#37: AI for Good

Ian Bowie
Hello and welcome to AI Unfiltered with me Ian Bowie and our resident expert Michael Stormbom, where we will be talking about everything to do with AI in our modern digital society and what the future holds for all of us.

Michael Stormbom
In this episode, we’ll explore some examples of using AI for the common good.

Ian Bowie
And it would seem actually that there is quite a lot of good that AI can do.

Michael Stormbom
Well for sure – we have a tendency to focus on all the negative aspects of AI on this podcast, but definitely there are

Ian Bowie
No I don’t think we do. I think we well to some extent. But no, I mean, we find the holes in some people’s thinking,

Michael Stormbom
Indeed and a lot of hype and I think it’s a good thing to look beyond the hype.

Ian Bowie
Yeah. Absolutely. Yeah. But yes, AI for good. I mean, you’ve got quite a few examples, I believe.

Michael Stormbom
indeed. And then we can start talking about AI for wildlife conservation, or wildlife monitoring, rather,

Ian Bowie
I think you have a particular company in mind that you’ve met Haven’t you?

Michael Stormbom
There’s a Polish company called epsilon. We have no association with them. We just want to highlight a company using AI for good. We’ll be sure to highlight other companies too in this podcast as we encounter them. And they have developed an app called Mbasa AI, which is used for wildlife conservation and tracking and the national parks in Gabon in Central Africa, for example. They use computer vision to analyze millions and millions of images from cameras stationed throughout the parks. And it would take years and years for a human being to go through these images and analyze them, which makes timely intervention where needed rather difficult. Also more timely analysis, for example, finding the science of disease in images and providing vital information on the animal populations of endangered animals for conservation purposes.

Ian Bowie
I suppose it can also monitor the effects of changing habitats as well. Yeah, sure. Yeah. If we get some unscrupulous loggers chopping down a whole load of forest, then we can immediately see what effects that has on the resident animal population.

Michael Stormbom
In a national park in a Zambia. AI is used to help prevent poaching by helping in surveillance using cameras and computer vision, creating alarms when actions needs to be taken to trial this doesn’t, you know, replace or take away human responsibility.

Ian Bowie
Well, you can have you can have a lot of information about poachers and poaching patterns, but you still got to put the rangers on the ground to go and sort out the poachers themselves.

Michael Stormbom
Indeed. So there’s no, well maybe that’s the next step. You will have those drones that go out and take out the poachers.

Ian Bowie
Well, yeah, that will…

Michael Stormbom
That will be the next step sort of automatic.

Ian Bowie
Instant zapping of poachers. Yes, that will be very cool, or just give the monkeys a machine gun. And then when the poachers come in, they’re gonna get a real shock on me.

Michael Stormbom
Yeah, so this particular example was around African elephants who don’t have opposable thumbs, but…

Ian Bowie
You could put heat sensing 50 millimeter machine guns on top of the elephant, that automatically…

Michael Stormbom
The only thing that can stop a poacher with a gun is a monkey with a gun.

Ian Bowie
It’s like that scene out of Crocodile Dundee. He got dressed up in a in a kangaroo skin when there were these nutters from the town, shooting kangaroos. And then he popped up in the headlights looking like a kangaroo with with a rifle. scared the living daylights out of them.

Michael Stormbom
So they could have a robot like, a robotic elephant hiding in the mis.

Ian Bowie
I mean, it’s pretty good. I mean, it’s like you said, you know, perhaps one day with poachers, you actually deploy these drones? Yeah, why not? But of course, the problem with poaching is of course, you’ve got to look, it’s like crime, which it is a crime, but you’ve got to look into the causes of it. And you know these poachers for the most part, they’re just poor farmers.

Michael Stormbom
No one is doing it for the joy, okay some hunters do, but.

Ian Bowie
Yeah, well, some of them do. Yeah, but predominantly white privileged, overweight males, who are just doing it for sport and paying for the privilege. Yeah, scumbags. The actual local poachers on the ground are often. They’ve got no choice. They’re trying to feed their families.

Michael Stormbom
So I mean you always have to look at the big picture in that sense. Yes.

Ian Bowie
Absolutely. Yeah. I think that’s probably where actually most governments fail, is they they they look at the crime was but they’re not looking so much at the causes of the crimes. And dealing with that.

Michael Stormbom
No but also protecting against poaching of, of elephants and also against illegal fishing, using a wide range of different types of sensors as well. So thermal, thermonuclear, I was about to say, but thermal imaging. Yeah. So I mean, you can cover a lot more ground with these technologies than a human.

Ian Bowie
Well, I suppose they act more quickly as well.

Michael Stormbom
Yes, because you can, you can go straight to where you need to go. I think that’s the.. Yeah.

Ian Bowie
Just instant zapping of bad guys. I mean, I suppose there’s AI for good as well. Making the streets of cities safer. Because you’ve got, you know, your facial recognition.

Michael Stormbom
We were coming into the Robocop territory here with that.

Ian Bowie
Well, yeah, but I mean, personally, you know, if you’re not a bad guy, and as long as AI facial recognition is let’s say 99.999% accurate, and why wouldn’t we want it on our side working for us to take the bad guys off the streets?

Michael Stormbom
What crime does it protect against?

Ian Bowie
Well, I mean, for example, if the police have got somebody that they know is wanted.

Michael Stormbom
Wanted for what? Well, I’ve been like oh, my God, that guy was smoking a joint over there. I think that’s, then you send Robocop on it, that soundes excessive.

Ian Bowie
Yeah. All right. Okay. Fair enough. I mean,

Michael Stormbom
It’s still about what laws are on the books and how they’re enforced as well so.

Ian Bowie
The seriousness of the offense. And yes, yeah. Of the offense and how many offenses the offender has committed and everything else. Well, yeah, you’d have to put all that into the system, wouldn’t you?

Michael Stormbom
Yeah. Yeah. But again, you can’t. They’re still human beings in the process and that needs to be there. There are human beings who decide the law at the moment. Well, we’re coming back to the AI dictator episode.

Ian Bowie
I was about just thinking about AI for good. Yeah. And AI for good automatically means taking out the bad doesn’t it.

Michael Stormbom
Or mitigating the effects of the bad? Maybe we’re perhaps getting a bit off topic if we’re now talking about the surveillance state and monitoring human activity. If there are people who just want to break the law, then the only solution is to have no laws.

Ian Bowie
Oh, my goodness. All right, great. And then you can do whatever you want. That’s called anarchy.

Michael Stormbom
Yeah, but then there are no laws for them to break is there.

Ian Bowie
Well, yeah, no, but then then then then you have to fall back on basic human morality, which is also highly questionable. Actually. There’s there’s a point if we talk about AI for good. What about AI for good in terms of controlling traffic? For example, what about if cars would have some kind of sensor that would detect if a driver has driven through a red light and then they could automatically kill the car engine or possibly even the driver.

Michael Stormbom
I feel like we’re taking to the death penalty a bit too frequently.

Ian Bowie
Driving through a red light instant death penalty. Yeah, I mean, I’ve always had the idea that for example, eventually technology is going to come to cars that for example, if you break the speed limit in the city, somehow the car will brought that back to the central traffic control center, which will automatically generate a ticket.

Michael Stormbom
That could be a possibility, but you don’t even need to have a thing in the car. I mean, they’re all are those speed cameras which work the way they work? Everybody knows where they are. Yeah, but I mean, if they’re everywhere, then there’s no place to hide. Or you could have sensors in the roads.

Ian Bowie
Yeah, that would be quite good.

Michael Stormbom
Because I mean, that’s in principle, how many traffic lights already work. There are sensors in the road to determine

Ian Bowie
If you’re driving too fast big spike coming from out of the road, yes. Yeah.

Michael Stormbom
Yeah, I was more thinking about using the sensors to optimize traffic flow, but… optimizing road safety.

Ian Bowie
But again, we’re talking about first world problems here. I mean, I think the developing world is where perhaps the more pressing problems exist. I mean, for example, food security,

Michael Stormbom
Food security. So I mean, optimizing food distribution or…

Ian Bowie
food distribution, but also farming practices. Yeah, absolutely. You know that you have the optimal amount of seeds spread but you’re not using too much or too little fertilizer. Yeah.

Michael Stormbom
One of the sort of example apps of TensorFlow, which is one of these frameworks for machine learning. So that’s, that’s an app where you classify, so you take pictures of your plants, and then it classify whether it has a particular type of a disease or not. And then if it has a disease, then you can do something about it.

Ian Bowie
Yeah, no, perfect.

Michael Stormbom
For example. So I mean, it’s a pretty straightforward image classification. Yeah. Yep.

Ian Bowie
And there’s also AI for water security, thinking that it can monitor how water is being used.

Michael Stormbom
Yes. And then it can be optimised, absolutely, yep. Yeah.

Ian Bowie
I mean, I suppose you could actually have aI enabled smart sensors in people’s homes as well. So you can sort of monitor how much water individual households are using.

Michael Stormbom
That’s how I get my water bill, at least it’s based exactly on how much I’ve used.

Ian Bowie
Yeah, the water in this association is averaged out over the whole association. But of course, you have got people who have a little bit complained because you’ve got a household have one or two, particularly older people. And then you’ve got a young household of mum, dad and two or three kids. And of course, you automatically know that they use a lot more water than the old people.

Michael Stormbom
So in my building before the renovation, it was a fixed price per person.

Ian Bowie
That’s what we have now. I think, well, of course, it will be fair. Yeah. But it’s not just that, I mean, it’s also like for example, an awful lot of water is is last through leakage. So you know, having sensors that can detect leaks, and then the quantity absolutely, yeah,

Michael Stormbom
Yeah. Well, coming back to critters, so I mean, save the bees. The world bee project. I mean, it’s been said that if the beast goes extinct then we’re totally fucked.

Ian Bowie
Well, we are, they pollinate everything. Yeah. And no more honey is not very funny.

Michael Stormbom
No, indeed, not so they’re a rather important species in the ecosystem. They go away, then we are…

Ian Bowie
Without the bees, hat am I going to put on my heart on my porridge in the morning? No, honey.

Michael Stormbom
Yeah, it’s not, I think other aspects. Yes.

Ian Bowie
But no, they pollinate everything. I mean, forget your apples. No, bees can’t pollinate the blossom.

Michael Stormbom
Forget quite a bit of food.

Ian Bowie
Yeah, absolutely. No we’re screwed. Save the bees people.

Michael Stormbom
Yes. Basically data gathering through various means. You’re gonna have to basically analyze, analyze hives and patterns and trends. No, no, no. Is it enough to save the bees, I guess.

Ian Bowie
Well, yeah. But again, you said that analyze and collect data and analyze data. There needs to be action.

Michael Stormbom
Yes. I think that’s the way we did research for this episode. I think it’s a lot about just collecting data, analyzing the data, but then it’s still a human being the need to do something with the data so there’s nothing that says automatic bee saver.

Ian Bowie
Oh, human beings need to act. Yes. That is a problem though. And you know, okay, you can you can have all the data in the world and you can even have analysts who analyze and formulate plans for action, but without collective mass action on behalf of, you know, whole societies. There will be no change.

Michael Stormbom
Yeah, well, I mean, climate change is case in point because yeah, we very well damn know what we need to do.

Ian Bowie
Yes, but nobody’s doing it. Yeah, or not enough people to do.

Michael Stormbom
This is the problem. Another example where as you were saying, Well, climate does different climate models that we hear about. So using AI to, to create those, but again, if the human beings don’t act on the information they get then…

Ian Bowie
They don’t, I mean, well, not enough human beings act, I mean, even the latest COP27 was a bit of a cop out, wasn’t it?

Michael Stormbom
Yeah. Well, but I mean, the same applies to a world hunger somewhere we can use all these technologies to to analyze, but then we do nothing. But if we don’t do anything with it, I don’t think that’s the… So I mean, there’s certainly a role for AI in helping us do good things, but it’s still up to us to do the good.

Ian Bowie
I mean, another good example is that we know that on mass as a species, we should move more towards a plant based diet than a meat based diet.

Michael Stormbom
Yeah, there’s no secret there. I mean, no, no, we know that cows are very bad for the environment.

Ian Bowie
Yet, meat consumption is only increasing.

Michael Stormbom
Another thing is, of course, the world population. We just passed 8 billion.

Ian Bowie
Well, yeah, but I think we know that there are just far too many people in the world. I mean, that is the single biggest problem with the whole world. It’s not lack of resources. You know, lack of access to clean water, food, etc. Simply there are far too many people on the planet.

Michael Stormbom
Yeah, yes, that’s it. I mean, we have this cost of living crisis, and of course the price of electricity, it hits a lot of people very badly, but I guess one positive aspect that we can find in that is that people are maybe a bit more conscious about their electricity usage and investing in monitoring your electricity usage.

Ian Bowie
Well, that’s true. Yeah.

Michael Stormbom
So I mean, even I mean, yes, of course, for many people, it’s about living from hand to mouth. I’m absolutely not saying that. But if, if something good can come out of that and maybe more consciousness about how much electricity and opt to minimize it. I mean it because it’s very much something you can do you can you can get to for example, what he said electricity price at this point, and you can automate it so to control your your heaters, for example. Yeah, absolutely. Yeah. Or push it. So that sort of thing you can do.

Ian Bowie
Well, yeah, I mean, it makes sense. Yeah, of course, it does.

Michael Stormbom
To what extent it will lower the electricity usage and other matter, but.

Ian Bowie
Well, of course, I mean, it’s quite difficult, especially as we’re going into winter. I mean, of course, it can control things like you know, lights. But I mean, if you think about, you know, how warm you keep your house in winter, then you know, I mean there are there are limits as to what you can do with the heating. Obviously, you know, if you’ve got monitors that have detected a lack of human activity in a room for a certain number of hours or whatever, it could perhaps, you know, turn the heating right down early, and then suddenly, if you know, there’s suddenly humans in the room again, the heating can go back up again. But you know, in the rooms where there are human beings, especially when they’re older people, you know, they have to have the heating at a certain level.

Michael Stormbom
That is absolutely true. There are limits to what you can do.

Ian Bowie
But then of course, having said that, you know, if you’ve got a big enough house, you might decide, okay, we’re not going to use certain rooms and just shut the heating off in those rooms.

Michael Stormbom
In essence optimizing.

Ian Bowie
Yeah, absolutely. As they did back in the day.

Michael Stormbom
Yeah. We’ve only scratched the surface in this episode. AI and accessibility is an important topic, for example, that we have discussed in previous episodes and which we will surely come back to in future episodes as well. But yes, I mean, there’s definitely tons of potential to use AI for good and I mean, for example, we were talking about screenings and wildlife monitoring, conservation. Yeah. Food safety.

Ian Bowie
Food safety, water safety, and also access as well.

Michael Stormbom
But AI is just a tool. And it’s up to human beings to use that tool for good.

Ian Bowie
You’ve been listening to me, Ian Bowie, and my colleague, Michael Stormbom. On AI Unfiltered and for more episodes. Please go to aiunfiltered.com. Thank you.

Transcribed by https://otter.ai