AI for everyone: how to open up public debate

Who does AI really serve in the end? Is it the people? The companies that profit from this technology? Or is the truth more nuanced? On this episode of the Swiss Connection podcast, Swissinfo science journalist Sara Ibrahim talks to Tammy Mackenzie, co-founder of The Aula Fellowship, a think tank and NGO with the mission to ensure that everyone can access the conversation on how AI is being implemented in society.

Journalists: Sara Ibrahim
Host: Jo Fahy
Audio editor: Michele Andina
Distribution and Marketing: Xin Zhang

SWI swissinfo.ch is a public service media company based in Bern, Switzerland. 

Transcript

Cold Open On AI Power

Jo Fahy 0:00

SwissInfo podcasts.

SPEAKER_02 0:03

I met Tammy Mackenzie, whose concern about the concentration of power in AI.

SPEAKER_01 0:10

I was a kid who was born cheerful and optimistic. The only thing that ever got me angry was injustice. If somebody was mean to somebody else in the class, then I'd, what are you doing? And I'd get involved.

SPEAKER_02 0:19

Tammy also leads a think tank, an NGO, with the mission to ensure that everyone can access the conversation on how AI is being implemented in society.

SPEAKER_01 0:31

There is existential threat to multiplying the ability of bad actors to act. There's existential threat to connecting automated decision-making systems to each other. You can make change. And the reason you can't do it alone is because it's all of society and it shouldn't be changed by one person alone. But it does change when we get together. It's how we've created society so far.

The Promise And Fear Of AI

Jo Fahy 0:58

Artificial intelligence is everywhere right now, and it's hard to predict how far it will go. It enables us to do certain things faster and better than before, but it also creates fears. Some people have lost their jobs to AI. Some are worried about AI-generated content and misinformation harming our democracies. But who does AI really serve in the end? Is it the people? The companies that profit from this technology? Or is the truth perhaps a bit more nuanced?

Davos AI House And Tammy

Jo Fahy 1:36

Hello, my name is Joe Fay, and you're listening to the Swiss Connection Science Podcast. And today I have with me in the studio our science journalist Sara Ibrahim. Thanks for joining us, Sarah. Hello. Sara, you visited Davos during the World Economic Forum annual meeting earlier this year to find out about current AI discussions. Tell me about what you heard and who you met.

SPEAKER_02 1:58

I was at the AI House in Davos, which is one of the main um venues for AI discussions. And their um different debates went on on a number of different topics from uh security of AI to how to make it uh more um sustainable and inclusive. And at one of these debates I met Tammy Mackenzie, which is a very unique and special um figure to meet in such a high-level place.

Jo Fahy 2:37

Okay, and why is that?

SPEAKER_02 2:39

Tammy is a professional with a background in accounting, but also a technologist and entrepreneur who's concerned about the concentration of power in AI. She has spent years trying to raise awareness through research and showing data and promoting campaigns. Um, and she also leads the Aula Fellowship, a think tank based in Montreal, Canada, that advocates for more inclusive AI. Okay.

Jo Fahy 3:11

So let's listen to what Tammy Mackenzie told you. And just a small technical aside here, Sarah had to record her interviews under very noisy conditions. So we did our best to improve the quality, but it's not always perfect.

SPEAKER_00 3:24

Welcome to the AI House Davos. Our next panel on the summit stage will begin in 15 minutes. If you are a confirmed guest, please proceed upstairs to the check-in and present your QR code for full access badge.

SPEAKER_03 3:40

Do you want to tell me a bit about you? What's your story? How you ended up in AI? Will you tell me more? I will. Um, thanks for asking.

Why AI Governance Lacks Legitimacy

SPEAKER_01 3:51

A few years before GPT came out, Montreal is a center for governance and for AI research. So that center there started uh work on the Montreal Declaration, and the people at the United Nations, the people who are working on this in the large organizations, OECD, UNESCO, and the institutes, recognize that there's a problem in AI governance because the people who have to make the decisions were sitting at tables with technologists, the technology companies, and nobody else. Uh, to their credit, both those people, those groups of people know that that's a problem, but they don't have the legitimacy to govern without the people having their say and being at the table. The trouble is, of course, uh AI literacy, access to information, access to the time and the luxury to be able to have conversations about esoteric things like AI. It's not very concrete. So they had identified plenty of problems, and they sent a call out for people who could help uh in terms of impact. They looked through a few thousand applications and picked 40 people to come bring them into Montreal, and we had a training there. Engineers, they had lawyers, they had policy people, they had academics. It's AI, it's interdisciplinary. You need all those people at the conversation. In fact, 40 is not even enough. You always need lots and lots of people at the conversation. And that just doesn't just mean specialists. The trouble is, is if you're trying to teach the engineer to speak to the lawyer and the engineer to speak to the accountant and the accountant to speak academic, it becomes very rapidly almost impossible, just like translating languages. So they also knew that. Um and in this case, a few of us were also identified as people who can facilitate. So that means the engineer doesn't have to learn to speak lawyer because I can translate fooder. So I'm not the only one. They made this group of us and they made us think tank through that week. And when we finished that training, which was short, it was about one week. They've done it now, I think, three more times. We've become the unofficial alumni organization because we walked out of there going, okay, yes, we're in. We want to collaborate, we know how, we think it's great, we've got projects that we want to do. Um, how are we gonna do it? And I stepped in there again because I have the management experience to organize groups and small businesses. I've been, I was 25 years or so a business consultant.

SPEAKER_02 6:03

That was the start of the Aula Fellowship.

Aula Fellowship And Inclusive Mission

SPEAKER_02 6:06

I think Thank an NGO with the mission to ensure that everyone can access the conversation on how AI is being implemented in society. Today, the organization already has over a hundred fellows, people from various disciplines and regions of the earth.

SPEAKER_01 6:24

Our mission is to ensure that everyone can access the conversation on AI. That's distinct from a think tank saying that we're providing advice on AI. It's not the same thing. It means that we use our specialties to make sure that everyone can access the conversation on AI. And by everyone, we don't, it's not a community effort, it's an all-of-society effort. So it means setting the table for all the people who are involved and the organizations and the institutions, but it's a common table. And the point is when we were doing those convenings, our fellows who are out in the world ensuring that people can access the conversation, what that looks like concretely, is that whatever they're working on, whether it's policy or law or research, whatever it is they're working on, um, they're already making sure that the people around them are having that conversation. They're having these conversations regularly about AI.

Four Reasons People Talk AI

SPEAKER_01 7:10

So we started to assemble that information, and what we found is that there's only four reasons why people get around have conversations about AI, and they're kind of progressive. The first one is AI literacy. They want to know what it is, how it works. The second one is how can they live with it in society, what's it have to do with them? The third one is how can they work with it? Like how is it affecting their work? How is how can it be used to maximize and optimize for? And the fourth one's shaping AI. So you can see that they're somewhat progressive. As you learn the first one, you might get interested in in deeper knowledge. So every convening, every conversation on AI comes under one of those four headings.

Jo Fahy 7:43

So a quick interruption here. This group isn't providing any solutions or giving advice, is it?

SPEAKER_02 7:49

Right. They want to facilitate discussions around small and big questions. The thing about AI, as I understand it, is that it's too complex. And to figure out what we want from it, people have to start having discussions and asking questions.

SPEAKER_01 8:06

Sometimes they're asking simple questions that can have an answer. That doesn't mean they have an answer, but it means it can be known. We can figure it out. Um, one example I use is will we one day be able to uh go interstellar transport, like get a ship out of the yeah. That's actually a simple question. The answer is yes. Engineering-wise, there's nothing blocking it. Uh we have knowledge to do it, it might be slow, it might take another 50 years or 100 years. But the answer, it's not complicated. The answer's not known yet, but it's it's still a fairly straightforward question. Um there's the other category of questions which should almost not be called questions, because they're so complex that there is

Simple Questions Versus Hard Ones

SPEAKER_01 8:42

no simple answer in visa, which you could not even imagine there being a simple answer. It's a complex circumstance. Um, who decides how AI guardrails go in? That is not a simple question. Who should decide? How should we decide? These are the types of questions that we need people to sit down and discuss about. So that's why we're doing that. We're ensuring that people can access the conversation on AI, use our specialties to make sure that the connections are being made between people in those decision-making circumstances so that they get the legitimacy to make good decisions. Uh the political legitimacy from people getting together and saying, this is how we want it, this is how we don't want it, creates for both businesses and for governance the legitimacy to act. So businesses will fill in that void, use that power that's been handed to them, the legitimacy of a group of people saying it. Um, and they will use that to make better products that are in line with what the will is of those groups.

A Lifelong Fight Against Injustice

SPEAKER_02 9:34

Tammy is now 50. Her engagement for a fairer and better world started a long time ago, during her childhood.

SPEAKER_01 9:42

I was a kid who was born cheerful and optimistic. The only thing that ever got me angry was injustice. If somebody was mean to somebody else in the class, then I'd, what are you doing? And I'd get involved from the time I was little. So I started thinking I wanted to be a lawyer for many years. I did pre-law, I did the LSATS, the whole thing. And as time went on, I realized that what lawyers do is not what I want to do. It's important. We need it in society. People need a voice, they need advocacy, but but that's not the same as pick your cause and work for it.

SPEAKER_02 10:10

Tammy built a successful finance career, advising small and medium-sized businesses. Along her professional career, she was involved in civil rights, anti-racism, and other community advocacy efforts. When she was about 30, she shifted her focus to climate action, helping companies work more sustainably. She earned an MBA and explored how improved accounting methods, business ethics, and AI could better measure social and environmental impact.

Existential Risk And Compounded Harm

SPEAKER_03 10:44

You said before you've been always kind of driven by the injustice that you want. So, what what is the injustice that you see in AI that you want to fight?

SPEAKER_01 10:56

The first one that I wanted to fight was this. There's two great people, thinkers in AI. There's many, of course. One of them is the scientist Joshua Vengio, who's the the founder of Mila, who uh started saying there's danger, there's existential threat. There is. This is non-deniable. We should not be denying that. There is existential threat to multiplying the ability of bad actors to act. There's existential threat to connecting automated decision-making systems to each other. So the effect it has on marginalized people compounds. So if you marginalize people with one algorithm and then that same group of people is applied to a different algorithm that also marginalizes people, it's not different people that are marginalized. It's almost always the same people that are marginalized, and they're more marginalized by this effect. So automating that just makes it much worse. So it's very much alarming. It means it becomes alarming in terms of um few events that we've seen as well that are not esoteric. We've connected two uh the two Google translation bots were connected together. And within minutes, they had invented their own language and both systems improved. They still don't know how that happened. So we have these effects that we can't predict that are affecting all of society, that by definition makes it existential to our society but also to our security. Um that so that was one aspect, that's Yashua Dangeo. Then you have uh Timnet Gebru, I don't know if you're aware of, who's saying that's a waste of time. It's already doing carves, it's already being used. Again, algorithmic decision making is way older than LLMs, um, but is very much uh a function of AI. And so she's right also. What struck me and what where I saw the injustice, the thing that we could do is that we shouldn't be, they're not mutually exclusive, those two problems. And it should never have been a debate or a fight between them. They're both correct. We have to deal with both of those things. And then I thought, oh, that's a huge job, right? Um but I do know how we can change things. Like I've been studying that, and it's it's not that tattoo, it's not imaginary, it's not um it's not Santa Claus or or wish for thinking, it's not even aspirational thinking. We just observe that's how society changes. And it heartbreaks me. I love tech. I've read every science fiction you can imagine. I stopped about 10 years ago, but you know, um, the future of science, like I think it's made us so much more powerful. We've we've got longer lifespans, healthier lifespans, um, just birth control. How about that for a technology that changed the world, you know? So I'm a believer that technology makes great change possible. And then to see it taken and not cared for, and uh everybody, every man for himself, uh let's try to get money out of it and squeeze it as much as we can. I'm a climate sci-right, a climate activist for 15 years. It's a major climate problem, and I'm not scared of dealing with big problems.

Neutral Convenings And Company Incentives

SPEAKER_03 13:39

We are here at WEF, and I feel sometimes out of my comfort zone here because you know, we are in a place that's where the creme de la creme, so just very rich people, powerful people, they can afford to come or they are invited to come. Yeah. So you you basically you want also to to connect with those big companies, big tech companies, because you think you can convince this pair that no no, I'm not trying to convince them that their business is going in the wrong direction. No, no, it's not activism.

SPEAKER_01 14:15

No, no, I'm not trying to convince them or anything. I have no answers. I'm trying to bring them to a conversation with everybody else. So there'll be veterans at the conversation, a red cross at the conversation, the mothers of Ukraine at the conversation. You see, I'm not responsible for the results of the conversation, just making sure it happens. And that it's not at Palantir's table or Microsoft's table, that it's a neutral space, a Swiss space. Come, talk, see together what you want from this.

SPEAKER_03 14:43

But my my question is how can then change for the people this legitimacy? Because at the end of the day, if in any case Palantir or anybody else understands that okay, I can go to this table, discuss, but still I'm still doing my business as usual because it uh gives me more money and more power, then why should I change?

SPEAKER_01 15:09

Because they think they'll get more money and power doing one thing than another. It might be everybody agreeing, uh, you know what, Palantir was right after all, you're doing fine, keep at it. Right now they definitely have legitimacy. People give them legitimacy because they're government and they buy their product. Uh-huh. That legitimacy is only conditional on the government saying it's a good idea, we should do this thing. So they know, just like government people know, companies know, that they need to keep legitimacy about their product for them to continue to work. Yeah. So they have to come to things like this. And that doesn't mean they have to leverage it, but they have to be part of these conversations or they lose legitimacy. So they come because otherwise they lose legitimacy. And once they're there, that's when they learn for their own benefit what type of products are gonna work that's gonna give them an advantage because they can modify their product based on what their constituents, their governments who hired them and what stated that they even want.

SPEAKER_02 16:02

It's it's not big changes, it's little changes.

Change Happens When We Gather

SPEAKER_02 16:07

It was nice to see Tammy's optimism. I asked her what key message she would like to give people.

SPEAKER_01 16:14

You can make change, and it the reason you can't do it alone is because it's all of society and it shouldn't be changed by one person alone. But it does change when we get together. It's the only thing that ever changes things. When enough people say they want something, it becomes the thing that happens. It's how we've created society so far. What we're facing right now, like companies pushing back against uh climate change, for example, saying, oh, we don't need to meet our climate cards. That's a bunch of people, the people who run the company and their clients deciding together that this is how it should be. So since that's true for the terrible things, we have to remember it's also true for the good things. It's why we have traffic laws, it's why why we have police service, it's why we have dental care, you know, it's why we have insurance, it's why we have fire insurance. You have to kind of have the faith that if the people want it, enough people will gather around it and then it will happen. When the people come, that's for sure. Things change. The only thing I can say is we make the table, and if people come, then things will change from it. I can't tell you how. That's not my job.

Next Episode And How To Support

Jo Fahy 17:19

Thank you so much, Sarah, for this interesting report, which I think can give us all a bit of hope that things will turn out okay with AI in the end. You're welcome.

SPEAKER_02 17:28

It was a pleasure.

Jo Fahy 17:32

Coming up in the next episode of the Swiss Connection Science Podcast, we'll visit the Swiss Federal Laboratories for Materials Science and Technology, the EMPA, where a young researcher is working on developing a green synthetic fuel that can be used in ordinary cars. Hit subscribe now in your podcast feed to make sure you don't miss out. For more science stories, visit our website swisinfo.ch. And you can help others to find our podcast by leaving us a five-star review or share our podcast with a friend. Today's episode was recorded and edited by our science and video journalist Michaeli and Dina. I'm Joe Fay, thanks for listening.