Skip to main content

tv   Washington Journal Chelsea Barabas Discusses Big Data in the Workplace  CSPAN  June 1, 2017 5:09pm-5:35pm EDT

5:09 pm
providing critically important information about our national security and our foreign policy interests and that information is not being trusted and acted upon by the primary consumer, the president and there's reason to be concerned about whether or not policies are based in fact and based on the information that we spent a lot of time and effort obtaining in order to ensure that we are making the best possible policies. that's a concern as well. >> matt olson, thank you for joining us. [applause] >> thank you everybody. >> i want to make a couple quick in athens. will take a quick break and start again at 2:30. our overflow room has a tv that works all the time except for today for some reason. given that there's a lot of seats available if some people want to come in and switch in better that would be appreciated. will start again at 2:30.
5:10 pm
>> next, conversation on the role of big data and technology in the workplace. from washington journal this is 20 minutes. >> welcome back. this week spotlight on magazine chelsea barabas is here to talk about her magazine appearing article appearing about the workers and technology changing. thanks for being with us. >> hi, glad to be here. >> when people think of technology transfer me the workplace they often think of automation but you write this big data which is guiding powershift to critics point out. >> sure, as you mentioned, a lot of the conversation around artificial intelligence and its impact on the workforce has really been around at the rolein it will play in automating all kinds of jobs in the workforce. a much less kind of talked about aspect of this is the thinking about the role that artificial intelligence plays will play in
5:11 pm
decisions about opportunities that are available to you in the workforce. so, a lot of my research has looked at the growing amount of data being used during hiring and recruitment for jobs, as well as evaluating people's c performance in the workplace to informed decision about promotion and things like that. some people think, in many ways, this could be an opportunity fo us to address long-standing biases that we've seen in the workforce around, you know, assuming groups of women, cult people of color, women, and denied opportunities because of these biases and hiringwi requirements, promotions, things like that. there's also a growing conversation around can data really be used to make more objective decisions at identify talents that may be overlooked otherwise and create more fair
5:12 pm
democratic labor market. >> you write in the article that both skilled and unskilled workers are increasingly adept at controlled not by human supervisors but by opaque algorithms. that with the future of work that will play out between machine and man and they will be between the workers who generate data in the people who are well? positioned to leverage it. if you could speak more specifically about how that data would be used in the workplace. how might it be used in the specific industry? >> for example, in like the high tech industry there's been a growing number of companies who have gone into the business of collecting information about the workforce and what individuals might have specific types of software development skills. for example, this is an area that is really right for big
5:13 pm
data analysis of various kinds because a lot of the kinds of works software developers do leave a very large digital footprint. if you're a coder, you'll likely to participate in forms online where people are discussing and answering questions about certain challenges they are encountering in their work and you have public open code depositories where people make available prior projects they worked on and things like this. what we have seen -- especially in last three years, a growing number of companies basically are in the business of gathering that data, taking it from the internet and compiling it in large data repositories that try to encompass as many of the potential workers in the high-tech field as possible. then what happened with that information is companies create dashboards and recommendations systems that recruiters and companies who are really trying to efficiently identify new talent can use to identifyfy skilled workers of a certain kind or people with certain
5:14 pm
kinds of characteristics that they are looking for that can range from highly technical thing, such as, been able to code in a specific language too much work but the concept like your ability to work wellin teams and your level of cooperativeness and things like that. what we are really seen happen is that these kinds of -- o because there's so many kinds of data generated that can be can fuel this type of recommendation system it's changing the way recruiters, even go about their work flow and thinking about identifying -- especially the employees at the top of the funnel. these systems are being used to service candidate that recruiters can go out, reach out to and tell about new job opportunities. this is very different from what we are familiar with and that most of us think of jobs in the way that we seek the job opportunity as being something that a company will post a javascript in and are able to
5:15 pm
apply. what were seen with these recommendations systems is that now recruiters and companies can focus more targeted attention on specific candidates. but they want to reach out to specifically. as opposed to relying on a public job posting and soliciting applications from their. >> chelsea barabas is and weree taking calls about big data., call (202)748-8000 and if you're not an intimate time zone call (202)748-8001.ar chelsea, what types of workers are at risk in as employers leverage big data. >> again, depending on what were focusing on automation or these different type of evaluation tools, automation is happening across the gamut. right now what we are seeing in terms of these automated decision-making and recommendation tools, we are seeing concentrations in a few specific industries. my work is primarily focused on the high-tech sector which is
5:16 pm
one in which, you know, the industry is growing and feeling the pains of feeling like the demands for skilled labor was outstripping the supply that they could find. so, with industries like that where there are chronic labor shortages and forcing the pipeline of skilled work massively expanding these types of systems are being used. another place is has been used for quite a long time are also low skilled work district for example, call centers. call centers most people who work in call centers the calls are recorded and constantly monitored and they are held accountable to hourly metrics for how may people they called and how many successful calls they've had and then their pay and promotion opportunities are based on those type of metrics that are fueled by digital data. i think a third an interesting area were starting to see
5:17 pm
increasingly in areas of work where were seen digital platforms play a larger and larger role in mediating interactions between customerst, and workers. interestingly one might be the transportation sector with the rise of companies like huber and left, taxi drivers now are generating far more data about their professional activity. everything from their professionalism on the job, the work they do on a daily basis and this is really valuable o information that, you know the driving platforms can be used as a basis for promotion and in such citation for certain kinds of behavior but also valuable data that could potentially be used for the worker to access different types of opportunities such as, for example, the purchase of a new vehicle. if they want to be able to finance the purchase of a large asset, hypothetically this typei of professional data to be used to give insights into their cash
5:18 pm
flow, their productivity and their level of professionalism which can fuel into new opportunities, new doors opening up for them. >> another point that struck me that you mentioned in your piece you write that big data comesie with all the. [inaudible] of creators. for example, one team a data analyst i interviewed said they listed lord of the rings of one of their favorite books on facebook were more likely to bee hired as successful cto's in tech companies. this insight helped detect which candidates were good for high-ranking jobs but the type of people who fit this ideal cto will also reflect the democratic who can hold the positions in the tech industry which means these metrics tend to reveal who is already in power rather than finding new people who are best equipped for the job. is there a way to mitigate that? >> i think were right at the beginning of theseig conversations. so, the early hope of big data,a
5:19 pm
a lot of the conversation was focused on this hope that we could achieve more objective metrics for basing decisions on. so, could we eliminate some of these sticky biases that have plagued our society forever. what we found no, actually, the data that we generate reflects the world in what we live in and this world is -- we have vastly different expenses according to our background and certain key demographic features. if we don't take that into consideration what we run the risk of doing is basically replicating the existing biases that we see already in the world today. the tech sector is a great example of this which is where that example is drawn from the tech sector over the last few years has taken a lot of heat for really not being very diverse, for being largelyng white, largely male industry at this point. especially on the engineering side of things.
5:20 pm
as a result, if you were to develop a recommendation systemt looking for key attributes and characteristics that might indicate who is going to be a good cto, those character sixwed mighlikely be skewed for the pee who are already in that position and most of them are white men, white men are more likely to come back. this is tricky because the way we sometimes look at this is thinking, okay, what if we eliminate race, gender especially from the equation and we can somehow take that out as a biasing factor. the trip with that is there is a lot of ways that race, gender, class, can get double encoded into variables that you might not think would have anything to do with those things. for example, zip code is a really important indicator
5:21 pm
oftentimes of race. our cities tend to be fairly segregated along racial lines. if you include something like a zip code in your algorithm for predicting who might be a good feel for a company you're likely encoding some bias that's going to be connected to race as well. >> let's take a few calls. john, from niles, illinois. good morning. >> caller: thanks for c-span again. chelsea, i'm a college administrator working in the career field with journalism students. i'm really interested to see what you found regarded linkedin for specific social media that we use with students for new grads and recent grads and in addition to mid-level people.th what do you think about linkedin and how that's been used for. >> guest: linkedin has been a massive platform for recruiters and sourcing talents. again, what's really interesting is figuring out what information is made available on these sites and how does that shape the decisions being made. another really interesting to
5:22 pm
look at this is university college admissions. so, the way that applications of structures, the digital dashboard that selections committees use in order to interview candidates, rank them and give them various scores are really important in shaping who gets in and who doesn't. that's similar to linkedin. it's the most dominant kind of professional network available and a lot of recruiters use that to find new talent. i haven't done a lot of work specifically focusing on that so i can't say anything to specific on the information other than i think it's probably very influential in important area these days. >> host: chelsea, you mentioned the big data algorithms have the power to sort people in piles of worthy and unworthy. that begs the question how regulated are the groups that collect and sort the data.
5:23 pm
>> guest: right now, very not very regulated at all. it's kind of the president for this kind of work and at the sign of scoring is comes from the fair credit reporting act in trying to regulate credit reporting agencies. the credit reporting agencies is a very specific kind of classification that really only is irrelevant or is only connected to the credit reporting agencies we've traditionally understood them. what we've seen is very quickly expanding group of third-party companies who do credit reporting like activity from a broader set of things, consumer screen activity. right now those really aren't regulated under any specific regulation or law. i think we can draw from the fair credit reporting act as a r useful president to think about
5:24 pm
how we might regulate these things in the future and why that might be important. so, for example, when the credit reporting started in the mid- 20th century some of the challenges particularly for consumers was their information was being gathered and sold to companies, potential employers and then those scores were being used as the basis for making decisions but there is no obligation from either the credit reporting agency or the person who is making the decision whether it be a potential employer, potential creditor to inform you that your credit had been used to make a decision. if you are denied an opportunity based on your credit score you really had no way of knowing if that's the case. if there was inaccurate information, you had no way of checking that, no way of contesting the decision that was made or fixing the record. the fair credit reporting act was created to create a due process framework to protect us as consumers and give us the opportunity to be notified when decisions were made on our behalf based on information
5:25 pm
collected. to to get access to that information, access to our credit scores and review the information is compiled about us. three to test either the accuracy of the information that was included in the report or to contest the basis of the type of decision that was made on your behalf. using that information. i think we can borrow thates framework and think about that similarly in this era of big data and algorithm. thinking about notification, access to your personal information that's been used and then a process for contesting those decisions and checking for biases and in discriminatory practices that might be built on top of them. >> host: john, from massachusetts. good morning. >> caller: good morning. how are you doing. >> host: what's on your mind? >> caller: i kind of have was at a software company that did some
5:26 pm
early on large database things and actually did some consulting for one of the big, big recruiting database systems prior to linkedin things. the whole notion that big data is doing anything that's particularly good and particularly systems that were built to parse and look at systems had been centralized into a couple of large companies and the whole idea of personal computers and individual liberty and all these kinds of things
5:27 pm
were thrown out the window. we designed a whole lot of systems to be able to make information available to people on an individual level and to get away from mainframes controlling and centralized control of the machine. unfortunately, people get their data to their systems and they put themselves in a thing where people make sort of group ideas about things and whether you talk about software engineers with specific talents on linkedin and hiring people on
5:28 pm
job performance instead of people who know the engineeringh is one thing that you have to understand that -- i'll give you an example linkedin recently wad brought out, excuse me, was bought out by microsoft where where essentially stock prices of linkedin was 43%, i'm just taking a number but that's probably about right and he was billions of dollars that was printed to get linkedin over to microsoft to control a lot ofro the technology. this is kind of how it works. what happened when the initial ipos for linkedin they limited it out to very few people and if
5:29 pm
you hire from linkedin you're only going to get certain kinds of corporate people and people that think a certain way. this is a typical and it's kind of really stops innovation andte if you look at where the information comes from the tech industry it comes from overseas engineers and things like this because they had to learn the systems from the ground up. >> host: chelsea, care to weigh in. >> guest: do i want to weigh in: there? sure, sure.aw i'm just trying to draw out a specific thing to respond to. perhaps you said a lot of things in their, you know, i'm not
5:30 pm
really sure what to say to that honestly. yeah, there was a lot there. >> host: let's go to al from lauren, ohio. >> caller: we definitely don't need marijuana smoking in the united states. that's for sure. i have two questions. first question on the magazine. when you got this spotlight in the magazine and they get your information do they sell it to other people? one for c-span, i have two? can they sell that information to other questions that have to do with magazines? second question is is c-span going to talk about the wiretapping on donald trump from previous president mr. obama because tomorrow will be a good day. thank you. >> host: i believe we have that conversation coming up tomorrow morning but your earlier point, chelsea can companies all the information they gather on people? >> guest: yes., in most cases, that is definitely true. they're very large secondary
5:31 pm
markets for the buying and selling of data. you know, either from third parties have collected it or services that you have used online and in many cases both services have the right to your data and can do whatever they want with it whether they use it for their internal purposes or sell it. particularly, if the company goes bankrupt they will oftentimes a very digital company, their data is their most important asset and they will sell that data to recover some of their cost or debt. >> host: let's move ahead to jack calling in from davenport, iowa. what's on your mind? >> caller: yes, you talked about sexism and stuff. in these videogames do people collect data and what the ratio is of men to women that playeded video games?
5:32 pm
>> guest: you know, i don't know a lot about that area but i do know there are some really interesting, there is some interesting work being done demt the changing demographics of the videogame players. particularly, around gender. i'm not hundred% on this but young girls are actually the largest growing demographic of the videogame players, i'm not sure where that data is because collected but because videogames are by their nature, digital, i'm sure there's a lot ofe is al information, a lot of information produced and it be interesting to figure out who harvest and collect that data. but i'm not familiar with that particular industry. >> host: chelsea barabas is our guest this morning purchase a host of digital currency and head of innovation for the currency initiative at mit media lab and she speaking about her peace and democracy journal. chelsea, thank you for your time this morning. >> guest: thank you.
5:33 pm
>> coming up as our former presidential candidate and secretary of state hillary clinton. she discusses her book at the publishing industry's annual convention bookexpo america in new york city. life coverage begins at 6:00 p.m. eastern on c-span two. on sunday, matt tybee will be our guest on in-depth. >> you grow up looking at thousands and thousands of faces until one day you see that face that you feel is put on earth just for you and that's instantly you fall in love in that moment. for me, trump was like that except it was the opposite. when i first saw him on the campaign trail i thought this is a person who is unique, horrible and amazing, terrible characteristics were put on earth specifically for me to appreciate or unappreciated or whatever the verb is. i had really been spending a lot
5:34 pm
of the last ten-12 years without knowing it repairin preparing fd trump to happen. >> he is a contributor to rolling stone magazine and is the author of several books including smelled like that elephant, dispatches from a rotting empire, the great derangement: a terrifying true story of war politics and religion, chris tobia, a story of the bankers, politicians and the most audacious paragraph in american history. his most recent book in insane clown of president, dispatches from the 2016 circuits. during our live three hour conversation will take your calls, tweets and facebook questions on mr. matt's career. watch in-depth with author and journalist met live from noon to 3:00 p.m. eastern on sunday. >> our next

20 Views

info Stream Only

Uploaded by TV Archive on