The ethics of AI in Academia. Benefits and promise of AI. What should you consider? Pitfalls to avoid?

Oracle Academy Tech Chat

In this episode host Oracle Academy manager Tyra Peirce speaks with Mary Olson Industry Executive Director, Education at Oracle. Tyra and Mary speak about the ethics of AI in education and give practical advice on how to best use the technology while avoiding common pitfalls. -------------------------------------------------------- Episode Transcript: 00;00;00;00 - 00;00;35;01 Welcome to the Oracle Academy Tech Chat. This podcast provides educators and students in-depth discussions with thought leaders around computer science, cloud technologies and software design to help students on their journey to becoming industry ready technology leaders. Of the Future. Let's get started. Welcome to Oracle Academy TED Chat, where we discuss how Oracle Academy helps prepare our next generation's workforce. 00;00;35;03 - 00;01;02;08 I'm your host, Tyra Pierce. In this episode, I'm joined by Mary Olson, executive Director, education at Oracle. Mary and I speak about the ethics of AI in education across multiple fronts administrative, academic and research. A little bit about my guest. For over 30 years, Mary has been using the latest technology to drive innovation and the education system. Welcome, Mary. 00;01;02;11 - 00;01;26;07 thanks, Tyra. I'm really happy to be here. So to start off, can you give me a little bit about your background at Oracle? Sure. I can tell you a little bit about that. I actually have a fairly short tenure at Oracle. I was recruited about a little over three years ago to lead the voice of the customer strategy for Oracle from A higher education perspective. 00;01;26;09 - 00;01;49;01 And I, I feel like I'm returning to my roots because I spent about 30 years at IBM before coming to Oracle. And my favorite job there was leading our education for North America. So I feel like I'm back home. I know today you're talking about A.I., and so let's kind of get a little bit more into AI. AI's constantly changing. 00;01;49;02 - 00;02;12;24 What are some of the ways you've seen AI used in education? well, you know, it is constantly changing and it's kind of funny because people think about it. You hear about it now a lot and they think it's new. Well, it's not new. It's been around for a really, really long time that the use cases are expanding and they're coming out in exciting new ways. 00;02;12;27 - 00;02;40;12 So when I think about how higher education and education in general is using AI, they tend to fall into some buckets. One of the buckets is streamlining the administrative processes and basically the running of the institution because face it, a university era or a school district is a business. They have to hire people, pay people, report their finances, do all kinds of reporting. 00;02;40;12 - 00;03;18;05 So there are a number of use cases around changing the way the institution is run. And that's that's probably the most significant use cases right now. They're also moving into research. So we see a lot of cases where researchers are really accelerating their discoveries wildfire research, cancer research, drug discovery, everything by using the generative and the analytical tools, machine learning tools that come with AI. 00;03;18;08 - 00;03;42;02 And so that's a really growing area. Another area that's kind of unique and that I'm finding more attention to these days is in providing students with services. So whether it's trying to help them with career guidance, what do they want to do after they graduate? Maybe help them write their resumes and polish those up? We're seeing a lot of applications there. 00;03;42;04 - 00;04;14;03 Another area that I think is kind of fun is helping to make it easier to actually teach. So we're developing use cases where they're a teaching assistant or a tutor that can help the professor actually answer questions for students. And there's a lot of applicability there. And probably I think one of the most surprising is it was two weeks ago I was reading in the San Francisco Chronicle and there was an article on how A.I. is disrupting the college admissions process. 00;04;14;06 - 00;04;37;11 And there's you know, it's actually kind of a big business to for parents often paying to help their students get into the right colleges with the college application process. And there are a number of startups out there that are saying, hey, don't pay this company a couple thousand dollars to get your students application filled out. Use our generative AI tools for free. 00;04;37;13 - 00;05;10;03 And when I looked at that article, I thought, you know, some people are saying you really shouldn't be doing that. But the California University of California has a statement of application integrity when you apply. And they're now actually saying that it is okay to use generative AI in your application process, provided it's used responsibly and ethically. Use it for content editing, for readability to make it, use it as a tool like it was designed to. 00;05;10;05 - 00;05;38;16 But they draw the line at actually plagiarism and just using it to write the product. So they say if they come back and they find out that you just used it to write the product, the application, then they will take action on that. That is so interesting for me. It's like when and I wonder, as I'm sure as we're getting out there and further, there are going to be ways to actually measure the tools in terms of like, are you creating a product or are you editing and, and things along those lines? 00;05;38;16 - 00;06;01;16 I think that that is just as you were telling me this, I was like, that is really fascinating that they're okay with it and they are disrupting that industry because what's the line of written by the code or written by you? And I'm glad that they're able to kind of to kind of take a look at that, which leads me into my next question. 00;06;01;18 - 00;06;26;11 What are your thoughts on faculty using AI in the classroom for teaching and learning? Well, I have some pretty strong opinions on that. So what I really think is that I go back to when calculators first came out and the faculty, a lot of people, a lot of teachers did not want their students using calculators because they thought they wouldn't learn the concepts of math behind that. 00;06;26;14 - 00;06;46;03 And so they were banned in a number of classrooms. But then, you know, the more they started to use it, the more they started to innovate with it. They learned how they could use the tool effectively. They could still teach the math concepts, but they could use the calculator to really accelerate the students learning and take some of the drudgery out of it. 00;06;46;05 - 00;07;09;11 I think we're going to see I used the same way. And to do that, though, I think if you're a teacher, what you need to do is you really need to understand how A.I. affects the subject you're teaching. So, for example, if you are teaching Python programing, you know you're in computer science, you're teaching Python. How can I help with that? 00;07;09;12 - 00;07;37;05 Well, I could just plain flat write the code, which happens in a lot of cases, or it can help maybe with debugging the code. And so if I were teaching Python, what I would do is I would have my students write their code, debug it, experienced the pain and agony of debugging it, and then run it through the generative tool to bring the original set of code through it to see how it debugs it and how much more effective that is. 00;07;37;08 - 00;07;58;11 And I would do that and make sure that I also included some failures. So I think it's important for them to know that just because generative AI may be writes the Python code, it doesn't mean it's perfect. You are still responsible, you still have to check it. And so I would be sure that there are some failures and ways that they can spot that. 00;07;58;13 - 00;08;28;24 And to do that, you know, just my advice to to instructors is experiment. Experiment with your own use cases. Have a couple of thoughts. I think of like I think of like my background is in is in writing and I think of, you know, a lot of times today especially I had a friend a few years ago that was an English professor at a community college, and she would talk about getting papers in where people are writing with emojis or writing with like text slang. 00;08;28;24 - 00;08;50;16 And so I think that this will help immensely. So, I mean, if you're not going in and plagiarizing, you're actually giving input there to see how writing should be done appropriately as opposed to not giving out what the saying L.O.L or TGIF or what have you. I think that that will help so they can kind of see what a structure supposed to look like. 00;08;50;19 - 00;09;14;24 The other thing that I think about with what you just mentioned is when I'm writing a Python code or a Java code or a database code, you're going to be if the AI is programed correctly, the app is learning correctly, you're going to avoid some of the pitfalls of security. Like I always think of sql injectors. I remember many years ago talking to some Oracle database expert about SQL injectors and learning how to code more securely. 00;09;14;26 - 00;09;36;14 And I think that this will kind of alleviate some of those and address those concerns as you're writing that code that it's automatically going to be coding more securely. Yes, it can definitely do that. And when you were talking about that, that also reminds me it can either of a professor who was telling me that he did have his student run the Python code through. 00;09;36;16 - 00;10;01;01 And what happened is he saw there's this moment where she wrote the code. She you know, she did that, but she didn't really understand it. And when it came out through the AI engine optimization and stuff, she said, now I get what we're trying to do. So it's a different way to tell. And just like your example of writing, you see good examples. 00;10;01;01 - 00;10;18;17 And so it's going to be your, your, it's going to elevate your own work. And it'll be interesting to see in terms of a database like if they're using inroads for data design before they're actually writing the sequel, that goes with it. So you have to understand the structure of it or you're not going to go you're not going to know what an idea is. 00;10;18;17 - 00;10;39;04 Like, you're not going to know how to name your columns more, to name your fields or what have you, without actually understanding the structure of beneath it. I think that that is where it is going to help. So it's not going to get rid of people like in our professions, but it's more going to enhance it so we can actually think about that thought leadership and strategy pieces as well. 00;10;39;06 - 00;11;01;27 That's exactly correct. Yeah. And getting back to the math example, yeah, I remember when I was learning math and they made me show my work. Yeah. It was always so frustrating to me because frequently I could actually get the correct answer but not really understand the steps involved. And so it's the equivalent of showing your work. So this goes long into my next question. 00;11;01;29 - 00;11;23;22 If faculty are allowing students to use AI to write computer programs or papers, what are some of the pitfalls you foresee? So some of the pitfalls and you touched a little bit on it with potentially, you know, plagiarism there. It's easy for a student to potentially fall into the trap of, hey, this looks pretty good. I'm going to going to use it. 00;11;23;25 - 00;11;45;18 And you want to make sure that that doesn't happen. There are a number of tools at Turnitin. One is one that's quite popular where you can run students papers through it to see if they've been, you know, created by generative AI or simply plagiarized, lifted off the Internet. You know, you can find that those tools aren't perfect, but they're getting better. 00;11;45;20 - 00;12;06;10 And I in fact, I was talking to a CIO at a university in Nebraska, and she mentioned that her daughter, when she was writing her college admissions essay, she ran it through the turn it in to see if it would look like it was plagiarized or something. And it came back and said it was even though it wasn't. 00;12;06;15 - 00;12;30;06 She wrote it herself, but it just looked too good. And so those tools aren't perfect. They have flaws in them, but those are ones that they can consider using to make sure that it's authentic. So on to my next question. What are some ways to ensure that students are still learning important skills instead of relying heavily on air to generate homework? 00;12;30;08 - 00;13;00;10 You know, so if faculty are actually encouraging students to use AI to help them write a paper program or something like that, there are some things that they probably should avoid. And one of them is not understanding clearly how to use it. So I recommend that they put clear guardrails in place and guidelines and explanations as to what is an acceptable use of AI in their class and in the project and what isn't. 00;13;00;13 - 00;13;22;05 And I know that that kind of shifts over time, and it really is very, very dependent upon the subject. So, for example, if you're going to be teaching, you're in grade school and you're teaching spelling, yeah, the students actually have to learn to spell and they can't just use an AI tool to write the words and spell them all correctly. 00;13;22;05 - 00;13;41;08 That wouldn't work. But if you're teaching creative writing, it may be acceptable to go ahead and use the AI to get some ideas about how to write and what you want to write about, but just make sure that they understand they can what they can and can't use it for. I think that's probably the biggest pitfall to avoid. 00;13;41;11 - 00;14;08;02 What's super interesting is, is I received an email that I could tell was written by generative AI, and it was because it was really clean. It didn't use the colloquialisms a normal writer would use, but it was surface level and did not have a deep enough understanding of the topic at hand. And so I went, well, this is written by the generative AI. 00;14;08;04 - 00;14;32;25 And I know that like, sometimes like I'm a formal writer, but I know when I'm writing about a topic, I'm writing about a topic in a certain way that doesn't feel like lots of words are strung together that make some amount of sense, but not a ton of sense. So I think that that's really important, like you said, of being the expert in a field, you're going to understand the topic a little bit better and, you know, generative. 00;14;32;28 - 00;15;02;22 I can be used to fix your grammar and make your sentences be a little bit more clean as opposed to someone who's using it to generate a full document that it's and they don't know what kind of prompts or endpoints to give to the A, I do actually help it write something that's going to be effective. To your point on that one, the depth of the work that they turn in will help clarify whether or not they actually understand the topic and that they can apply it. 00;15;02;26 - 00;15;25;25 So that's another good tip that I'd like to leave is make sure that they're applying that, applying what they're learning in new ways, not just to the same examples over and over. So then you can really tell they've understood the concept behind it. So one final question. If you could give advice to faculty or students, what would it be? 00;15;25;27 - 00;15;59;26 So my advice actually is to be welcoming, embrace what I can do as a tool in your in your classroom, but make sure you do so by keeping a critical eye on the potential exposures. There's bias in a lot of AI models, and you need to make sure that you recognize that bias and deal with it. Also, don't be like starstruck by, you know, this is the latest and greatest thing out there to try to avoid that. 00;15;59;26 - 00;16;30;27 The glamor of a new cool tool, but also remember that it is a tool and it's going to take practice to master it. So make sure that you start with yourself. Start by using the technologies, become comfortable with them, apply them in your discipline, and really embrace the opportunities that are out there. I think that's really good advice to practice with the tools and to make sure that you're using them correctly. 00;16;31;01 - 00;16;52;10 I think that's a really, really good piece of advice, especially for going to this new A.I. led and generated world in a lot of ways. So a big thanks to Mary for joining me and talking about A.I. and how to use it in education. If you're interested in learning more about Oracle Academy, visit Academy dot, Oracle dot com and subscribe to our podcast. 00;16;52;11 - 00;17;05;19 Thanks for listening. That wraps up this episode. Thanks for listening and stay tuned for the next Oracle Academy Tech Chat podcast.