Transcript

Global Lecture Series: Empowering Innovation

From ETH Zürich · Episode page · Watch on YouTube

· 1 h 2 min · 9,928 words · 94 timestamped sections

This is the full, timestamped transcript of "Global Lecture Series: Empowering Innovation", an episode of ETH Zürich published December 18, 2025 (1 h 2 min, about 9,928 words). It mentions 10 people, works and ideas, including Artificial Intelligence, Petra Airman, Chris Lupkamen, Stefano Brizoni, ETH Zurich.

Who speaks in this episode?

What is discussed, and when

  1. – 10:00: Petra Airman, Chris Lupkamen, Stefano Brizoni, Google LLC
  2. – 20:00: Ringy, Ping An
  3. – 40:00: Artificial Intelligence
  4. – 50:00: Google LLC
  5. – 1:00:00: Google LLC

Notable quotes from "Global Lecture Series: Empowering Innovation"

  • My name is Chris Lupkamen and I have the pleasure and the privilege of leading the Strategic Foresight Hub within the Office of the President here at ETH Zurich.
    about Chris Lupkamen · at
  • It is an incredibly important and vital topics today, the age of AI.
    about Artificial Intelligence · at
  • And until a few weeks ago, I was Chief Innovation and AI officer at Ringy.
    about Petra Airman · at
  • And our a Tejal colleague who is going to be joining us today is Professor Stefano Brizoni.
    about Stefano Brizoni · at
  • There's a company for instance, in China, Ping An, and the founder told the new incoming Co CEO, you're going to fail.
    about Ping An · at
  • So the question then is how do you actually bring in that experimentation And the way how I did that at Ringy was so when I started, I asked everybody saying 5 questions and I…
    about Ringy · at

Who and what is mentioned in this episode?

People

  • Petra Airman — Person · Main guest · 2 mentions · first at
    Featured speaker discussing how to build innovation systems in large firms and how AI changes innovation velocity and personalization.
  • Chris Lupkamen — Person · Host / Moderator · 2 mentions · first at
    Introduces the ETH Global Lecture, frames the discussion on innovation and AI, moderates Q&A, and manages time and audience interaction.
  • Stefano Brizoni — Person · Co-speaker / Referenced expert (Professor of innovation) · 2 mentions · first at
    Provides an academic lens on innovation as adoption and learning processes, obstacles to innovation, and how AI changes problem representation and upskilling needs.

Companies & orgs

  • ETH Zurich — Company · Setting · 1 mention · first at
    Host institution for the lecture series, providing the academic context for the discussion.
    Good afternoon, everyone, and welcome to the Audi Max and welcome to this edition of the ETH Global Lecture series entitled Empowering Innovation.
  • Ringy — Company · Business focus · 1 mention · first at
    Former employer of Petra Airman, where she served as Chief Innovation and AI officer, discussed in the context of corporate innovation.
  • Google LLC — Company · Referenced company · 4 mentions · first at
    Mentioned as a former employer of Petra Airman, highlighting her background in product innovation.
    And you were also at Google for a while.
  • Ping An — Company · Case study · 1 mention · first at
    Mentioned as an example of a company with a strong culture of experimentation and failure acceptance in innovation.
  • DKV Mobility — Company · Referenced company · 2 mentions · first at
    Mentioned as another company where Petra Airman serves on the board, showing her diverse board memberships.
    And so I'm in the board of directors at Bossard, which is a publicly listed company here in Switzerland.
  • Bossard — Company · Referenced company · 1 mention · first at
    Mentioned as a company where Petra Airman serves on the board, illustrating her involvement in corporate governance.
    And so I'm in the board of directors at Bossard, which is a publicly listed company here in Switzerland.

Concepts & topics

  • Artificial Intelligence — Concept · Core topic · 2 mentions · first at
    Central theme of the episode, discussing the impact and future of AI in innovation and business.

Full transcript with timestamps

Very nice. Each time it's a little bit different, which is sort of fun for me at least, and hopefully for you too. Good evening. Good afternoon, everyone, and welcome to the Audi Max and welcome to this edition of the ETH Global Lecture series entitled Empowering Innovation. Thriving, Not surviving, but thriving in the age of AI. My name is Chris Lupkamen and I have the pleasure and the privilege of leading the Strategic Foresight Hub within the Office of the President here at ETH Zurich. And most importantly for you, I'm the host to make sure we start on time and end on time.

So we will end at 6:30. The Global Lecture Series brings the world's leading minds to ETH Zurich to explore ideas that are shaping and will shape our future. We aim to foster broader understanding, enhance innovation and deepen impact through thought provoking dialogue and diverse perspectives. I think all of us in this room have chosen to be here, are very interested obviously in this topic today, both innovation and the AI part. It is an incredibly important and vital topics today, the age of AI. For some, this brings fear reactions, for some brings excitement and thrills, uncertainty, opportunity.

Where does that stand for all of us in the room tonight as we navigate this transformation, we need you and we'd like for you to consider where you stand and shaping and creating what it will take to thrive in this new era. Now, I'm going to invite our guests up for just a few moments and you will also have a chance to give us questions to ask. So we have a couple of helpers who also checked you in if you have a question. We have little cards, the cards which you can write the questions on and then they'll bring them up to me and we can integrate those into our conversation.

And I do that because sometimes 3 or 4 questions will come up at once and say, OK, we got to ask that one. And someone might ask a really random question about some algorithm that I've never heard of and then or we've might get to that one as well. So just saying who knows where we're going to stand with those now that's enough for me. Our special guest today is an ATI Lumna and also a digital shaper, Petra Airman. So Petra, come on up, join me up here. Let's give her a round of applause as he joins you on the stage.

It's great to have you. It's great to have you back again at ETH. Did you ever have a class in here? I didn't have a class in here, but I remember we had two events here after I graduated, which I moderated. Oh, very cool. Yeah. So what does everybody in the room need to know about you that's relevant for tonight's conversation? I've seen already some familiar faces here. So. But nonetheless, I think what's good to know about me, I'm an executive and board member. I'm somebody who loves to turn ideas into impact, who empowers teams to turn those ideas then into products and scale them.

And I'm a woman who loves to shape the future. I love to create what will actually lead to the revenue in the next 5 to 10 years. And so I'm in the board of directors at Bossard, which is a publicly listed company here in Switzerland. You should all know it. And actually my colleague from the board is also here very excited for that. And then also at DKV Mobility, which is privately owned based or headquartered in Germany. And until a few weeks ago, I was Chief Innovation and AI officer at Ringy. And I also see somebody from my former team here.

So thanks for joining. So that's good for for others to know about me. I would say I think that's really good because that gives you the the sort of the business side. And you were also at Google for a while. I used to work at Google before I was hired to Renee worked there in various functions. Basically, I think one of the thing that was constant throughout all those function was it was about creating new products that led to new revenue streams or LED to new user experiences. And so the last one was in augmented reality where you might have tried it out on Google Maps when you got the navigation in airports, shopping malls, train stations, because people spend 90%.

Percent of their time with in indoor venues, not outdoors. So that was important or before that an advertising helped Google shift revenue to YouTube through closing some new deals. That's great. So exactly that's this red thread for me for you has been really looking at user needs, user opportunities and and how do you innovate into that right and that's great. Great great. Thank you Pedro, if you want to grab a seat for a moment. Thank you very much. And our a Tejal colleague who is going to be joining us today is Professor Stefano Brizoni.

Stefano, come on up. Please give him a hand as he comes up. And the same question to you, Stefano, what does everybody in this room need to know about you that's relevant for tonight's conversation? Back a little. I I study innovation processes, obstacles to innovation and technical change since the mid 90s, looking at how digital technologies are developed and most of all adopted in user industries. I focus very much on this lengthy, difficult, nitty gritty kind of processes that enable users adopters to actually get competent at the new technologies.

And these processes are usually very slow, actually full of mistakes and and misunderstandings about what the technology is actually good at, as our first reaction with the new technology is always to fit it into the categories of the old technology. So we got computers and for 10 years we use them as typewriters without paper. Sure, but you can do a few more things than than that. And you can see this all the way from digital control systems and jet engines. That was my industrial thesis to robotisation, automation already, machine learning, blockchain.

You always see these patterns. That doesn't make prediction any easier about solutions do not, but the pattern is kind of clear. That's really great. And so Stefano, thank you very much. And we're going to get started right now with that conversation. So as I say, we have two parts to this, this dialogue really one is around innovation and then how we thrive in this, this time of AI. And if you think of the word innovation, we are at an academic institution. So we can start off with a little bit of a definition. The root of the innovation is around renewing, innovare to renew or to alter.

And so it's very interesting to think how we define innovation for ourselves. So how do you define innovation? So to me, innovation is giving longevity to a company, making it invincible so that it will thrive in the in the future. And basically to do that, you got to ensure that you build value for your customers. And so innovation, maybe it's easier to think about what is not innovation. So innovation is not a bunch of geniuses who sit in the room and they tinker something crazy. Innovation is not a fun colourful park with nice cushions etcetera.

It's hard work and innovation is not. You throw billions of dollars into this one brilliant idea and then everything happens. But innovation is really a lot of exploring. You're going to have a vision where you want to go, your North Star that you'll follow, but then you'll you'll explore because you don't know which of those bets that you'll have will work out. So you're going to have constant iteration. And now on the entire innovation spectrum, you can have like very efficient innovation, efficiency innovation or very transformational innovation.

And so for efficiency, it's yeah, you can see that basically in the car industry a lot. So you get a navigation system, it has an airbag, etcetera. So the car gets better and better and better, but it's still a car. Whereas on the transformational part, you enter a new market, you explore a new geography, new users, new services, etcetera, and both sort of has to work together if you really want to give that longevity to the company. It's fascinating, the longevity, the lung value to help it thrive, survive and make it through the different waves.

Stefan, what about for you? How do you define innovation? You're the professor of innovation. So at some level there is a statistical definition, a statistical office definition of innovation, which is the first commercial application or a new product or a new process, which is a beautiful definition because it's assumed as a way everything which is important, which is the process that lead to that first commercial application. It is important, nevertheless, because it does emphasize that we, we, we count. Innovations, if you like, the moment there is a market hit, do what Ian, there is a presence on the market with a new product or a new process.

And that's important because in the end, again the longevity of an organization, which is a very important concept comes from being competitive. And there is plenty of evidence that the main driver of long term competitiveness is innovation, right of any kind by the way. Then there is discussion there what is most important, when and where, but that is what makes companies sustainable in the long term. Again, the evidence is super clear. Now the challenge here is what that very few companies are really good at becoming systematic innovators.

A lot of companies, a lot of people are very good at having one good idea, which is excellent, but very few are good at being systematically good, right? And to understand that difference between the occasional innovation still important versus the the more systematic ability of remaining innovative in the long term, then you have to look at the processes that Petra was talking about. Because it's this process of trial and error and then error and then again an error where you stay persistent as an individual, as a team, as an organization that that really set apart the occasional innovators from the systematic innovators.

That is the most important thing and the most difficult thing to achieve. Because again, it's not one person is a whole bunch of people that work together or not in different ways with different backgrounds that over time have to remain organise and and committed, even when they're going to be facing more errors than successes on the, I like to add something there because I think it's so important what you just said. It can be everybody who can innovate, but a few of them do. And it starts with a processes. And I think processes are super important.

And also how are those processes implemented? There's a company for instance, in China, Ping An, and the founder told the new incoming Co CEO, you're going to fail. And what he actually meant was you're going to learn to experiment and a bunch of those experiments are going to fail. And if that happens at the very top, then they have authority to really try and then it can work. But it has to come from actually from the highest level so that it actually thrives. And there must be evidence that the highest level is actually serious.

When when then people fail because that's painful. Because again, innovation, if you look at the process, that's fundamentally a learning process. And learning processes are super important, but they're usually not efficient because they imply mistakes. So if you want to translate that is is a waste of resources in the short term. In the long term, Nope, That's a spring for swing board for something better and newer. But the the commitment and the persistence are fundamentally important to go from the bright idea of which there is no shore supply generally to the actual implementation in practice and at scale.

Because then there is also the issue of scale to, to bring in, which is another problem itself. Fantastic efficiency versus transformational, occasional versus systemic and the focus on the processes, you know, when I was doing focusing on innovation a while ago is always a practical application of a new idea. Let's just use that as my definition. And you've, I liked how you put it into really focus on a corporate standpoint. And you mentioned here a little bit about culture and we started this talk getting into culture and how important corporate culture, company culture, our own culture is.

So, so tell us a little bit about mindset and culture and the importance of that when it comes to the innovation to be me, I'm looking straight at you and Patrick. So you go for it. All right. So when it comes to culture, what Stefano, I actually mentioned is that that authority to fail, that's an important one. Permission to fail or permission to fail, That's just wondering. Authority to fail or I'm just wondering. I think there was a wrong word. OK, Excuse me. That is, that is super important and it has to come from from from the board.

So the question then is how do you actually bring in that experimentation And the way how I did that at Ringy was so when I started, I asked everybody saying 5 questions and I realised there is something new that I have to build if I want to bring in innovation. And Ringy, I'm not sure how familiar with a company. It has 140, fifty different brands in 21 different countries. So I can't just work one-on-one with all of them. So I launched an Accelerate. Called Ringy X and in that accelerator we brought this mindset into the companies and then the accelerator started to act like a multiplier into all the different companies because they were teams from from Ghana to South Africa to Poland, Slovakia, etcetera in the accelerator who experienced how to work in that way.

And So what we actually nailed them down is we had a scale of five different levels of evidence that they had to bring. And when everybody starts, you start at level 0 most of the times and then you move up the ladder. Whether you have user interviews is probably the weakest level of evidence. Then users who sign up is already stronger. Users who actually pay is the strongest. And so you move on that scale and you try to bring in that mindset, bring the evidence and then you can continue your product, building your product.

So this is how I tried to bring in a different mindset. It's very interesting. So how for me, it's an interesting question. Mayor to both of you, how important is it to create this also the safe space within a corporate setting to allow this to this to bubble a little bit before you send them back into their P&L WHIP, right. No, I mean, you know, it's it's interesting because you have to yeah, yeah, yeah. So it was a three stages process. They started in the seed program, then they moved on to start up and to scale and then once they graduated, they went back into into PNL.

So it was a transition phase phase there. But while they were within that incubated, they were like on a a bit cushioned and dependent then on us how much information we were passing on. And my team and Thong actually sits here coached those teams so they they could learn the ropes and start actually getting that evidence that was required. That's interesting. Sort of infiltrate a little bit. It's my word, not yours. It's just so you say, can I find out? Sometimes you have to do that. Stefan, what about your observation over the past almost almost 30 years of looking at innovation within a corporate setting?

Does it need to have its own little place where it can start typically or does it really does it work when the when the CEO, he or she or they say you shall innovate? I think it's always good to start small. And I think these innovation labs or innovation ecosystems who also open up boundaries toward external environment are useful. The challenge is that sometimes they're very much designed as, you know, following the the saying, whatever happens in Vegas, stay in Vegas. It shouldn't, it should go out. And that's the most difficult translation because again, you can train people to to become more innovative as individuals and in teams, you can get them to develop new ideas to prototypes repeatedly.

But then that thing has to go to some business unit existing or or new and that is the P&L whip that comes back. And and that's a very dangerous moment for any innovative ideas because that's bound to impact people with P&L responsibility that we must respect because this is how salaries are paid, right? And and they become suddenly careful because we got the good thing here going. Why do we need to change it? Because in X years it would be better. Sure. But look at my KPIs now. And going back to your point about culture, I think two things.

Culture is a very big object, right? And, and I like it to simplify it into three things for the sake of being able to design a study. So my life is simple. One is skills, which is the party we're referring to. We have to train people, to educate people to, to think a bit more broadly, a bit more innovatively, a bit more verbal looking. That can be done. But then there is the the piece which is about motivation and incentives. We have to give them the right KPIs and clearly, particularly in large organisations, the KPIs are bound to be oriented to financial outcomes.

In the short term. That's difficult, that's understandable. It's easy to criticise people. People like me like to criticise that because we have the luxury of designing project that can last five years plus, which is why I'm here. I'm very happy with this. I'm not complaining, but I also understand that people in industry don't always have that luxury. So the challenge here is realigning the skill development part with the motivation and incentive part. When you realign those two then you see different behaviours and then you can also scale and these three.

Many pieces, skills, motivations, behaviour. Are there moving parts that we change corporate cultures, but it takes time. Do you agree? Yeah, I, I do agree on that. And so you said skill. The second one was motivation for sentence and then behaviour will follow exactly. For the incentives, it's actually tough because most of the incentives, the whiplash is EBTA and the moment they can't get out, they immediately measured against that. And obviously then at the beginning they look really bad and then it depends how long of a lean way they're given.

But what actually works to turn that now into a positive is in one of the boards, we're measuring new products being developed and as an aggregate in the past three years. So every product that is up to three years old. And so I think that gives sort of an aggregate and it also gives an indicator at the highest management level how innovative the company actually is. Really is. Yeah. Or not. Yeah, that's really great. OK, let's let's lay on now the age of AI. It's three years, it has has an early birth date, it's much longer, blah, blah, blah, blah.

We know all that. But this is what we're here now to to think about how is this age of AI going to be impacting this innovation process? Pedro, how do you see this? The innovation process by itself, I don't think that it will tremendously change because you still got to build a product that a customer wants, needs, buys at the end of the day. So you still got to get the evidence. Now what will change is the velocity, how fast you're getting to that point. So with all the different tools, use lovable, use N8 N, use cursor, etcetera, you can build prototyping really fast.

Then you can throw them out and see where you got the biggest traction. But then you still got to build the product eventually. So the velocity is one of the key things that will change. And another thing that will change is we enter an era of hyper personalization. So I think the mass products we'll see sort of an end, but at the end of the day, you'll have your tailored product, your tailored news catalog, your tailored medicine to your needs, what you actually need. And that is possible through AI. And I think here we just stand at the at the very beginning.

OK. Stefano, what about you? I very much agree that that speed and and customization are two things that we'll need to monitor closely because that's where we are likely to see new things on the process. I agree the fundamental building blocks are not likely to change. One thing that seemingly at this stage at least and then in six months it will be different, is actually qualitatively hard to compare to beforehand. Is this so? We always use technology to innovate instruments of a different type. Take a pocket calculator, just as a very simple example.

A pocket calculator is a super useful instrument to solve complex problems. It's very much integrated in everything we do in many ways or computer if you like. Now, the peculiarity of the pocket calculator do is that in order to use it, I have to give it instructions, sum up these two numbers and gave me the result and then I plugged the result into whatever else I'm doing with with generative AI in particular, yes, you can use it that way, but it works way better if you share with the machine. I think it's important to remember this is a machine why we are solving that problem.

What is the goal? What is the audience that can potential customers? It's, it's a way to use it that requires that we share our problem representation, not only the specific step in the problem solving process, but the what is the overall problem representation that we have in mind at that specific point in time. Because if we can do that, that's what we do with prompting, by the way. Yeah, if we can do that properly, then the machine becomes way more effective, giving us more than the sum of 2 numbers and suddenly we are engaged in a conversation.

Which we didn't have with the pocket calculator that is actually shaping the problem. Not so, not so much of not yet the solution. I think that difference that we all experience again, when we prompt is my current candidate for the most important qualitative difference in the innovation process itself. And probably we've got to put even more thought in that than than before. And you've got to be competent, right? Because you, you have to be able to explain to a pretty clever junior colleague, I mean, very clever, very committed, but kind of incompetent at the job at the beginning what the job is about, right?

It's it's hard work and it does require a certain level of skills. One thing that came to my mind on the way here, it's sort of related. I read a LinkedIn post of her friend complaining about tons of AI generated LinkedIn post and that LinkedIn starts to become very cringe. And that actually goes to your point with AI, you can't just blast out so much different content and then the question is who actually wants it? So to your point again, yeah, you need to even think harder. What is it that you want to convey? So, So I think to me, I'm looking at two different issues here in a way.

OK, One reminded me when I was heading up the R&D group and I asked you guys, what do you do these days? We answer questions. This is your 2 + 2. I said, what do you do? And he said, well, we never answer the question. I said you just told me you answer questions, Said no, no, they asked. Someone asked us a question. We asked them why we don't give them an answer. This is what you're saying. We have to actually interact with this tool to help explain the context, to find out what really is the right question. And so wait, this is just a generative part to queries and challenges you and you're pointing out something which is other kind of tools which create this new span, this new level of spam which we haven't experienced yet.

And how do we deal with this proliferate, this capability of a proliferation which we never could have before? That's sort of the age of AI trying to understand this context. I'm thinking you mentioned front of house, but I'm thinking back of house. You're talking about talking about people who have new products. But I'm also thinking back of house where this age of IAI is, is will is impacting and will impact from an innovation standpoint. So tell me a little bit about your thoughts on the back of house part of it, not just, OK, that sounds good.

So what I'm observing is lots of companies, they go now into, oh, we can automate this process. We can automate that. We can make this faster. And this is all good. This is all the low hanging fruits that you got to pick. This is your homework. This is what you got to do. And what I also observe is irrespectively which company is the pleasure to do, Oh, we do 130 use cases. Oh, we have 66 use cases in AI. And just by the quantity, again, people are excited how much they do. But the big question is not how much you do, but what is the impact of it.

So what was important to me at Ringy is that we introduced high impact use cases where you want to focus on. So I think this is for the efficiency type of innovation. But then there is an even another untapped field where few are looking into. Most look into processes that they can change in the finance department, in the marketing department, in the comms department, everybody for themselves. But what if you think across different departments? What if you try to think about and you, how do you, how does the vision look like of your industry actually or so across the industry where you want to be at?

And how can AI, if you stitch 2 departments together and build a process, they are across the boundaries, how can that then impact potential customers down the line? That is where few companies are at yet. But I think that is what will fundamentally change. You will not win by just automating everything because everybody is doing that. So this is just being at at eyes level, but you got to think beyond that. So how can we actually transform across different silos, which makes many people very uncomfortable? Maybe yes, maybe no.

I mean, some people like it, some other people don't like it so much. Change. Change The strange beast. Yeah, We had just the conversation about that before. Strange. No, I mean, it's interesting because I think that's where some of the culture and some of the threat comes in. You're blurring A boundary. And I've liked my boundary because I have my chair. And I don't know about your chair. I like my chair. You know, this is sort of a very interesting thing. I would like to remind everybody who, did anybody have a question yet?

OK, there's a couple hands up. So if we have, we can bring the cards and pass them down to those who put your put your hands up so the cards can come to you. It'd be really fantastic. And we can leave them up there for a minute. So I see, I'm going to give you a couple of minutes to write the questions out. And I please beg you to use your best handwriting. All right, don't use Saturday night at the pub handwriting. Use your 7th grade handwriting class. If you had such a thing like I did, I don't know if you teach handwriting anymore, but it would be really helpful.

This is the first time I'm sitting in an audience where there are so many hands up. Well, right instantly. Oh, come on, Pedro, this is Etihad. This is look at this is this is our audience. This is our people here. So I want to as they're coming up to your questions, let me ask you both. I mean, just looking at that, is this a good time to start a new company? I think yes. I think now's the time when if you look at the top ten companies according to market cap, it's all tech companies right now. If you look like 20 years ago, it was the Exxon mobiles, Walmarts, etcetera, GE, etcetera.

So it completely changed to tech. And I think if you look 10 or 20 years out again, I think now is probably the time where new companies are getting started that will then that are being built right now, that will eventually show up in this top 10 ranking. I hope so they can start right here, right? This is the this is the important part. Stefano, what do you when you when you look at the innovation that I'm going to ask you about the same question about starting a company now, but back of house, front of house, you know, because Pedro's talking a lot about front of house.

How do you see it from an innovation and an AI standpoint now this differentiation between front of house, new kind of sexy stuff, back of house, the old processes. I mean, I think that the tension has always been there. I think irrespective of what kind of technology is now driving the wave of, of growth and start-ups and and stuff and both sides are important. I think we tend to to aneurysm in the second one. I think what I would like to see more in in in the back end or what I call the back end or the processes is really two things.

One, acceptance of the idea that to be innovative, you don't need to have a bunch of crazy people trying out randomly a bunch of new things all the time. You, you need a way of organising these people, right. And, and the nitty gritty part is important. Of course, the processes have to be flexible and all the sad, but you still need processes And one very important process that right now I think is very important, quite underestimated and generally disconnected from the discussion about innovation at the times of AI is their old learning and development training at scale.

So there is obviously everywhere there's requests for upscaling. We, we see it very clear at the age, but while the the the sort of the calls for upscaling are very broad and very loud, that doesn't translate in demand for upscaling. And the difference is willingness to take part of the course a class. And willingness means people don't have time because what times are difficult And the business cycle is, is, is is complicated because there is a willingness to pay argument as as one of my roles here is is vice director for for vital building.

So we do have offerings and we do have general interest. We see also applications but not in the numbers that you want to see. And we see companies being less and less willing to support their employees in in joining courses generally across the board. And we see people very cautious in taking time off their work to upskill because because things are quite complicated everywhere. People don't want to take time off because they want to be sure there is still a job to go back to in in the end. And I think that's part of a reflection that we should have also about what kind of programmes we offer.

Because the kind of programmes we offer, well, we design them in terms of formats, not content those we change, but in some of formats a long time ago. So that's kind of long and I think a bit cumbersome for for the needs in the industry right now. So I want to say on the skills thing for just a moment, because because I know, as you said, you're the Vice President. For continuing education and Petra, what kind of skills would you suggest that Stephanie should be making sure that he has and his continuing education classes in order to be really ready to be able to thrive in this age of AI?

What would you suggest with some of the skills which we need to be ensuring that our students and the Vita building students have? Well, if you go back to the discussion that that you actually started, the most important one is having some very deep knowledge that AI cannot have in an in a domain field. But then also having that critical thinking in order to make up your mind where you want to go. Because one of the most important things is you got to know where your North Star is. You don't know which path you're going to take at one point, but you got to know where to go.

And so that critical thinking is, yeah, crucial. So the, I think that is super important. And again, there are two sides to the discussion. One is the contextual knowledge and, and the contextual knowledge, we don't have it. That's people from industry who have it. The the, the, the activities, the programs we have tend to emphasise critical thinking and there is a bunch of ways in which that can be evaluated, assessed, even measured if you like, and then applied. But the difficult thing is really combining the contextual knowledge that comes from the participants of our studies to our programmes with the skill development abilities that we do have developed over time.

For example, one thing we are trying to increase as much as we can is, is project based learning because in the project the students can bring in the context. If we make up a fictional project, we can and it's good for developing critical thinking, but it works. But then that doesn't give you necessarily the ability of then explaining any general AI tool. What are the, the boundaries of the problem that you're trying to solve in that specific context. So we, I was referring earlier to these programs we ran over 18 months, which I love the early and they're excellent programs, but it's 18 months of a program and right now the pace of change in industry is, is a little bit faster than that.

So we are trying to breaking this down in smaller programs. We're going to go into micro credentials so people can take a few days of a workshop. But I think the challenge here is finding ways in which industry and university, not only us generally university and industry can Co design these programs because we look at different contexts at these high here. But to be able to interact with these machines cleverly, you need the understanding of the context at this level of depth. And and we need to somehow to bring the 2 levels together because otherwise this upscaling remains something that we can sort of deliver in a certificate good.

But in practice then there is another leg there are that needs to be kind of dealt with. Yeah. And maybe two other points. I think you made me think in terms of people need to become like AI natives, not just AI experts. And so probably that is one thing that is a bit more detangled from industries where ETH can go more at full speed. And the other one is probably communication slash team play. This is also important. I mean, there is always a technical core to our programs and I think people should become competent at different AI tools technically, but then they have to leverage that level of technical competence, which is kind of well codified and can be achieved with the context and with the critical thinking and the ability of engaging others.

It could be communication, it could be something else. But again, you were saying at the beginning that there is no one single innovator, always a team. And, and one of the biggest challenges that people face when they leave our programmes, I think is transferring back what they've learnt. And that transfer to some extent is engaging their colleagues also in learning and education on a real project. And that is to down to communication and to read the ability of explaining why this stuff is important now. OK, that's the case.

Just trying to take back. All right. You've done a beautiful job of filling the time so I could read this, read some of these questions. You know, I knew Stefano could do that, you know. Yeah. So in all, seriously excellent questions, really. And I hope I get to all of them. So, Peter, I'm going to read this because I like your mission to bring ideas to reality for the right impact. What are the biggest challenges you see in this journey? That's actually an interesting question for both. When on that journey, what are some of the biggest challenges you see on that?

So I'll start maybe with with the beginning when you have an idea you can argue against or in favour of that idea and whether the idea will work, nobody knows. So it's like, it starts like it's like an ugly baby nobody wants to look at and add, Catmull said that. And so parents want to look at the ugly baby analogy is kind of interesting because the ones who have the idea are in love with it. But then for others to adapt it, it's really tough. And there is where already a lot of ideas get filtered out or deleted that might have become the next big thing.

So it starts even at the very beginning. I see. But what about when you're like, when you're when you're pitching to something? I mean, you know, very often it's you part of communication, right? You have the best idea in the world, but if you can't share it or can't communicate, it doesn't really matter. It is, I think it's two way. It's one thing that you're sending the information and the other way is the information that people want to receive. You might pitch it to people who just don't want to hear that idea.

So around pitching someone's asked a question is how and you pitch an idea to a large company in the age of AI or algorithms rather than humans and HR screening candidates and retrieving proposals, what do we do? So it's, you know, very often, I know from my daughter, she sent out, you know, hundreds of resumes and she said she got very few responses that were actually seem to be written by a human. That's in this age of AI. What do we what, what do we do right when the algorithms take control of the gatekeeping. So I think if you're a student and you're trying to apply for your first job, obviously cold selling cold application is one thing, but it has never been the most successful way to land a job.

And it will become even less successful path down the line because of exact the reason that you mentioned. So I think probably starting being in touch with companies early, working your network will become even more important than before. Or maybe you have the opportunity to work with a company and say, hey, I have that idea that you should do. And you find somebody and you start just working with them on a occasional project to get your foot in the door and that they see you what and what you can do. So probably a shift in that path.

Do you have a view on this, Stefano? I think that is one of the single most pressing concerns I have about the use of AI, which in HR processes generally has been applied since before Gen. AI was there. The whole big data analytics things. And to some extent I understand the reason when you have hundreds of thousands of applications to deal with because much of companies you want to automate and make it more efficient. Now the challenge with that is this makes as you say the personal connections super critical. Now the problem with personal connections is that they tend to be conservative and clicky and and that reduces the opportunities the students that come from social background that put them in a position disadvantage can actually leverage.

And I think here is, is a very good example. I'm not saying that we shouldn't be using AII mean that's unavoidable. That's fine. But I think the process of designing these personal connections, this project is super important because as a higher education institution, our job is very simply that of making sure that at the end of their studies, our students have more opportunities than at the beginning of their studies, right? That that's a job that's a very simple, irrespective of where they come from, correct. And and some of these technologies that are deployed for understandable reasons are pushing against that.

So I think that is I don't have any solution to a firm or not. But in terms of problems, that's a big one. Interesting. More opportunities when you graduate than when you had coming in. That's a job. Yeah, I like that. That's good. That's I've never seen that in our job description, but I'll add that now into our entire strategic plan. I think that's good. No sure.

Yeah yeah yeah yeah yeah. Oh no, there's a lot of waste in in the process. Process ran this way, Absolutely, yes. I'm not saying but, and I think that is super important for us as an institution, for the companies, as institutions which are seeking new talent and they might end up actually limiting their surge in the effort of becoming efficient, which which is the the typical paradox when technologies are new. By the way, you looked, Petra, you mentioned looking beyond process innovation for AI use cases. Can you give an example of a cross department AI innovation that you could get excited about or you've seen one?

It's interesting because it's blurring of the boundaries, getting an example of what something could be. Yeah, with that, yeah. So I think there is what I mentioned at the very beginning is start with the user 1st and think about what the user actually wants. Because your job as a company is that your customer is happy. And so instead of thinking my customer comes to me here for doing this job and then comes here for doing that job, think about what is the spectrum of everything that we're offering from, I don't know, fuelling your car to providing payments, from reading news to a job to real estate, like whatever you have.

Like what does a customer actually want to do? And then offering those solutions to the customer right away. Instead of offering one product for this, offering one product for that merge them a little bit. OK interesting. And someone actually had a question here on saying the post app. So the post app can do so many things, but only people usually like 2-2 functions. So the question is, is, is that, you know, is that how often if you is it either push or pull? I guess it's it's offer or receive. So I'm trying to, I'm trying to think about that when we think of how's it going to know what I want?

How do you how do you know what the customer wants? Yeah. Does the customer know what he or she or they want? And that's dangerous. Actually, you should not figure out what they want because they changed their mind. And Steve Jobs actually once said, the moment you ask the customer what they want and you build it and you have the product and then they say, what the heck? The the The thing is rather which problem do they face? Yeah, that's a good. You got to figure out where are they stuck? Where do they spend lots of money?

Where is a process that is so painful that it takes them a team of seven people for two weeks to get this step done? If it's just providing a quote or if it's just providing, I don't know, whatever, and then figure out what is the pain point and solve that pain point. But if you start asking people what they want, then there you go. Yeah, I'm with you 100% on that, 100%. Stefano, you're looking, you're pondering. That's a professional requirements. Ja.

But now that I feel committed to say something, I think this is super important that the the problem and the need not not the want. Because honestly, we're not. I will not ever do underestimate the fact that the companies who are good at that are the companies who also have not stopped to invest in research and development. Because while the users and the customers are the sources of the problems we want to solve and then possibly monetize a bit, by all means, the source of solutions is still technology science and technology research and development.

And, and there is a tendency that we see across the board in different countries with exceptions like Google being obviously one in reducing the investment in research and development or shifting away from research, fundamental research and to our development. And we see that also in universities less at TTH obviously. But again, think about the average, which is important. Universities are becoming more applied because we also need to raise more funds from industry in collaboration with the industry, which is good important.

But again, once we know the problem a potential client might want us to solve for them, we still need to have the solution. And that comes from research and development. And a lot of that historically has been publicly funded. And we see less and less of that. And we see less fundamental research also in industry on average and more and more driven by few giant companies. OK, there are issues there of of of governance you want to be careful with. But generally, we need the problem that comes from downstream from the clients, but we need the solutions and, and those solutions come from science and technology.

And I think one of the crux is that you should invest in innovation when you still have the funds. However, once you still have the funds, you're actually thriving. So, So why, why should we innovate now? I mean, we can do that also later. But yeah, I think companies were successful. They're investing maybe 10% of their revenue into R&D and new business development constantly. Yeah, if you're doing, if you're investing 10%, they're on the high end of the of the spectrum. They're on the high end of the spectrum. But yeah, it, it's like it's a habit, like you go to the gym if you want to be healthy.

And so in the company, you yeah, have your toy. We see on that one, you know, 100%. So here's another question from probably one of the youngest members of our audience here. Where is the where's the most efficient, where is it in a company most efficient to use AI, Right. So within a company where I'm interpreting your question, young, young man, I'm just working on it, OK? It's where is it in the most, the most efficient as a company to use AI. So where would you suggest for a company, where should they be using AI?

Like if you're going to go into a company and say, OK, where are the first places they should begin to look to integrate AI Gen. AI or different tools, machine learning, where would you suggest? Did I get the question right? All right, good. Thank you. I think 1 is definitely interactions with customers and we're seeing that we all I think hated to be on the phone line and waiting and then answering your question and the person didn't know. But I recently had a very pleasant experience. I'm having this aura ring and the battery was broke so I chatted with the AI and it immediately detected your battery doesn't work and we'll ship you a new one.

Here's your address. Is it correct? And I said yes. And so that chat was literally maximum 2 minutes. And the next day I had a new RR ring at my doorstep. So those are very unpleasant experiences, but you can turn them into a brilliant experience. Five years ago, I would not have said I'm talking about that experience because that would have been, that would have been quite fun. But so in customer interaction, there's probably one one way you could start in market research or summarising large piles of documents. That is also where you can immediately leverage impact.

But this is this is one thing. And obviously, if you use a Gentek AI, you can actually have an entire Armada who actually creates your new events, writes you a report, etcetera. So there are other areas as well. Very good. Definitely you're nodding also professional requirement. But I think this is a very good point. I think it would be important to consider that AI is actually. Different things, right? And there are applications which are very specialised think of stuff like Alpha Fold for example, which has been delivering incredible support to to the the pharma industry for quite a few years.

Other tools are more horizontal in a way. And then different companies, a big like different users will have different problems, right? So this, this matching process is, is important to be to be made to be made explicit when one thing about AI, because again, there are many different tools at this point out there and the number of tools is increasing by the day. I think that's a really good point actually, when we're talking about a whole family with cousins and uncles and you know, all these, it's not just a one thing.

As you said, there's a general models of specific ones and here's someone's ask that says what can we do as individuals to thrive in this age of AI? How do we take charge in the face of this disruption? The question for both of you, how do what can we do as individuals to thrive? What advice do you have? I think AI should become like a daily habit or you, it's like your go to tool for doing this or that and you got to learn how to build your agents. You got to learn how to maybe automate certain processes in your inbox.

That's or wherever other tasks that you don't like so much or Yeah. So becoming really, truly AI native and use it constantly. Read up upon it as well. Do some courses. There are a bunch of cool courses out there. And also certain companies offer like hackathons over the weekend. So it's a great opportunity to be curious about it. That's the first. Be curious about it. That's the first thing. Yeah. That would be a great first step. Yeah, yeah, Stefano, what about you for this? Because it's actually a really important.

I think it's a really important question. Yeah. I, I, I will certainly build on the curiosity. On the what? On the curiosity. Yeah, being curious about it. Be sceptical. Going back to the critical thinking part, not cynical. Cynical is no good. It's very boring. Sceptical. Don't trust when people show up and say that this machine is going to solve all your problems because it will not. It will do some good, but not exactly. It's not exactly clear well. So be sceptical. To be sceptical and open minded simultaneously then there are options, very large number of options to become competent, to become native or at least to be not adopted into into the family.

But I think this is a good point in time to be a little bit sceptical because you are mentioning good time to set up companies. Very true, but absolutely true. But a lot of these new companies that pop up every other day make promises that they could deliver actually without AI. So if you you are engineers, scientists, check it out. It's like when there was the blockchain explosion. A lot of the stuff could be done with Excel, right? And then the market will will catch up over time. You know which is which, but we we don't have the benefit of time now.

So we are in the here and now. So there is that element of scepticism that you can leverage to find the time to take a course to become a bit more competent and really form your own opinion about what's happening and what is in it for you. Great, thank you. Now maybe adding also very, very, very short because we're running out of time. OK, very short. I think leaving grad school, you should be very optimistic because if you look at companies you're, you have such a, such a solid training and you're used to those tools inside out.

Nobody in those companies has those skills. So I would use that as a huge leverage to bring in a new perspective, new skill sets into that respective company. I liked it. That's a great answer. So we just, we're going to do a very, we're going to make a trial. These are short answers, like two or three words. And we've got 3 minutes of this because I'm going to try to get through the rest of these because these are really great questions. How do we resolve AI output? Explain ability since the architecture is heuristic and the output is not deterministic.

Short answer, we don't. You don't. OK, there you go. Pedro mentioned transformation and efficiency innovation. Stefano said we need systematic innovation. Can you please name one or two companies which exempt? This which we can take a look at and you have one or two companies, we could say they are great examples to take a look at. Can you name one or two in recent times? Google, certainly IBM, historically, we can always discuss about the ups and downs of IBM, but they've been bloody good. NVIDIA, the transition from video games to what they're doing now.

No, no plant. I mean Nintendo. Nintendo. OK, there we go. So we have 4. That was one or two. We got the name. One metric that should be used to decide to scale or kill. One metric. That's great. Yeah. Just like user adoption, user adoption, rate of user adoption, acceptance, there's a difference for AI.

Ultimately, I think this is the Apparo question. Whoever wrote that you go, you can come on the Apparo. Where does innovation really start today? Traditionally it was driven from customer needs. It says something changed, it's still customer needs or is something changed? So that can I say traditionally yes, if the body of science and technology was there to offer solutions you need both. And historically we need science and technology, the supply side and the the users, the customer side, we need both in different phases of evolution.

1 is dominant of the others by all means. But over time both always you can with customer needs definitely but also product offering that you have. OK, I still have 8 questions. We'd have no more time. So I'm very, very sorry, but you didn't get your question asked. We do have an opera afterwards and I think what I'd like to say thank you to both of you for a really fascinating conversation. Let's give them a a round of applause.

Now, this is our last global lecture of this year, but we are looking forward to a new year when we're going to have more global lectures coming your way. On Thursday, the 5th of March, we're going to be welcoming Pete Warden, the former director of NASA Ames Research Center. We're talking about space. And I think this is going to be a really, really, really interesting lecture. Very different domain that we're talking about tonight, but it's going to be a great one. So please can join us for that one. You know, when I, when I think about what we've talked about here, but innovation was an AI wrapper, it has struck out.

It has struck me how important both topics are for all of us in this room to really be curious about and to learn as much as we can in order to be prepared for what's coming. I think that it was very, for me, very inspiration. I want to thank you for that and for your willingness to explore these themes with me and with us. And the final thank you is to all of you to, for, for choosing to join us today. It is really wonderful to see you here. And also, please consider supporting ETH through the foundation because the foundation allows us to support students in their journey and faculty in their journey to make these kinds of dialogues become more meaningful, more impactful, and more real for all of us and everyday.

So with that, I'd like to say thank you very much for all of you. And one more round of applause for our wonderful guests.

Questions about this episode

Who speaks in "Global Lecture Series: Empowering Innovation"?

Petra Airman (Guest), Chris Lupkamen (Host), Stefano Brizoni (Co-speaker / Referenced expert (Professor of innovation)).

What is discussed in "Global Lecture Series: Empowering Innovation"?

In order: 0:00 Petra Airman, Chris Lupkamen, Stefano Brizoni, Google LLC; 10:00 Ringy, Ping An; 30:00 Artificial Intelligence; 40:00 Google LLC; 50:00 Google LLC.

What is mentioned in this ETH Zürich episode?

10 people, works and ideas are mentioned, including Artificial Intelligence, Petra Airman, Chris Lupkamen, Stefano Brizoni, ETH Zurich, Ringy. Each one on this page links to its own page with every episode that mentions it.

How long is "Global Lecture Series: Empowering Innovation"?

The episode runs 1 h 2 min; the transcript is about 9,928 words in 94 timestamped sections.

More from ETH Zürich

About this transcript: the episode's spoken words as captioned on YouTube, split into timestamped sections. Reclip.Me locates every person, work and idea in the text, keeps the time it is first named, and quotes the words around it verbatim. Last updated January 24, 2026.