(upbeat music) - Hello and welcome to "Prairie Pulse".

Coming up a little bit later in the show we'll see an Artifact Spotlight from the Kitson County History Center and Museum in Lake Bronson, Minnesota.

But first joining me now is Joe Kennedy.

Joe Kennedy, of course, you're from Concordia College where you're an instructional designer and academic technologist.

- Yes.

- Okay, well, welcome.

Thanks for joining us today.

But you're also one of the campuses' leaders in artificial intelligence and that's what we're here today to talk about.

But before we get into that, tell the folks a little bit about yourself, please and where you're from maybe.

- Well, I grew up with my dad in the Navy so I'm not really from anywhere.

I'm a husband, a father, a citizen.

I'm also a teacher.

And at Concordia College as an instructional designer my job is to help faculty incorporate technology to make their already good teaching even better.

- Okay, well, you are here to talk about what I'm gonna say is new.

And I don't know how new you'll qualify it, and somewhat controversial topic, AI, artificial intelligence.

I think the general public has trouble kind of understanding what it really is.

Can you maybe explain what AI is?

- Well, sure thing.

The idea of AI is to use machines to replicate what we think of as intelligence, to represent information to make decisions based on that information.

And of course, we've got ideas of AI from "2001 A Space Odyssey" and "Terminator" and movies like that.

And AI, at one sense does try to give machines the ability to reason, but since we don't really know how intelligence works in the human brain, AI takes multiple paths to try and replicate it.

What we're seeing now is a focus on what's called generative AI.

And generative AI seeks to create new things rather than just retrieve information.

And it does so by focusing on probabilities.

- Well, I may circle back around to that, but when you think about it, it's really concerning people quite a bit, you know?

So what are negative circumstances, I guess potentially misuses of AI?

- I'm glad you put it that way.

'Cause it really is, like AI itself is not a bad technology or a good technology.

It's gonna be how we use it.

The biggest concerns about the misuse of AI the most obvious ones are the creation of misinformation and disinformation, literal fake news, rather than just rhetorical fake news.

And a concern down the road is as we develop artificial intelligence if it becomes better and better, do we turn over some of our decision making and autonomy to a system?

For example, traffic lights right now are governed by a rudimentary AI system.

And so we've turned over decisions about when a traffic light should go red or green away from human control and saying every 45 seconds it'll cycle or every two minutes it will cycle.

So the long-term concern is as more and more systems adopt AI are we giving up the ability to make decisions?

But that's several, that's a ways down the road.

- Well, there's so much of this I think people don't understand it.

One of the things that I discovered a few months ago or somebody introduced me to ChatGPT.

Tell me more about that.

What is ChatGPT?

- Well, it is most people's introduction to artificial intelligence, because it is the first system that is available to anyone within internet connection.

So if you think about when graphing calculators first came out in the eighties and nineties and schools were trying to adapt those well ChatGPT is the whole world adapting to a new technology that is all of a sudden affordable.

And it is an interface that allows people to ask questions and get answers that may actually be new.

So instead of just saying, "Siri, who was the first president of the United States?"

or Googling the capital of Kansas, you can ask ChatGPT to take a whole bunch of ideas and turn them into an outline.

Or you can ask it to explain something to you as if you're a third grader.

So it goes way beyond just Googling.

On the backside, it's more of the generative AI.

It's creating new information, but it's important to note it's all done by probability.

So ChatGPT doesn't know that George Washington was the first president of the United States.

What it does know is that it read billions of pieces of data and almost every time the phrase "first president of the United States" appeared, George Washington appeared in close proximity.

So it's sort of like running the odds.

It's betting that it's right when it says George Washington was the first president.

- So what you're telling me there is ChatGPT can be wrong much like Wikipedia in the early days, it might give you the wrong information.

- Yes, it can be wrong and it can hallucinate, which is where it just completely invents things.

And that'll become fewer and fewer, but it will still always have that possibility.

It's just playing the odds.

- Well, let's go back to what are the origins of AI?

How did it get invented, if that's the right thing to say or ask?

- Well, it's not one specific thing.

It's been more of an evolution.

It's always been something ever since we've started working with computers that computer scientists have wondered is can we make this computer think like a human being?

So it's been evolving.

It's going to continue to evolve, but the reason it's on our minds right now is because one company OpenAI released this one product ChatGPT, and said to anyone if you can get to the internet, you can use AI.

And so that's why it's really hit our consciousness, but it's been evolving for decades.

- Well you told me this a few minutes ago.

I think generative AI.

So what are the differences?

What's generative AI and what's just AI?

Are they different?

- Right, so the very earliest versions of AI are embedded in computers to begin with.

In the Turing, going back to Alan Turing and the Difference Engines, the idea is if we could write hundreds, thousands, hundreds of thousands of instructions to a computer that say in this situation, do this, but in that situation do this.

If we could come up with hundreds of thousands of those we could take all the complexity of the decisions and human knowledge and turn it into a series of on or off, one or zero decisions.

And so that's where, that's what we started with for AI.

What generative AI says is, instead of requiring a computer to go down the left path or the right path, thousands of times to make a decision, let's give it a lot of data.

And based on that data it'll start to see probabilities and patterns.

And from those it can generate something new.

And we see that, it's actually even easier to see when we talk about products like Dolly or Midjourney which create pictures that have never existed before.

And that's why they're called generative.

It's brand new.

So it's easier to see when there's an actual picture, but it's doing the same thing when it gives you a textual response.

- Well, let's turn to AI and education.

What's the potential impact of generative AI on assessment in higher education?

- Oh, it's so amazing.

We can talk about just that for 20 minutes.

But to boil it down, probably three things.

First of all, we can use generative AI on the teacher side to provide better and more personalized feedback.

So the reality is, you've asked me here to talk about AI, there are all sorts of people who can talk about AI and do it really well, but if there was someone who was teaching AI, their true genius is that they can take what John is saying and the way John is thinking about it, and they can give you some feedback that will help the way you think about it, go more towards the truth.

Generative AI can help teachers do that.

It's the one thing that all of the machine systems we've got out there haven't been able to help.

So it can really help make the feedback more personal and make your learning better.

It also can make better decisions about what we should be giving you for the next step, because it's adapting to the way you think, the way you write and the knowledge you're displaying.

And because of that, this is where the real power comes in, because generative AI can help us make learning more personal and deeper, it really engenders that sense of curiosity that we have as kids.

Because now, no matter how old you are, you can keep asking why, why, why and you're not gonna exhaust the resources of the teacher because they have this other tool to help them.

- So what are the ethics of generative AI in the context of being a responsible and engaged citizen?

- Well, thanks for using that phrase, 'cause it's right there in our mission statement at Concordia.

It's almost like you planned that.

We worry first about the obvious ethics.

This is a tool that makes it easy lie.

If somebody wants to use this tool to lie, a fake news story, a fake picture, a deep fake video, they can.

And so we do worry about the ethics there, but the ethical concerns go further.

If we look at how is generative AI learning?

Well, it's learning by looking at this vast corpus of data and not everybody whose work is represented in that data gave permission for that data to be used for the machine to learn from.

And so artists like Stephen King are suing OpenAI and saying, "You're creating a product "based on my intellectual property."

And then the algorithms that take that data and teach the machine how to learn contain the human biases that we programmed into them.

So we have to worry about that as well.

- What are the likely generative AI skills students would expect to need when they graduate?

- The first one is gonna be prompt engineering.

That's on everybody's mind.

The director of our career center pointed out to me that back in March resumebuilder.com pulled a whole bunch of companies.

91% of them said, "We'll be honest, "we expect our new employees to come in "knowing how to use AI to teach our current employees."

And the most important skill is asking the right question.

This is just a tool that can help us organize our thoughts, find new things, come up with new works of art, but if we don't ask the right question it's garbage in, garbage out.

So prompt engineering, if I just had to answer in one phrase.

- But you know, are there ways that generative AI make teachers' jobs easier?

- Oh, yes.

It helps us as teachers because we can come up with better organized lesson plans, for example.

I teach my students how to use ChatGPT to create extra problems for them to work on.

So if a student knows they're really struggling with one concept, I don't have to come up with extra worksheets, extra quizzes, extra discussion prompts.

I teach them how to use one of these tools to create the prompts for them, and then they work through them and I can help them evaluate if it's right or wrong.

But it saves a lot of the administrative time in the background.

- Yeah, so it is being used in classrooms and how extensively is it being used?

You say you're using it, but... - Almost two thirds of the faculty at Concordia College at least talk about it.

Some of them like Professor Darin Ulness and a lot of the professors in our comm studies department use it very specifically.

They teach students how to use it to take what a student already knows and ask questions.

A lot of what ifs that lead the students into ways that they can better explore the questions that are really on their mind.

- You know, what's the potential of generative AI in misinformation and disinformation campaigns?

You know, thinking about elections and of course in matters of international concern.

How do you teach students and people to spot such campaigns?

Or can you?

- You can, I mean, it's really at the heart of a liberal arts education and the idea of critical thinking to begin with.

The first question that people should ask themself is does this seem believable?

Like, if you yourself are shocked when you see a photo or read a news story, you should be asking yourself, is it for real?

Because sometimes shocking things are.

I remember on 9/11 none of us believed what was happening, but it was.

But most of the time when we think I can't believe this is happening, then we're not being told the whole story.

So the first thing that people should do is ask themself is this shocking?

And then say, is there an apparent purpose?

Like when you tell someone there's been an earthquake, a lot of times you're just wanting to tell people there's been an earthquake.

But if someone comes to you and says, this candidate, I have a picture of them wearing black face or I have a picture of this candidate for office with Jeffrey Epstein.

People should immediately be thinking, well obviously if that's true, that's bad for the candidate.

So the person telling me might have a vested interest and if there appears to be a vested interest they should look further.

Honestly, with ready access to the internet, it's pretty easy to take a picture and run it through Google search, run a search for a key term in a news story.

And if a person does that and they don't immediately see lots of other stories that are independently saying the same thing, there's a good chance that it was used to be fake.

And more and more it's gonna be AI that's creating the fakes, because it can do it so much quicker than humans can.

- Yeah, we're hearing stories today about your voice can be cloned in a matter of seconds, and then with images of course.

So there's so much opportunity for misuse of AI at this point.

Is that not true?

- Unfortunately, there is a lot of opportunity and Candace King and Mr.

Beast and Tom Hanks will tell you, don't buy a product just because someone you like and follow endorses it.

Because in their three cases they didn't.

- Well talk maybe about some of the AI seminar the webinars and seminars that you've participated in and sort of what's come out of those I guess is what I'm interested in.

- Well, so at the end of November last year is when ChatGPT came out and it truly was a bright line demarking before and after.

And so the next few months most of the webinars were very much about, oh, this is going to ruin our ability to assess students because they can use ChatGPT to write their essays.

And so it's not just using Google or Chang to look up an answer, it's creating entire works of argumentation.

But then as the months went on the webinars became much more about look at all the ways we can use this tool to help students learn better.

And the ways that we can teach students to notice when someone is using this tool for nefarious purposes.

So most of the webinars now are much more focused on that because at the end of the day there's no technological solution to a human being wanting to deceive us.

- Well, you've told me many or maybe most students are using AI already, but are they concerned about the future of AI or maybe not as much.

They just, it's a reality of it's gonna be there.

- It's still too new to draw a lot of conclusions, but it appears that if you were to say, what are students most worried about in terms of large scale problems?

They're much more worried about the effects of climate change or in the United States and other countries they're much more worried about what appears to be an authoritarian creep than they are about AI.

- Well, in the most recent actors in writers strike, AI was an issue.

Did you follow that much?

Can you talk about what that was about?

- Right, the concern that SAG AFTRA and the writers Unions had was the motion picture companies and the content streamers have enough money that they could use the computing power.

It costs a lot of money to do a really good, a really good AI impersonation of someone.

But these entities had enough money, and if they wanted to they could say, well, John, we have hundreds of hours of footage of you interviewing people.

It's not that hard to create fake John to do an interview.

And so they wanted assurances that that would not happen.

Or in the worst case scenario that the studios would not say to actors who are extras, which is how they gain their entree into to film usually, and say, oh, we, we have your likeness as an extra now.

We're just gonna use that as an extra for the next a hundred crowd scenes.

And they were worried about that could have happened.

And the industry says, no, no, that wasn't our intention, but now it's actually gonna be in the contract from what we know on the outside.

- So with the crystal ball, I'm sure you've got one, no.

Are, are you concerned about AI or not as concerned as some of the more dire warnings we're hearing about, such as the end of civilization.

These kinds of predictions?

Can you comment at all on that?

- We are a long way from the apocalyptic vision of how or the machines in "The Matrix" or in "Terminator".

But I am worried about the human tendency to see a tool and want to know how they can use the tool to exert power over other people.

So that worries me much more.

- Yeah, so basically the genie's out the bottle and it's not gonna be put back in.

- Correct, yeah.

Where do you see AI in 10 years or so?

And you gotta do it real quick.

- I see it like an iceberg.

We're only going to see the most obvious things, but it's going to underlie everything that we do.

We already use it in very specialized ways like healthcare decisions very successfully and we just don't notice it.

- Well, we are out of time.

If people want more information, where can they go?

- They can go to OpenAI.com, which has resources, research and about us on their website.

Or they could contact the computer science departments at Concordia NDSU NDSCS.

- All right, thanks for joining us today.

- Thanks for the opportunity, John.

- Stay tuned for more.

(mellow music) In this Artifact Spotlight from the Kitson County History Center and Museum in Lake Bronson, Minnesota, Ed Jerome talks about his Metis ancestors who helped settle this northwest corner of Minnesota.

- Hi, this is Ed Jerome at the Kitson County Historical Museum at Lake Bronson, Minnesota.

And this is our Artifact Spotlight.

(upbeat music) My great-grandfather, Andre Jerome was born in the Red River settlement which is where Winnipeg is today.

And sometime in the early 1840s, they immigrated to Pimbina.

They were buffalo hunters, participated in the buffalo hide trade, and also making the cart trips to St. Paul.

I built a oxcart as a tribute to the ancestors that participated in the trade from the Red River Valley to St. Paul on the Mississippi.

Tried to build it a similar or replica of the originals.

They're built out of mainly oak.

And the hubs were built a elm because there's stringy wood and less chance of them splitting or breaking and they wore well.

And the stakes on the basket are ash.

Figured it takes about a couple hundred hours to build one.

That was with mostly modern tools.

1872 is one of the first settlers in Kitson County.

They raised a large family, had 10 children.

The Metis are the children of the fur traders and indigenous women from the area.

There's several generations like Andre Jerome was.

His great-grandmother would've been indigenous woman and his great-grandfather was a French furrier from Montreal.

This a pipe that was belonged to my great-grandfather Andre Jerome.

Wood from an ash tree and then the willow for the pipe stems.

In about 1983, we had a family reunion and a cousin brought it and donated it to the Kitchen County Museum.

This is a sash Metis wore.

It was originally used by the Voyagers and their canoe trips from St. Paul.

They wrapped around their waist to help prevent hernias 'cause of the heavy packs they carried.

We're proud of our ancestors.

They were opened up the country to parade.

We're proud of all our ancestors.

This is Ed Jerome, and this is our Artifact Spotlight.

- Well, that's all we have on "Prairie Pulse" this week.

And as always, thanks for watching.

(upbeat music) - [Announcer] Funded by the Minnesota Arts and Cultural Heritage Fund with money from the Vote of the People of Minnesota on November 4th, 2008, and by the members of Prairie Public.