# Will AI Make Universities Complacent? - Nabil Beitinjaneh

Episode 10 of AI with Sach, a podcast hosted by Sacha Gucciardo in Montréal, features Nabil Beitinjaneh, faculty Lecturer and Academic Program Coordinator, McGill University School of Continuing Studies. A McGill educator explains how AI changes learning, which skills matter, and why job displacement and weak regulation need collective answers.

- Show: AI with Sach, hosted by Sacha Gucciardo in Montréal (https://aiwithsach.com/)
- Episode: 10
- Published: 2024-09-09
- Length: 65 minutes
- Language: English
- Page: https://aiwithsach.com/episodes/10-will-ai-make-universities-complacent-nabil-beitinj/
- Video: https://www.youtube.com/watch?v=s-gC9irOk9o

## Guests

- Nabil Beitinjaneh: Faculty Lecturer and Academic Program Coordinator, McGill University School of Continuing Studies (https://aiwithsach.com/guests/nabil-beitinjaneh/)

## Summary

Nabil Beitinjaneh, who coordinates programs at McGill's School of Continuing Studies, talks with Sach about AI and education. He sees real gains: personalized tutoring, better course preparation, and predictive analytics that flag students at risk. He also names the costs. Students can feel they understand more than they do. Personalized tools can create filter bubbles. Learning, which is often hard and uncomfortable, may get skipped. He says nobody yet has a vision for the future of education, so universities must experiment, as his own courses do by letting students use generative AI to code.

The talk then turns to work. He expects AI to augment individuals but replace some jobs at the macro level, especially intermediate roles. He uses outsourcing, Uber and the Luddites to show how displacement hurts people and how hard it is to fund answers like UBI. He argues that ethics without regulation has no teeth, and that citizens rarely read the bills they are invited to comment on. He lists the risks as privacy, bias, concentration of power, opacity, and loss of human agency. For individuals he recommends AI literacy plus domain expertise. For organizations he recommends governance, clear accountability, and diverse teams.

## Key takeaways

1. AI can give you answers that create an illusion of learning, so educators need new ways to check what students actually know.
2. Learning that is hard and sometimes painful is part of how people learn, and personalized AI tutoring can remove it.
3. Pair AI literacy with domain expertise: without knowing a field, you do not know what to ask the AI.
4. Ethics discussions matter but have little force without regulation, and citizens rarely use the chance to comment on draft laws.
5. Job displacement will not be solved by one actor; it needs government, companies, schools and workers thinking together.
6. Organizations should set clear accountability for AI errors and give people a career path, not just add the technology.

## Verified quotes

Each quote is checked word for word against the recording.

> Well, people get this illusion of learning. You think you know a lot more than you really do. And this is one of the biggest challenges, I would say, that these AIs bring to the table or these systems bring to the table because suddenly you're dealing with text that might be generated by an AI or solutions that are generated or assisted by the AI.
> — Nabil Beitinjaneh, AI with Sach episode 10, at 6:30 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=390s)

> Learning is actually challenging and at times it's even painful, but that's how you learn. You have to solve a problem, you have to fail solving a problem, you have to get a case which you do not know how to solve and then you're going like, why?
> — Nabil Beitinjaneh, AI with Sach episode 10, at 10:38 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=638s)

> somebody might decide to apply to a company uh get interviewed uh go in for tests even physical tests or whatever walk in and out of the company get a job offer and never meet a human in their life it's all done automatically this is not uh the the far future this is today some companies do that
> — Nabil Beitinjaneh, AI with Sach episode 10, at 28:08 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=1688s)

> they would say but if i don't have a path to learn by coding how do i become a team lead and how or an effective team lead and then from there how do i become an architect or you know either path i don't have that path anymore you actually just killed it for me by outsourcing the work
> — Nabil Beitinjaneh, AI with Sach episode 10, at 30:13 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=1813s)

> a lot of discussions around ethics are amazing but ethics on its own does not have teeth the only time you have teeth is through regulation people started worrying about gdpr because there was a financial impact on the companies if they did not follow the gdpr
> — Nabil Beitinjaneh, AI with Sach episode 10, at 45:15 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=2715s)

> nobody has yet a vision of what the future of education would be. Like, there's all these ideas. And the only way that we will know what the right answer is, is through experimentation.
> — Nabil Beitinjaneh, AI with Sach episode 10, at 12:37 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=757s)

> i don't want uh you know an ai correcting an ai doesn't doesn't add any value not for the student and not for for me either like if you want to push to the limit and say you know you know the assessments are done or corrected via the ai well what does the student learn from this
> — Nabil Beitinjaneh, AI with Sach episode 10, at 15:02 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=902s)

> so who's paying the taxes if you're displacing all the jobs let's just take the extreme case where does the government make the money from are we going to tax the robots whatever that means that doesn't mean anything to me and doesn't and it's virtually impossible
> — Nabil Beitinjaneh, AI with Sach episode 10, at 34:10 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=2050s)

> you ended up displacing all these people and it destroyed society too because what ended up happening is industrialization created these factories which was artificial you know people in the older days you know when it came to uh the work in textile they would do it at home
> — Nabil Beitinjaneh, AI with Sach episode 10, at 39:47 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=2387s)

> so in a way we do have also a failure at a society level because everybody keeps talking about ai ai ai and nobody really has time to think about what the impacts are and how they will be impacted by it in the future
> — Nabil Beitinjaneh, AI with Sach episode 10, at 46:47 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=2807s)

> this is something that i always worry about which is what i call autonomy and human agency so who's making a decision and sometimes you don't even know who's making the decision anymore and if you ask an ai to do something for you are you learning or are you understanding
> — Nabil Beitinjaneh, AI with Sach episode 10, at 52:29 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=3149s)

> i was reading something from mckinsey recently they were saying that 50 percent of the um what i call the job descriptions that are out there have skills that people job applicants don't have
> — Nabil Beitinjaneh, AI with Sach episode 10, at 57:24 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=3444s)

> we had an issue recently with air canada where what are the chat bots gave wrong information who do you blame is it the developer of the chat bot is it air canada is it this that is the way it was done
> — Nabil Beitinjaneh, AI with Sach episode 10, at 1:01:04 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=3664s)

> when Bitcoin came out, there was all these companies that are trying to set up, you know, using up all this electricity. Now, with AI, we have the same problem, but we don't talk about it.
> — Nabil Beitinjaneh, AI with Sach episode 10, at 1:04:00 (https://www.youtube.com/watch?v=s-gC9irOk9o&t=3840s)

## Transcript

Automatic transcript; speaker labels are automatic.

[0:00](https://www.youtube.com/watch?v=s-gC9irOk9o&t=0s) **Sacha:** Today, I'm thrilled to be joined by Nabil, an international consultant and educator. Nabil brings a unique blend of expertise in business development, project management and leadership, with a current focus on data analytics, data science and AI. As a faculty lecturer and academic program coordinator at McGill University School of Continuing Studies, Nabil is at the forefront of preparing the next generation of data professionals. He also knows, he's also known for his ability to thrive in ambivalence, and create structure where others see constraints. So, welcome and pleased to have you on today, Nabil. Welcome.

[0:40](https://www.youtube.com/watch?v=s-gC9irOk9o&t=40s) **Guest:** Well, thank you so much, Sacha. It's great to be on your show.

[0:44](https://www.youtube.com/watch?v=s-gC9irOk9o&t=44s) **Sacha:** Okay, so let's jump straight into the first questions that we have today. How do you see AI transforming education in the coming years? And what are the most promising applications for students and teachers?

[1:04](https://www.youtube.com/watch?v=s-gC9irOk9o&t=64s) **Guest:** term, you know, today's term, we really see it transforming education at multiple levels. We first have the students, you know, there's huge changes for the students. Now students have access to a lot more power than they ever had before. And what we're seeing is a move towards what I would use the word personalized tutoring and the ability for students to learn at their rhythm and at their way of operating. At the instructor level, we're also seeing the same thing, but at a different way, where what happens here is instructors can do a lot more or a better preparation of material. And, you know, we have access to the latest information at our fingertips. We always had that through search and through the online books and a lot of things. But now you can even query these books and can query these systems and learn a lot faster to help the students learn even more.

[2:06](https://www.youtube.com/watch?v=s-gC9irOk9o&t=126s) So now you can get more up to date information, more comprehensive material and even, you know, create, you know, tailored courses to the, or tailored material to the students that you have in there. So it's, you know, as an instructor also there is, but this is a, we're still in the early stages, you know, how do we correct? How do we provide feedback? There's a lot of automation that could be done over there. So you can, the instructor can focus more on providing personalized and nuanced, you know, to really help the students move forward. There is a timely element to it. So if you can do these things faster and better, then students get to learn as they're, you know, the learning experience is enhanced because they're not waiting too long to hear back from the, you know, institution or from the instructor to do it. Now, I mentioned the institution. Now at the institution level, and this one here is where, again, all of these, what I'm talking, we're still at the early stages.

[3:10](https://www.youtube.com/watch?v=s-gC9irOk9o&t=190s) But there is this idea that we could get a lot of insight into student performance. We can start looking at, you know, the issues that they have, do predictive analytics to start saying, okay, which students are at risk here? When should, you know, who should you intervene with? How do you work with them, et cetera? And this can work also at the student level, and it can also work at the curriculum level where you can start looking at where the issues are, et cetera. So this is great on one hand, but on the other hand, we have to be very careful about, this is an enhancement to education as opposed to, you know, a focus on technology. Technology on its own does not fix everything, and we need to make sure that we're using AI to augment and help educators and not just replace them.

[4:06](https://www.youtube.com/watch?v=s-gC9irOk9o&t=246s) **Sacha:** That's a great point. And then you mentioned that AI can personalize learning experiences. What does it look like? Can you give us a few examples on how, so you touched upon it, right, a little bit? And also, can you give an example of how educators can leverage AI to really cater to these individual needs and even learning styles, right? I was a student not too long ago, and I've always wondered why education was said that way, where everyone learns in the same way, whereas I'm not the same as my, you know, my colleagues or my classmates. So I'm curious to get your take on that.

[4:50](https://www.youtube.com/watch?v=s-gC9irOk9o&t=290s) **Guest:** One of them is when you start having, when you can learn the patterns of the individual, and this is not something that can be done, you know, quickly, but it's something that could be done over time. Then you can start knowing what kind of a learner are they? Where do they struggle? Are they visual? If you're visual, okay, well then maybe orient them more towards, okay, why don't you look at these videos? Why don't you look at these? Is, you know, let's solve these things graphically versus somebody else who has a different learning style and you would teach them or you would show them the material in different ways. You know, some people prefer text. Some people prefer video. Some people prefer solving problems. Like some, you know, one of the students I worked with recently was someone that, you know, they were working with a course of fundamentals. They had to solve every problem. For them, the only way they can see the patterns is through what I would use kind of more of a brute force approach of solving all the problems to be able to understand all the variance that might happen, et cetera.

[5:56](https://www.youtube.com/watch?v=s-gC9irOk9o&t=356s) So then you adapt to it. Some other people, they get it very quickly, you know, move on quickly, you know, based on how they're doing. But at the same time, you can't move quickly without, you know, making sure that they understand what they're doing. And this is where the systems have to come into play. And it's not always obvious. Now, I flip this around when I talk about now that you have access to all this material, and I'm kind of thinking a little bit further down in my mind. What happens or these tools, if you want to be more precise? Well, people get this illusion of learning. You think you know a lot more than you really do. And this is one of the biggest challenges, I would say, that these AIs bring to the table or these systems bring to the table because suddenly you're dealing with text that might be generated by an AI or solutions that are generated or assisted by the AI. So how do you know what the student knows and what the student doesn't know?

[7:01](https://www.youtube.com/watch?v=s-gC9irOk9o&t=421s) And this is a real serious problem that we have, and I think many people will have because people tend to think that they know a lot more than they really do. Another issue we have to deal with here is when you're dealing with AI and using all these systems that can customize answers and can answer questions in real time. This is another one of the big differentiators, I would say, out of this is you're there. You can, you know, you don't understand something. You can immediately go and start asking questions on that particular area. Something that in the best of cases, you know, if you're in class, you can ask the instructor. That is, if you raise your hand and you feel that, you know, like the, you know, that urge to do it. Many people don't or they're a bit shy about it or they don't like to, they don't want to look, you know, as if they didn't understand it for other people. So here, this is a huge differentiator is that you learn as things happen.

[8:04](https://www.youtube.com/watch?v=s-gC9irOk9o&t=484s) You know, you don't understand something can go immediately and do it. Now, the problem with all of this is you can get into a bubble where you only, you know, especially if you're using systems to help you or assist you, then you might get into what we call a filter, you know, a bubble that where you're isolated. So you only get information that aligns with what you know. And we see that with social media. We have seen this with social media where you get into these, you know, kind of echo chambers where, you know, you only get to hear what you want to hear or the mindsets or the ideas that you were hearing from everybody else. So here we have to be very careful about it because the systems are, you know, as you probably know today are still in their early stages. So we don't know where, you know, how to assess the learning part and how to make sure we don't get into these areas or these, you know, isolation areas.

[9:07](https://www.youtube.com/watch?v=s-gC9irOk9o&t=547s) The other thing which is very important about learning, and we're talking here about, you know, how can we personalize the learning experience is what would happen is sometimes if we over focus on what they want, they will not learn skills which are very important for them. Learning when you go to university, it's not just about learning the material. When you're, it's also about learning how to work with people, how to ask questions, how to, there's all these other dimensions that come into the learning experience. And that part here does not come when you're working with, you know, if you're working one to one with a, let's call it a personalized AI, a tutor, whatever. So here, when we talk about collaboration, debate, those are types of things, learning from peers, learning from people who are different than you, hearing different perspectives.

[10:10](https://www.youtube.com/watch?v=s-gC9irOk9o&t=610s) And so, this you may not hear or you may not get from the systems that are out there today. One, I think, last point I'd like to bring to the table is when we're talking about learning, there is something that even with these systems or, you know, we have all this power at our disposal, there is something that people tend to neglect is that learning is never easy. Learning is actually challenging and at times it's even painful, but that's how you learn. You have to solve a problem, you have to fail solving a problem, you have to get a case which you do not know how to solve and then you're going like, why? Okay, what happened here? And then instantly you get this insight, oh, this is how I should do it, this is what I, and this is something that may, you know, we may lose as we're going forward. Having said that, I just want to, and I know this will come in a later, you know, question, I'm quite sure you're going to ask me about ethics, etc.

[11:11](https://www.youtube.com/watch?v=s-gC9irOk9o&t=671s) If you're customizing a lot, you're learning a lot about the individual, well, who owns that information? Is it the university? Is it the education provider? So even you forget about the word, educate, is it the government? You know, because you're part of the education system? Or is it the individual? And if it's the individual, how do you generalize? Or how do you, you know, like, there's a give and take because to have better customization, you need to have many different examples. How do you do it? There's all these challenges that are there, and these will impact how these systems are designed.

[11:54](https://www.youtube.com/watch?v=s-gC9irOk9o&t=714s) **Sacha:** All of these points were incredible. And I'm curious, it's a bit off of the typical question that I would ask, but are we going back to studying and, you know, doing everything on paper, again, in like, you know, a classical institution way? Or are we going more towards having, as you mentioned, kind of like a personalized AI for every student? Or is it something in between? How do you see the future of universities as we know them today? How, how, how are we moving on and adapting to, to these technologies?

[12:35](https://www.youtube.com/watch?v=s-gC9irOk9o&t=755s) **Guest:** Well, I don't think we have a, nobody has yet a vision of what the future of education would be. Like, there's all these ideas. And the only way that we will know what the right answer is, is through experimentation. The only way you learn is you try different things, and you see what works and what doesn't work, how far can you go, etc. So we've been experimenting in our courses with generative AI, you know, teaching students how to use generative AIs to code. Because a lot of what we're doing is at the basic level, you know, when some of the introductory courses, is it sufficient? Is it good enough? What can we do? How can we do it? You know, have the students go out and do it. And so there's a lot of learnings that happen, that's happening over there. And now we need to sit down and assess. And this is, we're still, like I said, in the early stages. Now, going back to paper, in some cases makes sense. Because in some cases, you may need to demonstrate your own knowledge.

[13:37](https://www.youtube.com/watch?v=s-gC9irOk9o&t=817s) You know, you cannot walk down the street, you know, if you have to communicate with people in Spanish, or let's just take a language thing. Eventually, you have to interact with people yourself, you cannot have the machine interact completely. Maybe you can, maybe, maybe that's the future where each one talks in their own language. And you have these agents or these, you know, UIs or phones, you know, you're talking to your phone. And then on the other side, the person hears what they, you know, the translated version, maybe that is the future. But we don't know how it works. And it's like, let's use the word chat GPT to answer an assignment. How much do they really learn? How much do they benefit from it? Now, I'm all for it. And in some of the courses I have, I'm using it completely where hey use it we only ask students to tell us that they've used it or not and we've added extra questions around that or we're adding extra questions asking them where did they run into problems because you know i don't want to spend my time correcting uh you know chat gpt or clode or whatever and you know and let's just push this a little bit further i don't want uh you know an ai correcting an ai doesn't doesn't add any value not for the student and not for for me either like if you want to push to the limit and say you know you know the assessments are done or corrected via the ai well what does the student learn from this you know and how do they learn from it so especially if they didn't use an ai to generate it so this is where it's it's very important from an education perspective to ask these questions and work with the students and the top of experiments to see what works and what doesn't work how you demonstrate knowledge a simple example is you know what do we need to know and what don't we need to know does anybody need to know how to use you know directions anymore you know everybody uses google maps or the equivalent apple maps or whatever do we need to learn how to create a map the answer is it depends on what your job is depends on what you're going to be doing and this is this is a challenge for everyone i would say you know because you need to define what is important and what isn't important this comes into coding you know now the big debate is are we going to have coders in the future or is it going to be everything going to be generated and up to now up to until today it's an enhancer it's not a remplacer that's

[16:31](https://www.youtube.com/watch?v=s-gC9irOk9o&t=991s) **Sacha:** a great point and it seems from from what you're saying is maybe that hard skills are becoming less and less important since you know coding for example uh we're going to have tgpt cloud whatever you may need to to help you with that so what kind of soft skills do you see becoming more and more important um to teach right students becoming critical of the information that you read and that these systems produce for example um what do you think so

[17:10](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1030s) **Guest:** we start with the first one we keep hearing time after time different educators different companies anytime you ask somebody what should people learn oh they should let be critical thinkers well and this has been the case since we've been going online and since before but especially when we started going online we should be able to identify what's real and what isn't real you know what is important what isn't important how are things done you know when you look at a visualization at the end of the line which is built on data you have to be able to understand or we should be able to explain how did you you know what are all the steps you went through to get to this visualization and understand what are the issues that have that might be in the data itself you know like because you have issues that could be all the way from the start all the way till the end of the the process and there's all these things that can go right or wrong along the way so critical thinking becomes for me something which is very very very important because you have to be able to challenge the understanding or how things are put together now aren't they same you know if we continue with thinking adjacent to it you have communication skills you know if you're going to be working with these you know you're going to be able to communicate very well you know the tool will not figure out what you want you have to give it very precise questions or you have to walk through as you know the prompt if you can use that term you have to guide it and you have to be able to communicate very clearly what you want so this is where it's incredibly important to be fluent in languages in you know especially the language that you're communicating at work or in your organization and you add to it the idea that now we have to be able to work together even more so emotional intelligence becomes i would say something that's i wouldn't use it was never under underrated but now it becomes even more important because now you have to work with people locally remotely now you're working with machines so you have to be able to kind of navigate all of those i add to that things like the ability to learn now learning is incredibly important now the problem is there's been an acceleration in the amount of at least in the area i'm in i'm i cannot talk about all areas but we you know by time you learn something there's two new things to learn by the time you figured out those two new things that you're learning some of the other some of the some stuff is absolute some new stuff is is there and you're constantly relearning and you know unlearning relearning and learning new topics so it's it's a challenge so to learn properly is something that people have to spend time thinking about and that for me is something that honestly even when i was back in university it was not something i i i hear that now they teach people how to learn how to research how to do these things but in my days it wasn't there it wasn't at least not a priority maybe i don't remember it but it's incredibly important important to learn and how to learn and how to you know retain what should you retain because you want to ask questions if you do not have expertise or knowledge you don't know what questions to ask you know the the the ai can you know if you want to use a search engine or the chat gpt or whatever the llm can answer so many different things for you but if you don't know what to ask you're not you know there's domain expertise which is important and this is something where you have to marry ai literacy with domain expertise you know you have to still decide this is an area i want to specialize in and i i have to know everything about it so i can go back and understand what's going on i can understand how things fit somebody explain something to me i can see it and

[21:38](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1298s) **Sacha:** you know what this last point of learning how to learn is is a great one and i fully agree with everything that you said but i want to dig deeper in this in this uh last point that you've mentioned and i'm actually with the right person to ask this because you're heading the continuing education program at mcgill university so what role does continuous education play in this you know age of ai so

[22:08](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1328s) **Guest:** this is at least the way i see things remember we we all have our own biases and i have something to sell so i'll you know take everything whatever people tell you always listen to them but you have to make your own decisions so we you know i am focused on or we at mcgill are looking at the shift that has happened education is you know how the shift is happening what is coming on etc so we are trying to adapt to these and save the students whenever possible or our potential students the time to come up to speed so there's a real we're looking at how things are done why they're done in a certain way and then trying to work to create new programs there's approaches to older programs to get things going so as i told you uh in and i'm talking about the programs i'm involved in we're involving you know like last time we talked in one of the first times that we start bringing in chat gpt into the classroom let's use code let's use chat gpt or whatever you want to use from a llm perspective to do your coding and so develop code troubleshoot code understand the code that's given to you by the instructor document it etc so there's all this our role is how do we make it easier and then now we're thinking in terms of one of the big challenges we have is that people don't have the time anymore or they don't have the breath you know they can't think about doing programs that take years and years and years to do people want to learn something now fast etc so now we have to think in terms of how can we offer that how can we make things more adapted to them so we're talking about things called micro credentials you know one or two courses or three courses put together that you learn the skill you know there's all these little changes that we're doing to make it make education more accessible to people so that those are the types of things that we're trying to do and it's it's not a simple one but again i'm giving you my perspective and it's i'm in a very specific area like we're in the technology area so it's uh we have you know our own set of constraints and our own set of challenges and but i would say huge opportunities in that area because you know we have access or we you know the technology is helping move these things forward and

[24:37](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1477s) **Sacha:** um i want to i want to go back to one topic that you said a little bit earlier there is a lot of talk going on around augmentation versus automation and i want to get your opinion on that so what what do you think is is going to like how do you think it is going to play out do you think we're more likely to actually replace fully some jobs replace some tasks or do we automate um i mean augment sorry uh human capabilities um what's what's your take on this debate

[25:13](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1513s) **Guest:** um my view is really at the individual level you're going to have ai augmenting human capabilities now at a macro level whether we're talking organization or society level we'll start we'll see both we'll see augmentation and replacement fortunately or unfortunately that's the way it works you have new technologies you have new ways of doing things so there will be a replacement that happens so people will use this technology organizations will use this technology or technologies together to build new ways of doing things that will remove or make jobs disappear so at an individual level we cannot focus too much at the well we cannot think about the macro we have to be aware of it but we have to think in terms of positioning ourselves where we can use the ai to augment our capabilities so it is you know and there's multiple studies that have been happened that i read uh in the past couple of years about what you know the use i think the current terminology is collaborative intelligence where you have humans and the eyes working together to do better so in the past if you want to say many six or seven years ago there was an article about human plus process is better than human alone so now when you have the ai you're you're moving that one even further so if you're able to work and augment your work with ai you should be able to do more on a day-to-day basis is it going to replace people this is in some areas yes will replace programmers completely i don't think so at least not yet um it's unclear to me if that's going to happen or not we have to wait we're still in the early stages some things will only we'll only see with time and we'll see how things play out however definitely with newer technologies it doesn't really matter what we're doing there's going to be a whole reimagination of processes and ways of doing things at the corporate level that will eliminate a lot of what i would use intermediate jobs in there

[27:37](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1657s) **Sacha:** and how do you see how do you see the workplace being reshaped potentially completely disrupt um and and like the new types of jobs that will become available do you do you see new career paths new jobs emerging or um some something completely different happening

[27:59](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1679s) **Guest:** there's there's a lot of different things that could happen now you can take a dystopian view and you can start seeing and it has happened in some places um where somebody might decide to apply to a company uh get interviewed uh go in for tests even physical tests or whatever walk in and out of the company get a job offer and never meet a human in their life it's all done automatically this is not uh the the far future this is today some companies do that and is this good or bad it depends on how you see things for me there is this there's an isolation factor here uh which is not good if you don't meet humans along the way then you're on your own you have your own uh narratives that go through your mind and there's other things that you know there's a lot of different things that come with it now the real challenge is how do you do this and keep the human like we saw a big shift and you know both of us are part of that in a sense you're working from home uh i work from home a lot a lot this is a huge change from the past we're still debating right now is that good or bad you know and some people will say it's amazing some people say no it's not and you know some people say let's go back to the office and we're talking mostly managers that say that some other people say no no we really like this this is amazing actually we should do more of it and we're gonna we're gonna see the same with ai you know where some areas people say it makes sense or not like we had a question about 20 years ago about outsourcing and coding like the simple debate in one of the companies i worked for was well if you outsource what i would use the word basic coding to india which many companies were doing how does a young developer develop their expertise you know you you know because the people that were inside the company many of them became architects they became team leads so they were guiding the people in india to do their work and but the the argument that some people had some are more junior people and i and i kind of understood it perfectly they would say but if i don't have a path to learn by coding how do i become a team lead and how or an effective team lead and then from there how do i become an architect or you know either path i don't have that path anymore you actually just killed it for me by outsourcing the work so we have a lot of challenges when it comes to working in the future now this even goes much further when it comes to areas such as you know when we talk about um you know automation you know what's human and what's not you know how far do we go with automation in some cases we are we you know about you know people being measured so much let's say the over example in some countries where they have to you know relieve themselves in bottles because they don't have enough time to stop and go to the bathroom because they're being measured in inhumane ways and this applies also to people you know where in many uh you know uh if you want to use the word work many not factories the expectation has moved up from an automation perspective that you know the the human is close to being a machine or being told what to do as if it was a machine as opposed to it being a human so we're there's there's a lot going on here and we as a society have a lot of thinking to do too

[31:42](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1902s) **Sacha:** wow um i want to i want to dig deeper into this this topic and and like the broader social impacts that the broader social impacts that we're going to see uh from uh this ai driven job displacement what measures can we can we take to ensure that you know it's actually equitable that we have an equitable transition for for everyone as you mentioned this this example of the uber drivers that you know need to uh do what shouldn't be done really and then is um universal basic income maybe a viable solution or what do you think okay

[32:37](https://www.youtube.com/watch?v=s-gC9irOk9o&t=1957s) **Guest:** so for the i'll start with the last one first because that's an area that i would say is beyond my area of expertise so it's easy for me to say i'll you know punt it out so ubi or the universal basic income has been an idea that has been floated many times in the past many years with ai we need to have ubi etc now it's not a simple thing so this is where i'm saying i'm not the best person to ask there's a lot of people that probably would be more qualified to ask this one because this is something that's done at a macro level at an economy at a society level and it's not it's the same thing as saying should we have um you know what we call uh we have right now ei you know unemployment insurance we have social assistance and even those two are very um how they call it problematic for many people should we have more uh ei should we have less ei it used to be unemployment insurance now it has become employment insurance and there's some a lot of nuances in between there you know should we have social assistance yes no this is very similar in the sense where if you're not working should you get a basic salary or something that that uh comes from the real question is who funds this this is not a simple uh question because the money has to come from somewhere so somebody has to be paying taxes if you want to use that term to fund this so who's paying the taxes if you're displacing all the jobs let's just take the extreme case where does the government make the money from are we going to tax the robots whatever that means that doesn't mean anything to me and doesn't and it's virtually impossible it's always been included inside companies it's like this is not something you know i can go back to the 90s and you had robotics inside companies um were they taxed they were taxed on their income you know as a company so what you have to do is you have to encourage companies to use ai and have them create economic value for the country now the other this is another challenge which is okay are you doing it locally are you doing it for somebody else that's a huge challenge because where do the taxes go if all the taxes are going to another country if any or we're giving tax breaks to all these companies which is what something that everybody asked for oh we need more people that know ai we need to fund people to learn which is i i'm not disputing that actually i agree with it fully but okay if you're funding something where does the money come from it's always that question that we have to deal with and this is done at a higher level and it's unclear how it would be done um and how it will look like in reality and how do you encourage people to get off the ubi etc which is something we saw with uh covid when people were getting money at home you know it was it actually stopped a whole you know people could not find people because it was easier to stay at home and not do work than go out and do work and so there's this thing there's a lot of thinking that has to happen there which goes beyond what i'm capable of abstract and i don't have the experience there now how do you make this equitable or how do you you know this is where it's important to understand the impacts and see how to do it and this is where government intervention is required let's say when uber came to town do we encourage uber which is a new technology you know whatever whatever this is amazing we can do all these things with uber or you know and you know the taxi drivers we don't like them too much because actually many of them were not very nice so actually some good competition would not hurt that part i don't disagree with but what happened was we destroyed the livelihood of many taxi people so what happened is you had this thing that existed in at the time which was you had to to become a taxi person you you know to run a taxi you had to have a vignette that you can take and use and this had value you had to pay for it and then suddenly you have this you have this company that comes out of nowhere doesn't have to have the vignettes anybody can be a driver and they're competing and taking away the customers of the taxi drivers so suddenly you there's a destruction of value that exists and how do you deal with that now there's no easy answer to that one because what happened was the government decided okay let it be you know let's uh remove the uh but the taxi drivers have and they actually now have a lawsuit against the government which we'll see how that one works out at one time are they going to make money they're not going to make money it's unclear where we're going with this one but this is just an example it's i don't really care well i do care which way it goes one way or another but what really matters here is how do you deal with the displacement because what are we talking about the most when we're talking about ai these days self-driving cars a lot self-driving trucks well guess how many hundreds of thousands of people are going to lose their jobs because of that if we end up with self-driving cars i'm not you know i haven't seen it yet you know even the best self-driving cars are not self-driving however if we you know assume that they're going to be there what are we going to do about it and that's the real challenge how do you when you have mass displacements of jobs how do you deal with this and this is something that has happened to us in the past you know the industrial revolution came and if you go back to the history there's actually a great book called blood in the machine which was written recently by a technology uh person who talks about the impact of what when they did the big changes in um you know in textile so you had all these crafts people that would spend they spent a lifetime or lifetime they spent seven years or so many years of their lives as an apprentice to be able to do a job so they had to learn the job they them for so many years or close to it to learn a profession and after so many years they would kind of you know same as a plumber would or somebody would get a you know you can do this and suddenly the machines come in they do similar jobs much cheaper and let's not argue about the quality the quality was less but people didn't see the difference anyhow most people didn't didn't care cheaper is better so you ended up displacing all these people and it destroyed society too because what ended up happening is industrialization created these factories which was artificial you know people in the older days you know when it came to uh the work in textile they would do it at home so they get everything they do the work they would they would give it to somebody who kind of an aggregator who put it all together and sell it fantastic so now you have to work in a factory you know you're getting paid minimal wage if not below minimum wage there was no minimum wage at the time you're paid very little and your social structure got impacted so there was actually even a pushback at the time by people called the luddites and this is why the book is so interesting because he in the book they they he covers what the luddites did and how they fought against the system and actually how they lost it was even more interesting to see that at end society cared more about making money and they outlawed everything you know from unionization to being part of the luddite groups or to uh whatever it reached a point that you know people got lost their life because of that you know you had people that were fighting against the system and despite the fact that they were trying to fight it on the legal side they lost their and they lost the the right to protest and fight against the system that the other way so we don't want to be there but we saw that also happen with the ubers you know if you remember well you had people you know knocking on the cars or you know uh you know stopping the ubers and yelling at them and stuff like that and you know we we felt that that was very unacceptable because that's not a way to operate but but you know like uh people lost a lot of money because of this

[41:49](https://www.youtube.com/watch?v=s-gC9irOk9o&t=2509s) **Sacha:** so that was that was a beautiful answer first of all what role should governments play then in in regulating you know this new technology do you think it should be as strict as what's happening in europe or should we should we be more proactive reactive how do you see how do you see the government playing a role here especially on on employments as well

[42:20](https://www.youtube.com/watch?v=s-gC9irOk9o&t=2540s) **Guest:** the government has to have an active role now how active how far this is something that only can be determined um through action so what happens there's a certain truth about like in one of the courses we give we actually bring in we have an ethicist come in and we have a lawyer come in to talk to us about how law works and law takes a long time to catch up with technology it's it's normal because you first have to you you cannot foresee everything that technology is going to be able to do if you go back four or five years and you talk you start thinking even if you go back 10 years try to think in terms of where we are today you cannot you know like all of what we're doing today or much of what we're doing today was totally not visible to us many years ago like what we can do today with chat gpt clode whatever perplexity let's let's keep on going down the list um we couldn't even see it three years ago like it it became like magic when when the first when open ai did it it was magic so or magic implemented even though we had all the elements and we saw how things could be done but we didn't know what the impact would be so law cannot preemptively think in terms of ahead of time this is you know we want to work on these areas so a lot of time they have to wait until something is out you'd see the impact you have to see in terms of and now i'm quoting my uh you know probably name ampaki who comes and teaches us is you have to see is there a for this area or not law itself like when horses first came out they displaced you know uh a lot of the way the market worked at the time you know well do you need to have the law of the horse or do you use existing laws and frameworks to answer problems that happen and in general you use the existing laws until you find that you cannot they do not work in certain cases and so what happens here is we have an issue in my books with fluency so the governments are trying their best to put rules and regulations so when it comes to privacy actually canada is pretty advanced quebec is incredibly advanced when it comes to privacy rules so we are doing good there now the real question is do we need more regulation now a business will always tell you you need less regulation because obviously you need less it's it impacts them and this brings up an issue when it comes to ethics etc a lot of discussions around ethics are amazing but ethics on its own does not have teeth the only time you have teeth is through regulation people started worrying about gdpr because there was a financial impact on the companies if they did not follow the gdpr you're talking to the european rules uh well so that scares companies scares them or at least gets them to think you know if there's a rule if there's a law they will follow the law so this is where the government has to make a decision how far do we want to go what's the society and what is the impact on society and what are we trying to encourage you know do we want to encourage companies or technology or are we trying to do the you know society or is a question of balancing between both it's not clear what way to do and this is what i'm saying this is this has to be done at a higher level and even at the societal level because it's not something like i have ideas or thoughts on the topic but i am not someone who's involved in the laws etc but both me and you are asked every now and then you know the government every time before they come up with a new bill or something they'll actually publish it and put it on the internet for both of us to input on it but i'm quite sure you never did it and i'm quite sure i never did it and most of the people i know never spent any time at all looking at the regulations and the rules that are out there so in a way we do have also a failure at a society level because everybody keeps talking about ai ai ai and nobody really has time to think about what the impacts are and how they will be impacted by it in the future

[47:03](https://www.youtube.com/watch?v=s-gC9irOk9o&t=2823s) **Sacha:** and and you know this is also like kind of the the objective of this this uh podcast is to drive discussion and i think if we have discussion with people like you and i and and you know different leaders then then maybe it has a snowball effect and results into something and and we get ideas going as well people you know listen and they get inspired maybe hopefully and then they they want to be proactive and i want to get into another question you mentioned a little bit about ethics and i want to ask you what ethical challenges do you see with with the ai with ai that um may pose like a risk to society

[47:45](https://www.youtube.com/watch?v=s-gC9irOk9o&t=2865s) **Guest:** so the issues are really they they haven't changed over time just that the the way we see them changes so we have privacy concern we talked about that um especially now that you're able to nearly break any system so the uh hackers are able to hack anything so on one hand you have everybody collecting data on the other hand you have the hackers how do you balance between both and how do you even balance what a company has you know there's all these discussions at one time another about you know what does meta know about a person what does google know about a person and some people say well they know so much right now we can't do anything about it well yes and no so that those are issues that are very important privacy then you have algorithmic bias side which is you have to be very careful about the algorithms and the data that's been used to start off with to build up all these systems and this is a whole area on its own um however if not if you don't look at it carefully we will tend to amplify the biases that existed from day one whether it's color whether it's sex whether it's uh age all these things get amplified whether it's education whatever everything you know you just get more of if you don't if you're not careful about it and we have seen the concentration of power you know if you look at how many communications systems you have these days you know you can talk about what's up you can talk about instagram you can talk about this you can talk about all the ones that are out there you don't have that many really and you heard recently what happened with telegram you know so it's a question of you know what's out there and what's you know obviously every system you have the good and the bad you know telegram was used for good and i'm quite sure has been used for terrible things too you know same thing is it what is what's up used the same way we don't know but it seems that what's up has a you know people do not look at it the same way for some reason we have tick tock the chinese you know this that there's all these battles going on um at a different level and it's unclear to me um whether they're these are technology battles or societal battles you know are we trying to have one country versus another are we having a big blocks fight china versus the us china versus russia i'm sorry us versus russia what what are the fights europe versus the world you know and this is you know if you take the the view of europe as being very you know they are very protective and the argument you can have from an american is well they're so big companies to to worry about so it's it's it's give and take at each level so we have a concentration of power which is scary and this was brought up in different ways you can have few companies that are doing certain activities and the other thing is and this was brought up in a book called weapons of math destruction you can have a few algorithms that impact a lot of people in the old days you can only do it at a school level or at a district level now you can do it on a country level and uh you know you have these rules or metrics that are put in place that look at people and are not good so we have a lot of issues there so we have transparencies ai systems are still considered to be black boxes so how do we work with them how do we understand what the output is where are the problem areas and this is something where when we talk about domain knowledge it becomes very important to understand not just use you know some people will use and i'm going to you know give you a little segue which is you can drive a car you don't need to know how the car works however you're going to have people like mechanics that need to know how all the pieces fit together and you need engineers who know not just how the pieces fit together but they need to design that these things all these elements to work together and come up with new ones so you need different types of people that are there um the one of the last issues or what you know right now that i can think about and i talked about data issues earlier on which is data quality how we gather data how we you know data is used uh representative or not but we do and this is something that i always worry about which is what i call autonomy and human agency so who's making a decision and sometimes you don't even know who's making the decision anymore and if you ask an ai to do something for you are you learning or are you understanding this brings us back to the start which is where we started if you ask you know an ai to do coding does it really are you learning how to code you need to know how it codes etc when you ask it to take you different places should you know how to navigate is navigating an important thing well what happens if we have a solar flare and the systems go down suddenly we don't know what to do anymore you know we can't even go somewhere because we don't know how to drive anymore you know let's assuming that we have cars that can still drive after solar flare and the electronics are not for fright but you know what i mean which is what do we have to know what is important to know and how much should people depend on the machines to do things and where do they where do they have to say no no no we have to do it now the perfect example comes in military applications should you allow a machine to kill another human or should you have someone that presses the button saying okay we identified this person this person is x we have a kill on him kill you know not to say don't get me wrong i'm a pacifist i don't even believe in killing i don't even believe in war uh that's not the way to solve problems however it is a huge issue which is where do we start and where do we end yeah

[54:30](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3270s) **Sacha:** and it seems to me that you know also with the examples that you've given a lot of everything we do as humans is based on trust so we trust the engineers to design you know functioning cars that won't crash um we trust we trust each other right like i trust you that you're going to give proper answers to my podcast and you're going to be as genuine and honest as possible everything is based on trust but it seems like with these ai systems that as you mentioned because they're in the hands of so many people that the trust becomes a lot more important right because now everyone has access to it and we need to trust that people will actually use them for good rather than for um for bad unfortunately and some people will use it for bad but i want to ask you and i want to transition to how how can you know individuals like what kind of strategies can individuals and even organization employee to navigate um this this transition that we're seeing right now which is really an economy powered by ai so

[55:45](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3345s) **Guest:** i think the basic points for individuals is we need to focus on developing ai literacy whoever you are you cannot escape ai so you need to understand what ai is all about capabilities and most importantly the limitations the fact that you we talked about you have to learn about learning and you have to also decide or you know you can no longer be of the school that i studied i went to university and that's it you know i don't need to go back to school for the rest of my life you're going to be constantly learning you know fortunately or unfortunately i look at it as fortunately i like learning but some people do not like to go back to school or think about it as being a chore etc now the we need individuals need to focus on what's human which is the critical thinking the working with people how to solve complex problems these are things that machines have more difficulty with because you need to break down the problem down into something more easier to work with so you can solve separate problems so this is really very important uh very important to do and i always say if you want to learn about ai go find a group that's working on different problems and join them it's good to work with other people don't i think ai is something that a lot of people talk about very few you know fewer people use go use it try it out try how things work etc now organizations is more they they have the same problem which is a lot of organizations i was reading something from mckinsey recently they were saying that 50 percent of the um what i call the job descriptions that are out there have skills that people job applicants don't have so what this is saying is there's so much that people have to learn so you can't just go out and get people like everybody thinks in terms of well this you can fire people and get new people well even people from the outside won't have the skills that you want so there's going to be a lot of work within the organizations to get people up to speed which is what i call ai literacy and there has to be the valuing the okay let's work with ais or you know like this if you use it show me how you use it show your friends how you use it use show your colleagues how to use it let's talk about it is there a better way to use ais etc so it's there is value there so people have to understand that this is coming so how to work it now within a company you have to have these discussions on ethics so now we can argue how valuable they are or not but the discussion on its own is worth a lot and that's something that people are not doing enough um so those are i think the big things that are need to be done at a company level now we talk also about within a company to focus on diversity and inclusion because when you have different people from different backgrounds different uh skill sets etc they will kind of show you or tell you what works or what doesn't work and they will give you different perspectives on um the development that you're doing you know if you're developing a product focused on women you can't have just men do it for example and it's better to have people from different backgrounds because it's you know different women from different backgrounds will use it in different ways you know i'm just giving that as an example companies have to be clear there has to be a lot of communications in there um you know now that i'm an education institution i'll say partner you know we shouldn't partner with education institutions because there's a lot that could happen or good that can happen between an organization and um a university or a university um um and a company there's a lot of back and forward that can happen a lot of good discussions that need to happen within the company we need to start talking about okay governance data governance ai governance how are we doing this um we have no choice part of it is going to come whether we want it or not through regulation but the other one is we need to understand within a company what we're doing and how we're doing it is and there has to be a way for people you know customers or users or potential users to give feedback on what's going on or on the products that are there this is something that's not always there because you have a problem who do you complain to or if you think you see something that's not working who do you go with and this is an issue also within within a company or outside of a company where does the responsibility lie is it the company is it the developer is it where where does the buck stop if you can use that term and this is a huge issue you always had that issue with multiple vendors inside a company now it becomes even more of a problem because you can have more areas that could run into problems if you can use that term and where where do you lay the blame or it's not even laying the blame how do you fix things we had an issue recently with air canada where what are the chat bots gave wrong information who do you blame is it the developer of the chat bot is it air canada is it this that is the way it was done there's all these things that could could happen so we need to have these discussions and there's legal implications in there but there's also ownership of what's going on i think that should give you an an overview of some of the areas obviously there may be more and i think through discussions i can probably if i go back and think i'll give you some more but if you start looking at the basic areas it help move move things forward you know you need need to think in terms of roles and responsibilities in a company who does what um are the roles well defined you know ai you can't just dump ai inside a company people that own it own the thing and even if you have people you bring in a data scientist you bring in an ai engineer or whatever you want to call them to work they have to have a path within the company a way to grow this is not always you know in the older days this was not a not always understood by companies they think oh we'll just bring in the technology no it doesn't work that way it's more than technology it's there's people there's processes there's a way of doing things there's all these things that come to play together and it's not just it's not technology on its own

[1:02:37](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3757s) **Sacha:** Nabil, I'd like to give you the floor to say any last remarks on this beautiful long conversation we just had.

[1:02:46](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3766s) **Guest:** Well, there is a lot of implications of AI-driven job displacement. And we do not know the answers to all of what these implications are. We don't fully understand who will be impacted and how fast because there's an impact on how fast. This situation that involves society, economics, government, and the companies that are doing this work here. So we need to look at different solutions. We need to start thinking about these things as a group. It's not a one person that can do it. And we need to start thinking in terms of, you know, do we want to look at UBI? Do we want to look at different solutions? How do we help entrepreneurs to implement AI in their work? How do we help companies put AI or get knowledgeable AI? And we need to step back sometimes, even one step further. You know, somebody you've probably talked to in the past, Sasha Lusiani.

[1:03:47](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3827s) She talks about the impact of energy. You know, all these AIs have energy implications. Are we thinking through all of those? You know, when Bitcoin came out, there was all these companies that are trying to set up, you know, using up all this electricity. Now, with AI, we have the same problem, but we don't talk about it. You know, training a model is very expensive from an electricity perspective and from an energy perspective. What are we doing about that? So we have to be very careful about all these things. So we have to think in terms of how do we, you know, it's not, these are complex problems. This is a system problem that has many different elements that interact. We cannot just focus on one little area and say, okay, this is what we're going to look at. We have to step back, and this is where people, you know, where you come in, where you bring in different people from different backgrounds to talk about the same problem or different, you know, facets of the same problem.

[1:04:51](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3891s) And then you have to get them to talk together, not individually. So thank you for bringing me to this, Sasha. I really appreciate it. And if you have any more questions, please, you know, you can do it online or offline.

[1:05:06](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3906s) **Sacha:** Thank you so much for coming in. I really appreciate it.

[1:05:10](https://www.youtube.com/watch?v=s-gC9irOk9o&t=3910s) **Guest:** Thank you.