LØRN Case #C0421
AI and transformation of the workplace
Jarle is originally a social scientist but became interested in technology when he was studying history and saw the transformational impact technology has had on society. He has a Ph.D. in innovation studies and now works with the development of long-term knowledge that can help Telenor with making good decisions in the future. They do this with the help of machine learning, statistics, and text mining applied to data about employees and customers. To name one example, they have developed a machine learning model which predicts voluntary resignations in Telenor. In the episode, you can hear more about this example and what its controversies are, in addition to other examples of how Telenor utilizes AI to transform its workplace.

Jarle Hildrum

VP Research

Telenor

"AI and machine learning have a great potential for transforming workplaces in a positive way, but for us to achieve these good effects, the introduction of this technology must be done in the dialogue between developers, employees and, managers."

Dette er LØRN Cases

En LØRN CASE er en kort og praktisk, lett og morsom, innovasjonshistorie. Den er fortalt på 30 minutter, er samtalebasert, og virker like bra som podkast, video eller tekst. Lytt og lær der det passer deg best! Vi dekker 15 tematiske områder om teknologi, innovasjon og ledelse, og 10 perspektiver som gründer, forsker etc. På denne siden kan du lytte, se eller lese gratis, men vi anbefaler deg å registrere deg, slik at vi kan lage personaliserte læringsstier for nettopp deg. 

Vi vil gjerne hjelpe deg komme i gang og fortsette å drive med livslang læring.

En LØRN CASE er en kort og praktisk, lett og morsom, innovasjonshistorie. Den er fortalt på 30 minutter, er samtalebasert, og virker like bra som podkast, video eller tekst. Lytt og lær der det passer deg best! Vi dekker 15 tematiske områder om teknologi, innovasjon og ledelse, og 10 perspektiver som gründer, forsker etc. På denne siden kan du lytte, se eller lese gratis, men vi anbefaler deg å registrere deg, slik at vi kan lage personaliserte læringsstier for nettopp deg. Vi vil gjerne hjelpe deg komme i gang og fortsette å drive med livslang læring.

Vis

Velg ditt format

Varighet: 39 min

Ta quiz og få læringsbevis

Du må være medlem for å ta quiz

Ferdig med quiz?

Besvar refleksjonsoppgave

Du må være medlem for å gjøre refleksjonsoppgave.

Hva er det viktigste dere gjør på jobben?

Vi jobber for langsiktig utvikling av kunnskap som kan hjelpe Telenor til å gjøre gode veivalg for fremtiden.

Hva fokuserer du på innen teknologi?

Maskinlæring, statistikk og text mining anvendt på data om ansatte og kunder.

Hvorfor er det spennende?

Fordi det gir store muligheter til å utvikle kunnskap som ledere kan benytte direkte i å forbedre arbeidsbetingelser for ansatte i selskapet, samt styrke prestasjoner.

Hva synes du er de mest interessante kontroverser?

Etiske implikasjoner av økt bruk av AI i samfunnet. Vi kan se dette tydelig i prosjekter hvor vi bruker maskinlæring på ansattdata.

Dine egne relevante prosjekter siste året?

Utviklet en maskinlæringsmodell som predikerer frivillige oppsigelser i Telenor og en maskinlæringsmodell som predikerer «dropouts» i e-læringsprogrammer. I tillegg har jeg jobbet med en tekstanalysemodell som automatisk kategoriserer fritekst-tilbakemeldinger fra kunder på tema og sentiment.

Dine andre favoritteksempler på din type teknologi internasjonalt og nasjonalt?

Recruitment robot hos Amazon, recruitment analytics hos Reitangruppen og retention analyics hos Experian.

Hva tror du er relevant kunnskap for fremtiden?

Måter vi kan benytte maskinlæring til å forbedre arbeidsbetingelser og prestasjoner i organisasjoner uten å gå på akkord med ansattes personvern.

Hva gjør vi unikt godt i Norge av dette?

Reitangruppen og Elkjøp har interessante eksempler på bruk av analytics og maskinlæring i rekruttering.

Et favoritt fremtidssitat?

Livet kan bare forstås baklengs, men det må leves forlengs.

Viktigste poeng fra vår samtale?

AI og maskinlæring har et stort potensiale for å transformere arbeidsplasser på en positiv måte, men for at vi skal oppnå disse gode effektene må innføring av denne teknologien gjøres i dialog mellom utviklere, ansatte og ledere.

Hva er det viktigste dere gjør på jobben?

Vi jobber for langsiktig utvikling av kunnskap som kan hjelpe Telenor til å gjøre gode veivalg for fremtiden.

Hva fokuserer du på innen teknologi?

Maskinlæring, statistikk og text mining anvendt på data om ansatte og kunder.

Hvorfor er det spennende?

Fordi det gir store muligheter til å utvikle kunnskap som ledere kan benytte direkte i å forbedre arbeidsbetingelser for ansatte i selskapet, samt styrke prestasjoner.

Hva synes du er de mest interessante kontroverser?

Etiske implikasjoner av økt bruk av AI i samfunnet. Vi kan se dette tydelig i prosjekter hvor vi bruker maskinlæring på ansattdata.

Dine egne relevante prosjekter siste året?

Utviklet en maskinlæringsmodell som predikerer frivillige oppsigelser i Telenor og en maskinlæringsmodell som predikerer «dropouts» i e-læringsprogrammer. I tillegg har jeg jobbet med en tekstanalysemodell som automatisk kategoriserer fritekst-tilbakemeldinger fra kunder på tema og sentiment.

Dine andre favoritteksempler på din type teknologi internasjonalt og nasjonalt?

Recruitment robot hos Amazon, recruitment analytics hos Reitangruppen og retention analyics hos Experian.

Hva tror du er relevant kunnskap for fremtiden?

Måter vi kan benytte maskinlæring til å forbedre arbeidsbetingelser og prestasjoner i organisasjoner uten å gå på akkord med ansattes personvern.

Hva gjør vi unikt godt i Norge av dette?

Reitangruppen og Elkjøp har interessante eksempler på bruk av analytics og maskinlæring i rekruttering.

Et favoritt fremtidssitat?

Livet kan bare forstås baklengs, men det må leves forlengs.

Viktigste poeng fra vår samtale?

AI og maskinlæring har et stort potensiale for å transformere arbeidsplasser på en positiv måte, men for at vi skal oppnå disse gode effektene må innføring av denne teknologien gjøres i dialog mellom utviklere, ansatte og ledere.

Vis mer
Tema: Digital strategi og nye forretningsmodeller
Organisasjon: Telenor
Perspektiv: Storbedrift
Dato: 190619
Sted: OSLO
Vert: SS

Dette er hva du vil lære:


Innovation
Diversity
AI
Machine learning models

Litteratur:Competing on Talent AnalyticsRichard Rosenow’s blog on LinkedIn

Del denne Casen

Din neste LØRNing

Din neste LØRNing

Din neste LØRNing

Flere caser i samme tema

#C1118
Digital strategi og nye forretningsmodeller

Karl Munthe-Kaas

Styreleder

Oda

Anne-Lise Fredriksen

Utviklingsleder

NKI

#C0294
Digital strategi og nye forretningsmodeller

Ole Gabrielsen

Direktør for teknologi og endring

Sarpsborg Kommune

#C0309
Digital strategi og nye forretningsmodeller

Helge J. Bjorland

CEO

Globus.ai

Lytt #C0421

Tekst for Case #C0421

SS: Hello and welcome to Lorn. My name is Sylvia Seres and this is a Lørn podcast with Telenor. Our topic today is artificial intelligence and its applications in H.R. human resources. And my guest is Jarle Hildrum V.P. of innovation and organization at Telenor research. Welcome.

JH: Thank you. Thank you for having me.

SS: It's a real pleasure having you here. You have a super exciting background for talking about what we are going to focus on today which is going to be both how you set up innovation in a very international company, with a past that's so glorious that it doesn't even know itself how glorious it is. We Norwegians are not very good at celebrating our technical heroes and we need to do more of that. And then the other thing we're going to talk about is how you analyze people's skills, behavior and needs and help them optimize the way that they use their competencies through A.I. at their workplace. And before we do that, I'm hoping that you can tell us a little bit about who you are and what drives you.

JH: Okay. Thank you. I'm a social scientist so I've been working with research on innovation technology and society basically topics around that

SS: You know a techie like me, we are super arrogant so I don't even know exactly, I know, but for our listeners... Let's say what is a social scientist? You know I think of a designer as a person who always likes to play with colored pencils and post-it pads social scientists. What do they know?

JH: Well social scientists would know how individuals interact in groups in order to achieve certain objectives. How groups of people end up in conflicts, how they would... It could end up in situations where they can be productive together and create better societies. Social scientists would understand how elections work. For instance, I have political science as a background, or how decisions are made in political institutions or in organizations. Social scientists can know how organizations work, how you can plan or collaborate within companies that work well, how culture evolves.

SS: So how do we get people to do what they need to do. So, where psychologists are very good at looking at the individual.

JH: Yes.

SS: You look at groups’ relations, interactions at the bigger scale.

JH: I would say so because I think that psychology would be a bit of.... It's a clinical discipline. It could also be a social science. But normally when you talk about social science you talk about sociology; you talk about political science and you talk about social anthropology. Those three are the typical areas that are pointed to.

SS: So, these are things that we believe are not automatable.

JH: I believe that some of the insights that come from this area could be hard to program

SS: In algorithm.

JH: Yes, because it has to do with social interaction, it has to do with trust. It has to do with empathy and things that, at least at this point, are difficult to decipher and program.

SS: So, we'll go back to your level. Just going to put one of my big kinds of theories of life and love and peace in everything on the table. The older I get the more I believe that the reason why we love people around us, why we love places, why we love cultures are their imperfections. As the opposite to the perfect machine. And it's about being in a way fond of these imperfections in their very human form and then complimenting them in the right way with the right tools for efficiency in organizations is the way forward I believe. So, we'll get back to that.

JH: Yes yes.

SS: So Jarle, you are a social scientist.

JH: Yep.

SS: And what did you do before Telenor?

JH: Well so I worked at the University of Oslo, Stanford University for a little while. Commissioned research institutes where I basically worked on trying to understand how companies can plan for innovation within innovation collaboration among people with very different skills and competencies who sit in different places, who are different culturally, ethnically and to be able to provide conditions for them to basically succeed in creating new things and new ways of work that benefits the organization.

SS: So, then I think I understand I'll have to go into the pain and joy of diversity with you the perfect man. But why Telenor?

JH: Well from... Since I started studying, and this is back in the early 90s, I have been watching Telenor from the side. When I started studying Telenor was slightly overstaffed by the state bureaucracy and I was watching this company as it transformed into a super successful multinational corporation that delivered not only voice telephony, but mobile telephony, Internet connectivity, platforms where Internet of things, financial services, online classifieds. A lot of different services to a lot of people in different countries

SS: Almost unprecedented international growth as well. Including a Norwegian start point.

JH: Yes. It's completely unique. I think in that sense both for the growth and also for what I see as a strong ability to experiment and to really venture into new ground. So, I see it as a company that is brave in that sense. I'm proud to work in the company. I always wanted to get the job there and then I had this knack for research and I wanted to do a PhD in these things so I did that first. But for me it was a goal to get a job in the company because I really think that we have a heritage of strong innovation and a strong ability to change. I think it's a fantastic company in that sense.

SS: I am a big fan and one of the things that amaze me is that Norway has certain technical innovation hothouses. We don't celebrate them nearly enough but some of the environments such as FFI, Kjeller, Norsk Regnesentral, Sintef, Kongsberg, your people here at Telenor. These people have been moving around and they are really world leading in radio technology, sensor technology, connectivity internet of things, materials technology. There's so many things that we at some very basic nerdy level do exceptionally well because we know how to process technology in this country. And now we've translated it into the digital world. And this has happened and we've seen the growth of Telenor. But we haven't kind of really admitted to us as a society. What's the basis for this kind of incredible growth and development? And I really believe that this needs to be spread as a story of our current welfare state even you know. This is what's built the prosperity that we base our future on and so you have a company that has a history of very brave leaders. I think even starting with Tormod.

JH: Yes

SS: I think I did a great job with basically rolling things out internationally. He saw the risk and he still did it. And I think we underestimate how much courage it takes from a leader to do that.

And then with Fredrik and now with Sigve, people who I think venture into unknown territory in a very exciting way.

JH: Mm hmm.

SS: So how does one translate this into the whole organization?

JH: I think that you know that the rest of the story about my entry into Telenor is that sometimes I'm presented with people who do not know that history and that's understandable because you know this is local in Norway. And the Norwegian part of the company is quite small. You know in terms of the whole organization. And I often hear that we are slow and we're really stagnant and we stop. And I'm amazed when I hear it and I tried to explain that look you have to look at the past. You'll have to look at where we come from. And you really have to see that. This is what we've done today is based on competence building that goes back 40 years because you had this strength for instance in radio planning that goes back to Telenor’s research efforts back from the 60s actually. The televerkets formed several hundred eight hundred at the maximum employees who were working systematically internationally on this.

SS: They were the central core. The standards that were set for this kind of mobile technology.

JH: Exactly. So, for the GSM technology they were central to defining those standards and really were leading. And even though we don't have that scale of a research unit in the company, we still have the competencies still there and we are still living off that. Not only that, I think the concrete kind of technical ability to do these things but also just kind of have the focus and the understanding that this is an area that we know and that we can evolve further. So, I think that this is to spread this into the whole organization. We basically have to do what we're doing now is that we have to work with really good storytelling. The company is expanding very fast now, probably it's going to be almost twice as big and we need to, I think, to try to build the image of the company around this story. I think that's very important.

SS: What are your main pieces of advice for spreading innovation throughout the company making it everybody's concern if not business and across cultures? What are your main kinds of goals in terms of getting Telenor to understand its full innovation capacity?

JH: Well I think that first of all I think it's important for people to understand that they are actually doing innovation without knowing it because very often innovation is associated with very fancy new technology whereas innovation in fact is improvement of processes. How we handle logistics, how we get our products and services out to the customers in an efficient way so they can appreciate them. It is about…

SS: Through all of our parts?

JH: Through all of our parts everything that we do. And it's also about updating our services and products slightly. But you know people out in the business units of Telenor will tweak the commercialization campaigns, the products, the additional perks that you can get on top of your subscriptions. If you buy a subscription you can get additional things like kind of access to music libraries so you can get online classifieds linked in there where you're in the market but this is not something that you easily access. So, people in the business units they're actually working on these incremental small innovations all the time that in sum are extremely important for us in order to be able to serve all these markets. So, I think that first of all I think it's about just making people more kind of self-confident that they are doing innovation too. And then I think secondly, it's about enabling because we are in the industry that has an extremely strong cost pressure and the people have to do more with less time. And then it's a lot about trying to enable people and free up time so they can work in projects together to try to make improvements in different areas.

SS: I think just understanding what industry are in now and where the industry is going and getting that kind of lifelong learning aspect flexible learning into people's everyday life is something

that Telenor does really interestingly

JH: Yeah

SS: I think

JH: Yes

SS: Because I mean are you a mobile first company? Are you a digital first company? Are you an A.I. tools companies eventually that deal with our smart homes or smart cities or, you know,I think there are so many ways where a company like Telenor can be going?

JH: Yes

SS: it's important to understand this transformation while we're living it.

JH: Yes absolutely. I think that this is partly the way that we're going. At least six years I worked in the company. I think that you know it's partly about planning and deciding where you want to go. And partly it's also about experimentation because it's an industry and business that changes extremely fast. Technological development is super-fast. So, artificial intelligence has really just… It has progressed so rapidly during those years that when I was hired it was hard for us even though we were working on an AI then to see the potential of this today and to see the potential of AI combined with the Internet of Things for instance and 5G. That was very hard to see back six years ago. So, I also think about having the flexibility to actually take a turn when that's appropriate. And I think that's actually a strength of Telenor because you could say that sometimes Telenor is criticized for being inconsistent in the sense that you have one strategy and then you'd make a rapid pivot right. But then I think this is part of the execution power of Telenor. Just like if you look back in the past you know some of the successes that Telenor has had have been decisions that have been made almost on the fly when the opportunity arose. Say Grameen Phone for instance and then those types of expansions and then we may look back at them. It would look like they were kind of well-planned and recent but really they were about taking that opportunity when it came. And I think this is also what's happening now with our new strategy.

SS: I think refusing to pivot when the facts change because one has said something else earlier is simply a bad strategy.

JH: Yes.

SS: You know you have to, and if you're quick to adjust to the new facts then you have a huge strategic advantage.

JH: Yes but of course you have… That's the upside and then that the thing that you need to also work on with that which I think back to your former question about what would I do to try to enable the company to deliver on the innovation potential, it has to do with competence. Because I think that when you have this ability to pivot fast and also say that culture to pivot and execute you could lose some competence on the way, because when those changes are made very fast and abruptly they can also alienate employees who thought that look I thought my competence was supposed to be used for this purpose. Now there's a new purpose. Right.

SS: And can we talk about that for a moment. As I said I'm going to ask you a little bit about the pain and joy of diversity. And I was thinking first and foremost of the cultural diversity that Telenor has and I think actually works really well. I've spoken to Rouga among other things the Montenegrin CTO. We talked a lot about how different cultures collaborate to actually create some new magic here. But I would like us to talk for a moment about the subject, the professional diversity that is happening. I think that in the world that we're going into and now we are moving into the future where you know we're all going to be changing jobs all the time. Not every fifth year you know twice a year in the way the content of our work is going to be continually changing. And I think the center of our competence has to evolve.

JH: Yes.

SS: And getting people to understand that that's OK. I just need to keep having. That's what everybody needs to do. It's not an easy thing to convince people that believe that you know well I'm the specialist and that's what I am. But do you believe that I'm right? Do you believe it? And I think getting people like you who do social science to start working with A.I. that I thought was my territory I think is the right way going forward. So how do you make sure that people actually are comfortable or encouraged or enthusiastic about learning stuff that they believe they're going to have to learn and develop with it?

JH: I think that it's about trying to keep a core of competence there.

So if you use the example of me working with machine learning and analytics, I think that I can keep a core of my identity as a social scientist working with that topic, because it's for me it's almost like a liberal art in the sense that when you create models that are supposed to be used and communicated through the organization they have to have some sort of an aesthetic around them

SS: A story.

JH: Yeah, they have to have a story. And this is part of how I was trained. So, I think I look at myself not in any way as a technologist but I know certain techniques and I can apply the background or the height I have with those techniques. And I think that's the same thing when you want to have kind of a traditional marketer moving to say a very advanced digital marketing you say look your former competence is extremely important. We are going to help you out with a couple of new skills on that and it's going to make you even more able to do your job. It's going to be exciting for you. And I think that's always important to honor the competence that people have and I think also looking back at Telenor and trying to look at the successes of innovation that comes in addition to the kind of standard products. If you look at for instance financial services that are not well-known that we do but we actually do it quite extensively in Asia. This is something that kind of goes back 25 years. So, it's kind of a base competence that we have in the company but it has been applied in very different ways. So, it's originally you know, it was about marketing and marketing people, economists and some technologists who have kind of morphed this into a skill and a competence that has been applied many places but it's not the same as it was.

SS: So I think my mental image when I listen to you talk about this is a T shaped individual with you know the vertical leg being your core competence that gives you your professional identity and then the top horizontal bar being there all these new tools and you are continually adding to your basically toolbox.

JH: Yes.

SS: And A.I. or machine learning is just one of those tools and you're playing with how can it help you understand your core competence questions in a better or a new way.

JH: Yes.

SS: And maybe there will be other vertical legs you know. Maybe you decide that some sort of I don't know ultimate automation of the workplace is another core competence that you have. But again I think it's extremely important to have some of those deep legs.

JH: That's absolutely true.

SS: That makes you proud of what you do.

JH: Yes and also to be relevant I think as I say as a collaboration partner we have we're talking about collaborating across disciplines. I think that really what I think is that any person who has the passion and competence in an area can recognize that you know the same passion and competence in another area and then it's much easier to relate and to interact and collaborate. If you have people who are generally just generalists who are jacks of all trade and not really have a base it's harder to do that because it becomes more of a kind of rhetorical interaction rather than you that you actually kind of going deeper into there. Yes. And more able to connect.

SS: Experts recognizable to bingo very quickly.

JH: Yes.

SS: So now what is the social sciences guy like you want to do with technology like machine learning applied to the workplace what’s the project?

JH: Well the project is to I think to be modest it is to try and transform the workplace. I strongly believe that machine learning has a great potential to improve the way that we are helping employees to do their jobs better. This can happen in several areas. I already mentioned automation that you know you can. Certain tasks that we are spending a lot of time on the kind of the people HR side like reading resumes like creating reports like analyzing data also can be auto parts to a great extent this is already being done by many companies. Google is great. That's the ultimate icing in H.R. work so this is one area. We have not started or we’re working on robot cessation in some aspects but for the machine learning part. They are in Telenor right now. Three examples that I'm working on so I'm working on predicting voluntary turnover. So that's we're actually building a machine learning model that can identify who is going to leave the company voluntarily

SS: based on what you can do.

JH: Well you look at past data on the people who choose to leave voluntarily and then so you could have us. You have to have a certain size.

SS: but do you look at the salary or do they start coming late to work? What's a relevant spot?

JH: Those are relevant. So you could say that the companies that have come the furthest in this area work with 250 different variables that are among these are the ones you're talking about. Yes. So if you looked at Oracle and they have been exploring with data on LinkedIn updates whether the employee has been updating his or her LinkedIn account whether you know he or she

has been sick.

SS: about their CVs.

JH: Yes all of these things are right. In Telenor we are very concerned with privacy and it's kind of like when you work with machine learning in the area of HR. It's important for us to come to work with the ethics of that.

SS: Yeah.

JH: So if you take this example, what we're looking at is really data that's available to anyone. So when I build this model I have data on the employees where the employee works his or her tenure his or her age. I look at the number of managers that have been replaced in his or her team. I look at whether there has been restructurings in the team in the past. I look at and all this is theoretically kind of founded and I look at e-learning whether the person has completed or invested in him or herself.

SS: Is that a positive or negative signal?

JH: It's actually a positive.

SS: They’re staying

JH: they're staying. So yes. And also stock ownership. Whether you own time or stock again it's constantly positive in the sense that people stay for that reason.

SS: And if you discover that you're in danger of losing people you don't want to lose.

JH: Yes.

SS: What can you do?

JH: Well then you can get two things out of a machine learning model like that. The first thing is that you get the prediction score for every employee based on the model. So you can segment teams and you could say so the risk is actually higher there. Let's do something. The second thing is that you get the ranking of the predictors. So the factors that…

SS: ...you understand what really matters to people.

JH: Yes. So the factors that are the most important the model will tell you are those for that specific segment. So say if that is e-learning in a certain area that you could actually provide more competence and training and more time for that. Well that you can do you can plan in advance and then you can you get the better outcome say for instance in Telenor Norway we have too few women in tech. So if we can improve the accuracy of this model we would be able to see in advance you know certain things that we could do in order to kind of keep the diversity.

SS: Yeah. So help them stay helping us.

JH: help them stay basically.

SS: Come and help them stay.

JH: Exactly so. So that's the positive. And then there are the ethical concerns on the other side that we are also working with that are less positive.

SS: Yeah.

JH: So you could say that if I'm getting a score saying you know that 89 percent likelihood that I will leave within one year and the wrong manager gets that information. Well he or she could think or say that look I'm not going to invite these guys to a meeting about sensitive stuff. Why invest in training?

SS: I'd be very careful who gets to see what you know.

JH: or you have to regulate how that information is used. So we are interacting a lot with the unions. So the unions are really cheering for this project because they think that it's a tool that we can use to kind of keep the diversity in the company.

SS: Yeah.

JH: And it will kind of be a more objective kind of consideration if you know what's going to happen up there. But they are also setting up you know or want to set up together with management. So it's a two part collaboration. The rules of the game. So how is this information going to be used?

SS: Can I ask you about something else? So you have data about what people in a way feel about their work but what do you have. Can you somehow figure out what they know and how you can teach them more? You know I'm trying to think about whether all of us are going to need to go back to school in a virtual sense. None of us have time to go back to the university or you know an MBA every year. But you know our workplaces and our road to work, journey to work in our home is going to be this kind of a learning environment that we have. But it'd be nice to get some help on what I should learn. Where do I start? How do you make a structure for me, my virtual syllabus?

JH: Yes.

SS: And I believe that if you look at the content on LinkedIn combine it with I don't know stack overflow in the number of thumbs up a programmer has got and so on. You're beginning to get a good sense of who's a real expert that could teach others what their next opportunities for learning are. Is that something we could do?

JH: Yes I think that to scrape those kinds of social media accounts is somewhat complicated. But we actually have some great resources internally and so we have a workplace which is the Facebook kind of Internet version which is a place where a lot of the experts are really showcasing their competence. They have communities where people are gathering around them. And it's possible by using the information that's on that site. Given that we have the permission and we're allowed to do it to build up a map of you know the experts in different areas. That's one. The other one is that we are running something called the 40 hour challenge now which is the CEO saying that everybody can take 40 hours training in the area where they need to kind of a scalar themselves. And we had that for last year and this year. And it's an amazing uptake.

SS: People participate.

JH: People participate. So the one thing you keep out of it is that you can just by looking at what people are searching and what they're trying to kind of train on you can get a certain map of the needs. That's one thing. The other thing is that you can also learn about how you can enable people because that's the second machine learning project that I'm on. We see that in different places you have higher rates of dropouts. So people would kind of go into the courses and they would work on it and then they'd drop out for some reason

SS: Why did they drop out?

JH: Exactly. So why do they drop out. Well if you link that data on kind of the usage patterns you know and kind of how they are working how often they log in and out when they're training and then you link that also to the situation of the employee how many projects are they on. You know what type of role they have. What time. What typical time of the week or do you know of how close to the two quarterly reports are they. When they drop out these things can help us. You know first of all to predict who's going to drop out when. And secondly you know to see that for different segments again you know what seems to be driving the dropouts and how can we address those things.

SS: And we need to get the land podcast on these platforms so they can see how many podcasts there are on the way to work.

JH: That's what we should do. Yes

SS: So in a way you're making a social model of the organization and you're optimizing it for people to stay in the right kind of motivation and right skill level. And there are other companies that are doing this. Amazon has the recruitment robot right and Groupon has the recruitment analytics and Experian has the retention analytics. What's the best thing you can say about each of these? Why are they interesting?

JH: Well they're good then bad things. So if you look at Amazon they actually scrapped the recruitment robot because it discriminated against women. So it was trained on 20 years of Amazon data and it kind of reproduced some of their prejudices and some of the biases that existed in the company. So it's a prime example of the consequences of A.I. actually

SS: because the robot was basically analyzing past successful hires and for some reason because one was hiring more often men it assumed that that was a predictor for future success or something like that.

JH: Or rather that men if the measure of success is promotions and men are promoted more often than women. And the model learns that you know then it will promote that it will choose the

men.

SS: Yeah.

JH: So…

SS: ...self-reinforcing prophecy in a way.

JH: Exactly. So I think the company I think is the most inspiring in terms of using machine learning and advanced analytics on people is Google. So they've been working on this since the early 2000s. They have an organization called rework where they kind of makes kind of exposing all of the things they do and they really have a strong focus on hiring because it's extremely important for them to hire people who will be able to succeed in the company but they're not automatically seeing it. It's a combination of kind of people who can evaluate what the model is doing. And of course models. And the second thing they do is that they work on optimizing teams. So not only being the individual but what combination of people in terms of kind of the chemistry and you know social skills and interests and competencies do you need in order to have a team that's successful in innovation. So this is something that Google has been working on for a long time. And I think that again going back to innovation you know and having a component that's so big and diverse maybe you know for the future this is something that we could do. You know how do we set up a team that's you know the optimal team. How can we learn how these

SS: what is the anatomy of an optimal team.

JH: exactly or at least improve or ability to do that.

SS: I recently read that NHS has done a similar. So you have call centers and you analyze some of the feedback that people write to your call centers. And try to improve the quality of service and do that. And I think that's a super interesting Ali as well. And I read recently that NHS has done some A.I. based analysis of the feedback and complaints that people in England have stated about their healthcare experiences. And you know it's so interesting surprising in a way points from this. So it was by far not the you know the professional level of the doctors and the quality of the medical services they got. It was always whether they were greeted in a polite way or whether you know people were too late whether they had to wait too long compared to the schedule the time or you know. So it's a really interesting way to learn what people really care about when they consume your service.

JH: Yes.

SS: So we are overtime of course a lot what A.I. improvement of Silvia would be to get her to understand the clock. So you were working with A.I. to improve the work conditions and the performances of your parts of organization and you do this internationally or mainly in Norway.

JH: Well it's international in the sense that we collaborate with some business units you know Grameen Phone is actually doing lots of interesting things on this. So we work with them and we also work with the central H.R. organization that has responsibility for all of the units. So but I'm located in Norway.

SS: OK. So you are kind of a cyber-Hunter for talent and you recommend the guy called Thomas Dave Put as a good read. Why?

JH: It's because he's really great that he has a broad span of writings on A.I. and organization management. So he's one of the people who can write intelligent kind of intelligently about the use of A.I. in people and human resources. And he's a Harvard professor and there is a real it's an example of non-technical guy who has a very good grasp of both the technology and kind of that.

SS: The problem...the organizational kind of context of the technology.

JH: Yeah. There is a guy from Facebook as well. Richard Rosenald. Yes. So..

SS: ...he has a blog?

JH: Richard Rosenald he is the head of Facebook’s people analytics team and used to be at least, and he is publishing. Facebook is not as open about what they do, they don't publicize to the same degree as Google does on their work on A.I. machine learning on the people's side. But they do lots of interesting things. And this guy he has created. He has written a primer on A.I. and analytics in H.R. and it's published on his selected site. So it's kind of for the treasure for people are interested in that.

SS: Yeah we'll take that. Would you like to leave us a quote as a parting gift to our listeners?

JH: Huh that's hard.

SS: That's hard for a social sciences.

JH: Yeah. Well we've been talking a lot about predictions but I fundamentally believe that you need to look at the past a lot in order to get your insights. So I would leave you with a certain creepy guy who said that you know life can really only be understood backwards but it has to be lived. Say four links...

SS: ...forwards.

JH: Exactly. It has to be lived forwards. So that's the in the sense that the constraint that we're all always facing even if we have A.I. we really have to look at our past and understand that in order to proceed in a good way.

SS: Yeah. I love that word. If people are to remember one thing from our conversation we talked about many things. What do you think is the most important point we have.

JH: I think that the most important point is that we need to have discussion and the dialogue on artificial intelligence that looks at both the vast opportunities that are there and the challenges for privacy of people you know the well-being of our employees all of the potential difficulties or problems that we can fight to get into it. It's a balance it's both. And we need to talk about all things at the same time and then we need different people not only technologists not only social scientists but all.

SS: I have a quote that I usually kind of bother my children with searing a motto maybe care there deliver.

JH: Yes

SS: And I think we talked about care through the unnecessary pain and joy of diversity. I think we talked about there through this. You know the precious courage of innovation that you talked about and we talked about delivering where this needs to be scaled in a very operationally concrete way.

JH: Yes.

SS: And it's really impressive what you guys are doing.

JH: Thank you.

SS: Yeah. Jarle Hildrum V.P. of innovation and organization at Telenor. Thank you so much for coming here to Lorn and inspiring us to learn more about A.I. used for HR.

JH: Thank you Sylvia.

SS: Thank you for listening

Quiz for Case #C0421

Du må være Medlem for å dokumentere din læring med å ta quiz 

Allerede Medlem? Logg inn her:

0

C0421 AI AI and transformation of the workplace - med Jarle Hildrum

1 / 3

How did Jarle describe social scientist?

2 / 3

Where was Jarle working before moving to leaving to where he is working recently, at Telenor research? He was working at

3 / 3

Why did Amazon actually scrap the recruitment robot? It was scrapped because

Your score is

The average score is 0%

Du må være Medlem for å kunne skrive svar på refleksjonsspørsmål

Allerede Medlem? Logg inn her:

Vi bruker Cookies for å forbedre brukeropplevelsen av sidene. Les mer om personvern & cookies her