AI's Growing Impact on Work and Society with Mark Daley

Pauline James, CEO of Anchor HR, and David Creelman, CEO of Creelman Research, are back as guest hosts to discuss the impact of AI in the Workplace. They are continuing their conversations with AI technology experts, striving to educate and support our community remain abreast of advances and considerations related to AI.

In this episode, they speak with Mark Daley, Chief AI Officer at Western University

Questions for Mark include:

• Impact on Workplace: How do you see AI impacting the future of work?
• Economic Performance: It’s been noted that organizational performance is lagging. Do you believe AI can help close this gap? 
• Implementation: How can companies leverage this technology effectively? What skills are necessary?
• Ethical Considerations: What ethical considerations should those governing this technology keep in mind when implementing AI technologies?
• Societal Impact: How can organizations support this moment in history as proactive and healthy contributors to improving society?

 About Mark Daley 

Mark Daley, Chief AI Officer at Western University and a Professor in the Department of Computer Science with cross-appointments in five other departments, The Rotman Institute of Philosophy, and The Western Institute for Neuroscience. He is also a faculty affiliate of Toronto's Vector Institute for Artificial Intelligence.

 Mark was named in the Maclean's magazine "Power List 2024" of the top 100 Canadians shaping the country and in Constellation Research's AI150, a list of top 150 top global executives leading AI transformation efforts.
 
Mark has previously served as the Vice-President (Research) at the Canadian Institute for Advanced Research(CIFAR), and Chief Digital Information Officer, Special Advisor to the President, and Associate Vice-President (Research) at Western. Mark is the past chair of Compute Ontario and serves on a number of other boards.

Chapters

• 0:03 Impact of AI on the Future
• 13:14 Navigating AI Ethics and Governance

Transcript

Speaker 1: 0:03

Welcome to the HR Chat Show, one of the world's most downloaded and shared podcasts designed for HR pros, talent execs, tech enthusiasts and business leaders. For hundreds more episodes and what's new in the world of work, subscribe to the show, follow us on social media and visit hrgazettecom and visit hrgazettecom.

Speaker 2: 0:31

Welcome to the HR Chat Podcast. I'm Pauline James, ceo of Anchor HR, and I'm thrilled to be back with David Krillman, ceo of Krillman Research, as guest hosts. We are continuing our conversations with AI technology experts as we strive to educate ourselves and support our community and remaining abreast of advances and considerations related to AI. David and I were pleased to have the chance to connect with Mark Daly. Mark is the Chief AI Officer at Western University and a professor in the Department of Computer Science, with cross-appointments in five other departments the Rotman Institute of Philosophy and the Western Institute for Neuroscience. He is also a faculty affiliate of Toronto's Vector Institute for Artificial Intelligence.

Speaker 2: 1:14

Mark was named in the Maclean's Magazine Power List 2024 of the top 100 Canadians shaping the country and in Constellation Research's AI150, a list of the top 150 global executives leading AI transformation efforts. Mark has previously served as Vice President Research at the Canadian Institute for Advanced Research and Chief Digital Information Officer. Special Advisor to the President and Associate Vice President of Research at Western Mark is the past chair of Compute Ontario and serves on a number of other boards. Mark given, you're a leading expert on AI who has been influential in supporting innovative and practical applications in the Canadian and international market. We know that you will have some tremendous insights to share for our audience. Beyond our introduction, could you take a few minutes to tell the HR chat listeners about yourself?

Speaker 3: 2:14

Sure thing. So I left high school and I went to Berklee College of Music. I was going to be a professional musician and I loved it. I had an amazing time in Boston. But I realized that I missed mathematics, which is kind of a weird thing, and I thought maybe math is a better career and music's a better hobby than doing this the other way around, because math's hard to do as a hobby, less so now than it was in 1995. I ran away from music school to become a mathematician, ended up taking some computer science too, got very interested in computation as just a fundamental process in the universe. That got me interested in the brain, took a sabbatical, did a master's degree in neuroscience after becoming a professor which is unusual but was a fantastic decision and then got involved in academic administration and sort of realized the only thing that I loved more than than doing research and teaching myself was helping lots of other people do that.

Speaker 3: 3:09

Thanks for listening to this episode of the HR Chat Podcast. If you enjoy the audio content we produce, you'll love our articles on the HR Gazette. Learn more at hrgazettecom. And now back to the show. And now back to the show.

Speaker 4: 3:26

So why do we look at the impact of AI on the workplace? Maybe I can ask for your perspective on how you see AI impacting the future of work, both in the short, mid and long term.

Speaker 3: 3:38

Fantastic question In the face of what right now, looks like an exponential change curve. I'm not even going to talk about mid and long-term, because that's sort of on the other side of the brick wall and anyone who thinks they know what's on the other side of a brick wall you should be skeptical of In the short term. I'm a big fan of strategic foresight methodologies, so I think of the short-term future in terms of scenarios, and the two most likely scenarios right now are there's one scenario where AI progress sort of plateaus and whatever we've got right now we get slightly better versions of it, but this is kind of it. And there's another scenario where we get even another two or three years of exponential improvement in the capabilities and capacities of these models, and I think it's important for any organization to be preparing for both of those worlds. So in the sort of dull future, there's still lots to do, lots of excitement, lots of things that chatbots and generative AI can do to assist us. So still lots of work to do there, and a lot of people are doing this.

Speaker 3: 4:45

A lot of organizations are doing this, which is great. My advice to them is just take five minutes a day to lift your eyes up to the horizon and ask okay, so I need to be doing this now. That's my hedge against the dull future. But what if the exciting future happens and so many of the tasks that we used to require humans to do can now be done really easily by a machine? How does that affect my business process? How does that affect my core line of business? How does that affect how I deploy humans? So many of those questions, and it's going to be unique to your industry, but take just a little bit of time to think about what might happen.

Speaker 4: 5:29

And you know it's a bit amusing how difficult it can be for people to find five minutes to lift up their eyes. If we move right to the present, what AI applications have you seen that are already deployed and are of proven value?

Speaker 3: 5:44

The biggest one everyone's using is document writing and report writing, email answering. So when you have something that is relatively prescribed almost a rote process like bureaucratic reporting, this is important. It has to be done, but do I actually need to spend a week at a typewriter writing up there? Well, no, the report's already got a standardized format. I just need to provide the data. I need to provide the oversight. But turning that data into a rote narrative is something current generation generative AI is really good at. That's just a one-off.

Speaker 3: 6:19

The reality is, the answer to that question is highly dependent on who's asking it both the role in an organization and what they're trying to accomplish, and their personal preference, because there are places where an AI could do this for me, but I actually want to do this myself, something I derive value from doing, and so it really is hugely variable, not just across sectors, but even within a medium-sized organization. The way someone working reception is using it very different than the way someone doing business development is using it, and so there is that sort of personal. How can this technology help me?

Speaker 4: 7:00

And so, at the personal level, have you seen applications that are delivering significant value? At the organizational level, it's not just a matter of how an individual is using this technology.

Speaker 3: 7:12

Of course, everyone wants to think at the organizational level. An analogy I've made before is, you know, imagining university campuses a few hundred years ago, when running water first starts to come in. Was there a chief running water officer? Was there a running water strategic plan, or did we just sort of acknowledge people doing lab experiments are going to use it differently than we use it at the hockey rink, but this is sort of a core enabling technology. And then the same is true for electricity.

Speaker 3: 7:39

Trying to control top down what the use cases are is going to be tough, and we've seen this recently in digital transformation. There's a lot of companies who, early on in the era of digital transformation, tried to create a top down digital transformation strategy, and there's now good literature in the business literature showing that almost to a one. That failed. What succeeded was understanding digital technology as a transformative technology and focusing on how do I provide this technology to everyone I serve, to my community or to my organization, educate them on how to use it, but then let the use cases come bottom up, and with AI it's similar. From an organizational level, you have the necessity of creating policy space so people know when and where they can use it. Obviously, you shouldn't be putting PII into ChatGPT because we don't have a data sharing agreement with ChatGPT, so guardrails around. What acceptable use is education on what it can do for you, but then providing an environment where individuals can experiment and feel empowered to experiment.

Speaker 3: 8:51

One of the things we saw is a lot of our staff members at Western had really creative ideas, but some of them were hesitant because they were like am I going to get in trouble if I use AI? So you've got to get over that. No, you're going to get in trouble if you share data you shouldn't share with a vendor we don't have an agreement with, but you already know that. But we're actually going to encourage any successes with generative AI this week, and so, instead of that being like I don't know if this is allowed, you're actually getting rewarded and encouraged for experimenting and then sharing with your peers what that's done for you and then to get cross-team, cross-functionality discussions. We've created communities of practice that allow for those who want to engage, listen and share what they've done.

Speaker 2: 9:48

Thank you. I like that both structured and organic approach. You've made the comment that if you're someone who loves your job exactly how it is, this next short while may be more of a struggle. I'm wondering if you could expand on that.

Speaker 3: 10:01

Yeah. So the reality is that so far, ai has continued to become more and more capable, and to really make this visceral, google's DeepMind had an AI compete in the International Math Olympiad, and that is a gathering of the very brightest young minds in mathematics. The problems are fiendishly challenging and Deep Mind's entry won a silver medal, which is a huge accomplishment. It was actually one point short of a gold, but Sir Timothy Gowers, who is one of the most important mathematicians of the 20th 21st century, was a judge, and he was interviewed on the BBC a week later about it and his conclusion was I didn't used to feel this way, but now, after seeing with my own eyes what was possible, I don't know if there's room for humans in mathematics in the future.

Speaker 3: 10:51

I think machines are going to do mathematics and I don't know how I feel about that. I feel kind of weird and it's worth listening to the interview. So that's something that used to require human intellect alone. As far as we know, dolphins don't do mathematics and we'll have to learn how to talk to them, but humans are the only entities on earth that do mathematics, and now we have machines that can do mathematics, and if you're a mediocre mathematician like me. That's really exciting because now I can prove theorems that I wouldn't be able to alone, because I've got someone to help me.

Speaker 3: 11:20

But if you've built up, all of my self-image is I'm the smartest person in mathematics and all of my joy comes from doing this one thing. The fact that a machine can now do that thing better than you is going to be very difficult, very difficult. If your life is broader than just one of a small set of tasks and you're curious and there's things you want to do, impact you want to have on the world, this is going to be incredibly empowering because things where you used to be weak, you're now going to have a partner to shore up your weaknesses there. And so, if you want your job to stay exactly the same as it's always been, the next five years are going to be very tough because everyone's job is going to be changing at a really rapid pace. If you're a lifelong learner, if you're agile, if you're enthusiastic about trying new things, and if you're focused on impact that you want to have as a person rather than process, the next five years are going to be empowering like nothing else.

Speaker 2: 12:22

Thank you, and building on the idea of shoring up, we shift gears and get your perspective on economic performance and the opportunity of AI tech to help drive and improve performance. It has been noted that organizational performance is lagging in the Canadian economy and I'm interested to hear if you believe AI can help close this gap and, if you do, if you have suggestions on how companies can leverage this tech effectively.

Speaker 3: 12:42

Yeah. So I'll start with the pessimistic reality, which is AI is not localized to Canada only, it's global. And so, in terms of, can AI assist with productivity in many different verticals in an organization? Absolutely Any organization that isn't actively adopting that right now is going to be completely outcompeted. So it's dangerous to think, oh, just because we've got AI, now Canadian companies are going to catch up. No, that's just going to be table stakes, to sort of stay where we are.

Speaker 3: 13:14

What we do have in Canada is a first mover advantage. In the very early days, the government of Canada, through NSERC and CIFAR full disclosure, former organization invested in AI research even during the two AI winters, when it was sort of out of fashion to fund AI and, as a consequence, toronto and Montreal. But Canada more broadly has an incredible depth of talent in this space. Canadian companies have access to a talent pool that goes back a couple generations now. So we've got really good professors and they have good grad students, and now there's startup companies like Cohere that have come out of the University of Toronto. We have all of that opportunity. Canadian companies that want to get more innovative have the talent pool right here locally, and a lot of Canadian kids want to stay here in Canada. I hope my kids do when they're ready.

Speaker 3: 14:07

We do have that opportunity and what's required from Canadian business is to just be willing to take on a little bit more risk. We used to be a country of risk takers, and when people sort of look back to the days of Nortel and Research in Motion, blackberry, say, well, where's the next BlackBerry? Those were risk takers who founded those organizations, and so we just need a little bit more risk appetite. We've got the talent here. We have the latent capacity. I encourage Canadian companies just take on a little bit more risk, bring in that talent, give them a bit of freedom. There's no reason why we can't catch up to where we were 30 years ago.

Speaker 2: 14:46

Thank you, and I appreciate the call it on the talent that we have within the market that we want to really tap into and leverage. In addition to that, are there skills that you see as necessary that we should be encouraging within our organizations at this juncture?

Speaker 3: 15:02

Absolutely Curiosity, self-directed learning and critical thinking. I can't predict what AI is going to be able to do five years from now. I don't think anyone can. The only safe bet is probably more than it can do right now, and so, to be able to thrive in an environment where the pace of technology is changing so quickly, you really have to have a growth-oriented, agile mindset and a real desire to engage with every day. Okay, the technology couldn't do this for me yesterday, but there's a new model. Maybe it can do it for me today.

Speaker 2: 15:35

Thank you. I could also like your perspective on ethical considerations and what those who are governing this technology within an organization should keep in mind.

Speaker 3: 15:47

There's this weird thing happening right now, and it's probably because we've anthropomorphized AI as chatbots where someone will torture ChatG GPT and get it to say something inappropriate and then post it on social media and say look what chat GPT did. And we're seeing this really right now with X has released the Grok 2, and it's got Black Forest Labs Flux image generator built into it and there's very few guardrails on it, and so people are generating images that are sort of wildly inappropriate and the reaction is look what X's AI did. And I think that's an absurd reaction. It's as absurd as me writing hate speech with a ballpoint pen and then saying Bic is a terrible company. How could they make a pen that would allow me to do that? How could they make a pen that would allow me to do that?

Speaker 3: 16:41

So organizations already have governance and policy in place to govern human behavior, and that's true of governments too, right, like at the highest level, and we need to get over this.

Speaker 3: 16:51

I can blame the technology for something that a human made it do. If you produced a terrible image with Grok 2, that's not on the AI. The AI is just a tool that is on you, the human being who prompted that, and so in organizations you have to remind people just because it's a really cool technology it does more than a ballpoint pen could, but it's still a technology, it's still a tool. You, the user, the human agent, are ultimately accountable for what you do with that tool, and if you do something inappropriate with that tool, that's on you. You don't get to turn around and say, well, no, chat, gpt made me do it. So I think the good news is we already know how to govern human behavior, we know how to create good policy, and none of that changes in the AI era. All we have to do is get over this anthropomorphization of the tool and allowing people to sort of blame the tool for something that is actually their choice and their decision.

Speaker 2: 17:49

Thank you. Related to that, I would also welcome your thoughts on how we're going to balance the need for an appropriate level of governance and I appreciate your caveats around what that looks like and a reminder that we have a lot of tools and structures in place that continue to apply, but with that, the governance that's required, with the balance of the need to keep pace.

Speaker 3: 18:09

So the approach we are taking at Western is not to have a monolithic AI policy.

Speaker 3: 18:15

What we are doing is, where it's appropriate, updating existing policies for the AI age. So, obviously, use of AI in the classroom is on the minds of instructors and students and instead of creating an AI policy, what we did is I went back with our vice provost academic and our associate dean's academic and we made changes to existing academic policy updated for the AI age, and some policies don't need to be updated. Our data governance policies, shockingly, don't need a lot of updating for the AI age because data governance is still data governance and just because the piece of software is AI instead of some other piece of software really doesn't change all that much. So, by relying on what we've already built, by focusing on human behavior and structural expectations and not othering AI and thinking, we need to do a whole separate set of policies and rules, that actually allows us to be more agile, it allows us to build on what we already have and it saves us from the trap of creating parallel streams of governance, which is a nightmare for everyone.

Speaker 4: 19:24

Well, this has been great. I'd like to close by talking about the societal impact. So we're going to give you an opportunity to wave a magic wand and tell us how you'd like organizations to show up and support this movement, to be proactive and healthy contributors to improving society.

Speaker 3: 19:42

So I think it's going to be different for every organization. It's going to depend on your mission. I think if you are an organization that is focused on the impact you want to have on the world, this is going to be tremendously empowering. It's going to be disruptive, things will change, things will be different, but I think the agency that each of us has as an individual will be amplified by this technology.

Speaker 3: 20:06

Things that would have taken me a week to do completely manually, I'll be able to automate pieces of that and maybe do that in half a day now, and so that means that the value of human time is going to be much higher.

Speaker 3: 20:23

My ability to change the world around me is going to be higher, and then when you start assembling humans into organizations again, you get that same scaling effect of the technology and at the same time as people's jobs are changing, you are going to have those colleagues like I just love my job exactly the way it is. Organizations are going to have to be tremendously empathetic and sympathetic to humans who feel disrupted, and they will need tools to educate and adapt to a changing reality. And for sure, humans can't adapt, but some of us won't want to, and that's okay, and we have to acknowledge that. And so sure humans can't adapt, but some of us won't want to, and that's okay, and we have to acknowledge that. And so I think we do have a duty as organizations to those we serve, to our communities, to help them through a transition. That is really exciting but is also, for some, going to be challenging.

Speaker 4: 21:13

Yeah, and helping people through a transition falls right into the bailiwick of our main audience, which is HR professionals.

Speaker 3: 21:20

Absolutely Ironically, AI is going to make the need for HR even more apparent than it was pre-AI.

Speaker 2: 21:29

Thank you Really appreciate the insights. You've offered the practical guidance. This has been a short conversation. If someone would like to learn more about your work, how are they best to do so?

Speaker 3: 21:42

If you want to see what we're doing at Western aiuwoca is our one-stop shop for that and if you want to know what I'm thinking, you can follow me on LinkedIn or my sub stack.

Speaker 2: 21:55

Thank you Really grateful for your time.

Speaker 3: 21:58

Thanks so much.

Speaker 1: 22:02

Thanks for listening to the HR Chat Show. If you enjoyed this episode, why not subscribe and listen to some of the hundreds of episodes published by HR Gazette and remember for what's new in the world of work? Subscribe to the show, follow us on social media and visit HRGazettecom.