And, and some manipulation of that photography. And, I look at that and as much as i love it, i see all the reasons why it's not the original.


Like this, put get rid of it, you see the shadowing and yeah. So i just made it part of the picture because, yeah, you can't. So, Is by working with this stuff. That you become literate in it. There really is no other way. And ultimately, That's what fostering data literacy means, working with The data.


Gathering it manipulating it using it to create something, whether it's a scientific paper, whether it's a picasso or anything in between, And like i said, draws all of these things together, that's the melon country. I'm so sorry. Thank you.


So if you need to leave, that's fine, models to the end of the session. The questions for Steven. So i missed the very first two slides story. We went another session, we had to wait till this time to come over, but are you starting with people that are going to work with data in their jobs?


Or is it more broad on? It was more broad. It's more broad. Yeah. Yeah. We, you know, we, we were interested in everybody in in the organization. Including and then these stresses including frontline soldiers who need to know, whether they're being fed misinformation or not. Yeah, so we we have a project that we've looked at for citizens in general education, general people, we start a little bit more basic.


Exactly what is data. Have more sort of awareness that there's data and data B collected around us and it's going up like that. Because part of the part of what we found was like in the businesses as well, people who encounter anger in their every day job, they may not be working with the data but they should be a critical thing.


So something that comes across the desk asking for data or asking they could read something, they need to be able to according on it. So i think i agree with that. 100% false. These models are on many different levels, many different. We try to explain ones that we're going to work with data and all our studies.


So i will share with you our


But thank you for a very interesting talk, especially liked your idea about looking at data, losing not just as it individual and organizational level. But actually at a role level and this idea that you brought up about assessments, how it's possible to analyze pretty much every single communication and interaction of a person and compare that to the data literacy requirements for a certain role.


First thing that came to my mind just about pattern recognition. Recurrent, neural networks are very good at doing this type of pattern recognition, not just all of your communications, but all your interactions online. And there are countries that now score people based on all of their interactions online. To determine who might be Best suited maybe for a certain study program at university.


Scary for me, as a computer scientists. You might as you know, i have no evidence of this. But i would put very large amounts of money on the proposition. That companies are already doing this with social media virtually certain that this is the case. And linkedin. Um yeah you're specifically designed for that purpose.


In strength papers right here. So, i just want to confirm like was Different. Meaning, for the same thing. That's a good question. No no, no, you're right enough when i say that's a good question. I mean that's a really good question. Right. Is there a singing thing? That constitutes gated literacy that they have different meanings for Or.


A reach of them using a different meaning, for a different thing that we're sort of loosely categorizing, when i believe the second is the case. Right, so each each of the 20 papers. Looks at data literacy. Frames it like this and then looks at it. But they each do it a bit differently and they're looking at different things.


But they're overlapping in a family resemblance. And we, because we're humans with neural networks. We go. Oh, yeah, that's all the same thing. It's like vixenstein in games, right? What is a game? The again, you have The gazillion definitions of game. But each definition looks at some things chest, not a game.


Address again. No chess is a sport now. Anyhow Time in space messing with other again, right? Yep. And like the colon codes in the last table. Why don't we might codes? Around here in the in the, in the team mock. Yeah, here it was the course. Sorry. Yeah. Yeah, domains after it's oh oh i'm sorry.


Yeah. Um, that refers to ethical foods. Nice. Yeah. So And as part of another paper, Um, i did a study of 70 different Um, Ethical code to practices for artificial intelligence data, ethics. Legal. Ethics journalistic, ethics. And so on. Um, to examine the clean made by Two minute, felt someone.


Um, i think it was felt saying, oh no, we have a common set of ethical principles that we all adhere to. And that's what should define ethics for artificial intelligence. So, i examined all these codes I got a pretty similar result to what i got here. There is no single principle.


Eating leaving aside, the definitions of the principles. There is no single principle adopted by every code. It's cute, that's good, that's good. They're not even close. Right? And, you know, right off the bat, look at the major principles. Right virtue. Ethics duty ethics benefit or Consequentialism contract, ethics and meta ethics.


Or is my old professor? Used to say met or ethics. Okay, you never get out early. And, and on. Right. So yeah. So yeah. So the this this diagram interestingly, Works from the upside hand. And that's how my paper is designed as well. Kind of on the one hand from all the different applications.


Again, my little list, big long list. I do list and then working into the center. And then from the different codes, analyzing what the ethical implications were. Running them through a filter of duty of care ethics. I mean, in the center and this is the last chapter of the paper.


Ethical practices for Uh, artificial intelligence. Thank you. You're gonna have to end the session now but because we're a little bit over time, but please continue to discussions after the sessions with our speakers. I want to thank all of us.


Stephen Downes Stephen Downes, Casselman, Canada
stephen@downes.ca

Copyright 2024
Last Updated: May 30, 2024 01:59 a.m.

Canadian Flag Creative Commons License.

Force:yes