Latest / Paths of Curiosity in an AI World / Q3 - AI in Data, Behavior, and Human Systems with Aaron Klein, AB, MPA

Transcript

Q3 - AI in Data, Behavior, and Human Systems with Aaron Klein, AB, MPA · Paths of Curiosity in an AI World

• 0:04
  SPEAKER_02: Welcome to Passive Curiosity in an AI world where we explore the world's careers through the lens of AI.
• 0:13
  SPEAKER_02: For decades, banks discriminated against borrowers in person.
• 0:18
  SPEAKER_02: And algorithms inherited those same biases and made them invisible.
• 0:23
  SPEAKER_02: Aaron Klein says AI could finally fix that.
• 0:27
  SPEAKER_02: Or make it far worse.
• 0:29
  SPEAKER_02: Let's get curious.
• 0:36
  SPEAKER_02: Aaron Klein is the Miriam K.
• 0:38
  SPEAKER_02: Caroline Chair and Senior Fellow in Economic Studies at the Berkings Institute, where he focused on financial technology, regulations, and the future of payments.
• 0:49
  SPEAKER_02: Before Berkings, he served as Deputy Assistant Secretary for Economic Policy at the U.S.
• 0:55
  SPEAKER_02: Treasury Department, where he helped craft the Dloyd Frank Wall Street Reformation Act after the 2008 financial crisis.
• 1:05
  SPEAKER_02: Before that, he was the chief economist of the Senate Banking Committee.
• 1:09
  SPEAKER_02: He holds a BA from Dartmouth and an MPA in Economics and Public Finance from Princeton School of Public and International Affairs.
• 1:18
  SPEAKER_02: Mr.
• 1:19
  SPEAKER_02: Klein, thank you so much for joining us.
• 1:27
  SPEAKER_02: So most of our listeners are starting to financialize right now.
• 1:31
  SPEAKER_02: First credit card, first bank account, maybe their first loan.
• 1:35
  SPEAKER_02: An AI algorithm will likely make the first call on whether they get approved.
• 1:40
  SPEAKER_02: What's happening when that decision gets made?
• 1:44
  SPEAKER_03: So it's a very good question.
• 1:46
  SPEAKER_03: The name of the AI is called FICO, which stands for Fair Isaac Corporation.
• 1:52
  SPEAKER_03: Fair Isaac has been around for a very long time, and what it does is it takes a bunch of information mostly from things called credit reporting agencies.
• 2:03
  SPEAKER_03: These are three major companies, national companies, that collect information on types of transactions people have.
• 2:12
  SPEAKER_03: They then, this FICO algorithm in a big mysterious black box, puts it all together and spits out a number if it can.
• 2:22
  SPEAKER_03: That number is kind of made to look like an SAT score.
• 2:26
  SPEAKER_03: Not entirely.
• 2:27
  SPEAKER_03: 800, you can do a little bit better than that.
• 2:29
  SPEAKER_03: It's not perfection, but it's closer to 850 as a perfect, which is an odd number to have as a perfect, right?
• 2:37
  SPEAKER_03: How many tests do you have that are out of 800 something other than the SAT?
• 2:42
  SPEAKER_03: And like the SAT, it's not terribly accurate, but it's very important in your life.
• 2:47
  SPEAKER_03: So with this number, uh people decide whether to give you credit, and if so, at what rate.
• 2:54
  SPEAKER_03: Generally, the higher number is better.
• 2:57
  SPEAKER_03: Some of the big factors that go into this include how long you've been in the system, that is, how many years.
• 3:04
  SPEAKER_03: So that's not really good for younger people.
• 3:06
  SPEAKER_03: Kind of younger people start out at a bit of a disadvantage.
• 3:11
  SPEAKER_03: Other factors include how much credit you have, how much of that credit you use, uh hint, if you're using everything you got, maybe you're a little bit riskier.
• 3:22
  SPEAKER_03: Um so it's very good for younger people to start a credit history early, to use it responsibly, and to pay on time.
• 3:31
  SPEAKER_03: One other big factor is have you ever been late in a payment?
• 3:34
  SPEAKER_03: So you say, oh, oops, I forgot something.
• 3:37
  SPEAKER_03: I was a little bit late, I missed a deadline.
• 3:39
  SPEAKER_03: I know when you're you ask your teacher, oh, can I turn this in late?
• 3:42
  SPEAKER_03: And sometimes there's a money penalty on a credit card, but in this world, there can be a credit penalty.
• 3:47
  SPEAKER_03: A person who's constantly late, or even late a couple times, particularly when you're younger, you're gonna you're gonna hurt your credit history, and then things are gonna be more expensive and difficult.
• 3:58
  SPEAKER_03: Importantly, a large number of people, particularly young people, don't get a score because there's just not enough information.
• 4:05
  SPEAKER_03: So you're kind of considered unscored, and that can be very difficult then for for getting other credit.
• 4:12
  SPEAKER_03: So having something like a bank account early on in life gets you into the system, and that can be uh uh helpful as you get older.
• 4:26
  SPEAKER_02: So you're suggesting that people start their credit history early.
• 4:33
  SPEAKER_03: Yeah, uh you want to be you want to have something in the file.
• 4:36
  SPEAKER_03: Uh particularly, it doesn't mean that you have to borrow to start your credit history.
• 4:40
  SPEAKER_03: You can have a bank account or a credit union account.
• 4:43
  SPEAKER_03: Uh if you're lucky enough to have a parent that can add you to a credit product as a co-signer, that can also be very helpful.
• 4:51
  SPEAKER_03: Your parent uses that responsibly.
• 4:53
  SPEAKER_03: I should say a lot of people inherit problematic credit histories because their parents needed to put somebody's name on to be able to get something, and and uh that's part of the the uh historic inequities and inequality in our existing system.
• 5:08
  SPEAKER_03: So, you know, like many things in life, you can't control your parents, uh, but life's a lot better when you have good ones.
• 5:16
  SPEAKER_02: Before AI, how did banks decide who got alone and what fundamentally changed when the algorithms took over?
• 5:25
  SPEAKER_03: Yeah, so the beginning of who got alone is something called desktop underwriting, which kind of happens around the 1980s.
• 5:35
  SPEAKER_03: So we can debate when this algorithm becomes artificially intelligent or not.
• 5:41
  SPEAKER_03: It's a little more complicated of a question, but it's a simpler question to say what happened before we had computers.
• 5:48
  SPEAKER_03: So before we had computers, there were kind of two things.
• 5:52
  SPEAKER_03: One is there were regional credit reports, so these big three credit reporting agencies still existed.
• 5:58
  SPEAKER_03: In fact, there were more than them, and and people didn't move around as much.
• 6:02
  SPEAKER_03: Information was harder and more expensive to gather.
• 6:05
  SPEAKER_03: So there were sometimes to query a local area and get some a file that the bank would then make a judgment on.
• 6:14
  SPEAKER_03: And the banks had uh loan officers and credit qualities.
• 6:18
  SPEAKER_03: A lot of this was done in person.
• 6:20
  SPEAKER_03: So when you wanted a loan, you had to go down to the bank, you had to meet the banker, the banker looked at you, they decided whether they liked you or not, they decided whether they knew you.
• 6:30
  SPEAKER_03: Uh back in this time, there were no banks in multiple states.
• 6:34
  SPEAKER_03: It's difficult for young people to think that Bank of America wasn't everywhere in America.
• 6:39
  SPEAKER_03: But uh, do you know when they allowed banks to have branches in more than one state?
• 6:46
  SPEAKER_01: No.
• 6:47
  SPEAKER_03: No, it's it's not until the mid-1990s, with the passage of a law called the Regal Neal Act.
• 6:54
  SPEAKER_03: In fact, before that, several states like Illinois had what was called unit banking, where a bank could only have one branch.
• 7:01
  SPEAKER_03: So most states banks could exist, like Bank of America was in California, up and down the state, but in some states like Illinois, you could only have one branch.
• 7:09
  SPEAKER_03: Um and so in 1980, we had about 16,000 community banks.
• 7:14
  SPEAKER_03: And part of the bank's job was to know the people in the community.
• 7:17
  SPEAKER_03: And so the bank's job was to kind of know you.
• 7:20
  SPEAKER_03: Uh often they'd also know your parents.
• 7:23
  SPEAKER_03: Uh, and so those judgments, and they were judgment calls.
• 7:27
  SPEAKER_03: Um uh this led to a lot of biases in the system, not surprising.
• 7:34
  SPEAKER_03: Uh uh, you know, uh, and so uh Congress in the 1970s passed some laws making it illegal to extend credit on certain functions.
• 7:45
  SPEAKER_03: We'd call them protected class, race, gender.
• 7:49
  SPEAKER_03: It included marital status.
• 7:51
  SPEAKER_03: Unmarried women were having a hard time getting credit.
• 7:53
  SPEAKER_03: Bankers would say, Well, you seem like a good credit risk, but tell me about your husband.
• 7:58
  SPEAKER_03: And uh they said, Well, we're not married.
• 8:00
  SPEAKER_03: Oh, I don't know about that.
• 8:02
  SPEAKER_03: Uh, I mean, that's very difficult for your generation to believe, but trust me, that was such a widespread problem that that Congress passed a law um called the Equal Credit Opportunity Act or ECOA.
• 8:14
  SPEAKER_03: But before computers, it was done very much on a person-to-person basis with some information from these regional credit reporting agencies, and a lot was based on reputation.
• 8:24
  SPEAKER_03: Starting in the 1980s, computers could kind of plug this system in and it became more automated.
• 8:30
  SPEAKER_03: And then the computer algorithm started, and this FICO number started uh as what's called a credit score.
• 8:37
  SPEAKER_03: So you have a credit report which has the information.
• 8:41
  SPEAKER_03: By the way, everybody can get a free credit report.
• 8:44
  SPEAKER_03: We wrote that into a law I helped write on called the Fact Act.
• 8:47
  SPEAKER_03: So you can ask for a free report.
• 8:49
  SPEAKER_03: They'll try to sell you the score, which is the quantified number.
• 8:54
  SPEAKER_03: Uh, but a lot of these reports have errors.
• 8:56
  SPEAKER_03: So I do recommend uh, you know, find out what what these credit bureaus know about you or think they know about you.
• 9:02
  SPEAKER_03: And it and as a hint, the more common your name, the more likely you are to have an error.
• 9:08
  SPEAKER_03: There's an Aaron Klein in New Jersey who messed me up.
• 9:13
  SPEAKER_02: You've been chief economist of the Senate Banking Committee, a senior official at the Treasury, and now a senior fellow at workings.
• 9:22
  SPEAKER_02: Look at us from where you were as a student.
• 9:24
  SPEAKER_02: How someone ends up doing those careers and what the common thread was.
• 9:40
  SPEAKER_03: So it's a very, very good question.
• 9:43
  SPEAKER_03: Uh I was a a high school student in Silver Spring, Maryland at Montgomery Blair High School.
• 9:51
  SPEAKER_03: Shout out, Blair Blazers, my eldest daughter, is a sophomore at Blair herself.
• 9:56
  SPEAKER_03: Uh the building she goes into is a different building than the one that I went.
• 10:00
  SPEAKER_03: Uh and this is where my journey in public service started.
• 10:04
  SPEAKER_03: Uh, up until that moment, I was kind of interested in making money, corporate America, uh uh different, you know, my my dreams of being a pro athlete had dissipated uh when I realized I was uh my athletic ability was limited.
• 10:20
  SPEAKER_03: But I was pretty good at math.
• 10:22
  SPEAKER_03: Uh, and so my math ability I thought could take me pretty far, certainly further than my athletic ability.
• 10:29
  SPEAKER_03: And uh, but I noticed something, which was the high school I went to was really run down.
• 10:35
  SPEAKER_03: The building was from the 1930s, we had uh holes in the walls, rain when it rained, the hallway was wet, it was super overcrowded, we didn't have a parking lot, we didn't even have a single athletic field.
• 10:49
  SPEAKER_03: But I lived in a very rich county.
• 10:51
  SPEAKER_03: Uh and my school was was the uh highest share minority in the county, uh, and incredibly run down.
• 10:58
  SPEAKER_03: And the school where my sister had gone was sparkling, really nice, had a domed basketball gym, and really rich.
• 11:06
  SPEAKER_03: And there was a recession in the early 1990s when I was uh in high school, and the county could only afford one new school.
• 11:14
  SPEAKER_03: And so they choose to rebuild the rich school and let our school continue to suffer.
• 11:20
  SPEAKER_03: And that struck a deep cord in me as being morally wrong.
• 11:25
  SPEAKER_03: And from the experience of organizing, I became high school class president, I became vice president of our countywide student government, and organizing with other people in the community, we let a community rally to put pressure on our elected officials to change their vote.
• 11:41
  SPEAKER_03: And in point of fact, we did.
• 11:43
  SPEAKER_03: They voted from five to four not to rebuild our school to six to three to rebuild our school.
• 11:48
  SPEAKER_03: They built our new school, which opened well after I'd gone, but my daughter goes there now.
• 11:53
  SPEAKER_03: And that experience was so deeply meaningful and moving in my life, I felt a call to public service.
• 12:01
  SPEAKER_03: I said, I've been given some ability in life.
• 12:04
  SPEAKER_03: I was born into a, you know, uh with a wonderful mother, uh, and and I want to use this to make the world a better place, not just to make myself uh uh rich and happy.
• 12:16
  SPEAKER_03: And so I went off to college uh with a desire for public service, and I wasn't entirely sure how to do that.
• 12:23
  SPEAKER_03: Uh, and I went to Dartmouth College, wasn't quite sure what I wanted to do.
• 12:27
  SPEAKER_03: I got won these elections, I thought elected office was cool, I did some internships, and I got a wonderful opportunity to intern in the United States Senate for then a somewhat middle-of-the-road seniority senator named John Kerry of Massachusetts.
• 12:42
  SPEAKER_03: And John Kerry's office in the Senate showed me that the legislature was where it was at.
• 12:48
  SPEAKER_03: And that's what I really wanted to do.
• 12:50
  SPEAKER_03: I wanted to work in the Senate.
• 12:52
  SPEAKER_03: And my time in college showed me that my mathematical skills would be best used with this tool and framework of economics as a way to see the world, to advance the public cause, to advocate for people who are not as fortunate, as blessed as me, who didn't have the ability to dedicate their life for public service for whatever reason, but to speak for those who couldn't.
• 13:19
  SPEAKER_03: I also saw that to get a good job in the Senate, the kind of job where I thought I could make a difference, I needed a graduate degree.
• 13:26
  SPEAKER_03: Uh, or to work a very long time and work my way up.
• 13:29
  SPEAKER_03: And uh when I finished college, uh, the economy was booming in the late 1990s, and it was a little bit easier to get into grad school.
• 13:38
  SPEAKER_03: For those of you uh uh who don't know, uh uh it's easier to get in grad school during a good economy because other people work uh when the economy is weak.
• 13:47
  SPEAKER_03: People are like, oh, my options are aren't as good in the office, so I'll go back and pick up a grad degree.
• 13:53
  SPEAKER_03: Uh and so I was very fortunate to go straight through and get accepted to Princeton School of Public and International Affairs, where I focused on economics even further.
• 14:04
  SPEAKER_03: And then when I graduated, it was really hard to get a job because I had all these fancy degrees and very little work experience.
• 14:11
  SPEAKER_03: But eventually, uh good fortune struck me one time again where the senior senator from my state, the great senator Paul S.
• 14:18
  SPEAKER_03: Sarbanes, was looking for a young economist.
• 14:22
  SPEAKER_03: Uh, and uh I was a young economist from Maryland, and he was the senior senator from Maryland who had who ran the staff of the Senate Banking, Housing and Urban Affairs Committee.
• 14:33
  SPEAKER_03: And so lo and behold, uh slightly before my 25th birthday, I got hired to be the professional staff economist on the committee, uh, where I rose to become chief economist.
• 14:44
  SPEAKER_03: Uh, and I'll tell you, uh I've had these four jobs.
• 14:48
  SPEAKER_03: You mentioned the uh chief economist on Senate Banking, Deputy Assistant Secretary of Treasury, which I got lucky enough to get an appointment when President Obama was elected in 2008, and I got my appointment in early 09.
• 15:02
  SPEAKER_03: And then after my career in government ended, uh I got several jobs at uh think tanks.
• 15:08
  SPEAKER_03: Uh and and you know, uh Katie, I like to joke that I've been unqualified for all of the four jobs I've had, but I think I've done a pretty good job in each of them.
• 15:18
  SPEAKER_03: So, does that mean I was unqualified or not if I did a good job doing it?
• 15:23
  SPEAKER_03: Uh uh each of the times, some of the people who've had those jobs pre and before me uh have been incredibly uh successful and qualified people, but you know, I I was very fortunate to have a stint in government.
• 15:36
  SPEAKER_03: What I will say is after 12 and a half years in government on the Senate Banking Committee and then in the Treasury Department working through the financial crisis, I came to realize a couple things.
• 15:48
  SPEAKER_03: Uh one was I was tired.
• 15:52
  SPEAKER_03: The financial crisis took a lot.
• 15:55
  SPEAKER_03: I once worked 30 straight days and 30 straight nights uh in the fall of 2008, uh, helping pass what's best known as the TARP program, and then going to implement it.
• 16:08
  SPEAKER_03: I also believe one of the strengths of American democracy is that people take turns.
• 16:13
  SPEAKER_03: Uh, if you go back to the ancient Greek and Roman republics upon which this country was founded, they often stressed the need for temporary government service.
• 16:23
  SPEAKER_03: That government service is best done when it's shared.
• 16:28
  SPEAKER_03: Uh, nobody has a right to a career in government.
• 16:32
  SPEAKER_03: And there was a line of people out the door that wanted my job.
• 16:37
  SPEAKER_03: Those jobs are very competitive.
• 16:39
  SPEAKER_03: I'd had four years as deputy assistant secretary, I'd done a lot.
• 16:42
  SPEAKER_03: I'm very proud of my record, but I also believed other people take a turn.
• 16:47
  SPEAKER_03: Uh, I had my second daughter on her way, uh, and I was also realizing, you know, I need to think and reflect a little bit about what I've done and give someone else a turn.
• 16:59
  SPEAKER_03: Uh so I left my appointment.
• 17:01
  SPEAKER_03: I joined the think tank world where I continue to write, research, and uh uh promote ideas to try to impact the world.
• 17:09
  SPEAKER_03: So I've not left public service, uh, I've just taken a different uh uh path into influencing that, uh, and one in which uh has given me some other opportunities in life and given other people opportunities in life in government.
• 17:25
  SPEAKER_03: I I kind of wish that some of the older generation would realize that they've had their turn uh and and give some make a little room for for new thoughts and new people.
• 17:35
  SPEAKER_03: And look, maybe maybe the person who took my job after me was worse.
• 17:39
  SPEAKER_03: It's possible, but it's also possible they were better.
• 17:42
  SPEAKER_03: Uh uh, and you know, I think we all need a little humility.
• 17:45
  SPEAKER_03: Uh, you know, rarely are you the greatest in the world at doing something.
• 17:51
  SPEAKER_02: There can only be one greatest.
• 17:54
  SPEAKER_03: There's there's only one goat, and uh uh uh and that's fine.
• 17:58
  SPEAKER_03: You know, but but by the way, it's uh that that's fine.
• 18:02
  SPEAKER_03: You don't have to be the greatest to make a massive impact in the in the world.
• 18:07
  SPEAKER_02: I love how the need for public service is what drove you.
• 18:12
  SPEAKER_03: Yeah, look, uh there are only two professions that people talk about having a calling for: the priesthood and public service.
• 18:20
  SPEAKER_03: Right?
• 18:20
  SPEAKER_03: Divinity, there's a calling, and public service people say, I've never met anyone who said I had a calling to be an investment banker.
• 18:27
  SPEAKER_03: Like, look, those are great jobs.
• 18:29
  SPEAKER_03: I'm not putting anything down.
• 18:30
  SPEAKER_03: And again, public service has a lot of sacrifices, and not everybody's in the same position to make those sacrifices, and that's great.
• 18:37
  SPEAKER_03: You got to do you.
• 18:39
  SPEAKER_03: Um but for me, it there's a fulfillment.
• 18:43
  SPEAKER_03: Uh, there's a very good article written by um uh the Dalai Lama and Arthur Brooks.
• 18:51
  SPEAKER_03: Arthur's a professor at Harvard and previously was the president of the American Enterprise Institute, which is the kind of more conservative version of the Brookings Institute, located next door to us in Washington.
• 19:03
  SPEAKER_03: And they talked about the need the value of work in fulfilling purpose in life.
• 19:10
  SPEAKER_03: And what they said, I think, is very true, which is that to be happy as a human being, you need purpose.
• 19:17
  SPEAKER_03: And purpose can come from many things.
• 19:20
  SPEAKER_03: It doesn't have to come from work.
• 19:21
  SPEAKER_03: Purpose can come from being a parent, purpose can come from being a good child, purpose can come from taking care of an animal.
• 19:28
  SPEAKER_03: There's lots of paths to purpose in life.
• 19:32
  SPEAKER_03: But work is one way in which people get purpose and which people make a contribution.
• 19:40
  SPEAKER_03: And so I encourage every young person to think about what gives them purpose and then fulfill that because that's how you receive a higher level of fulfillment in life.
• 19:52
  SPEAKER_03: You you you will not find your purpose and happiness just from you know eating Doritos and playing video games.
• 19:59
  SPEAKER_03: That can be fun.
• 20:01
  SPEAKER_03: That can be a period and a moment where you do a little bit every day, but that's not gonna both those things are gonna leave you empty inside, even if they taste good in the moment.
• 20:11
  SPEAKER_02: Well, that's deep.
• 20:14
  SPEAKER_03: Well, yeah, it is.
• 20:16
  SPEAKER_03: When I first started reading some of this stuff about work, I thought, well, you know, this is a little bit of mass owners of capital telling the masses of society, do stuff for me.
• 20:26
  SPEAKER_03: But when you read the Dalai Lama, uh, and he talks about it's very true.
• 20:30
  SPEAKER_03: Everybody gets purpose from a different thing, and that's fine.
• 20:34
  SPEAKER_03: The people I know in life who are happiest are the people who found their purpose.
• 20:38
  SPEAKER_03: And for me, public service is a core part of that.
• 20:41
  SPEAKER_03: Being a parent is the other.
• 20:44
  SPEAKER_02: So now to talk a little about your work, your research shows AI credit systems can analyze what device you use to apply, where you shop, even your browsing behavior.
• 20:55
  SPEAKER_02: What might that system already know about a person who just applied for their first credit card?
• 21:01
  SPEAKER_03: It could know quite a bit.
• 21:02
  SPEAKER_03: It could know where you go to school, it could know where you live, it could have a rough guess as to how much your parents make.
• 21:09
  SPEAKER_03: Uh, the amount of information that's out there about people is wild.
• 21:13
  SPEAKER_03: And it's a little bit of a black box.
• 21:16
  SPEAKER_03: Uh, what can and can't be used in terms of your history is unclear.
• 21:22
  SPEAKER_03: Uh what my research has led me to believe is that established banks uh tend to be a little more conservative in the data they put in, and newer upstart fintechs, and other lenders tend to be a little more aggressive.
• 21:35
  SPEAKER_03: Non-traditional data can be a lot more accurate in terms of predicting what will happen.
• 21:41
  SPEAKER_03: It can also be a lot more biased.
• 21:44
  SPEAKER_03: Uh, there's a paper I'm quite fond of from a professor at Duke, where she looked at people who were uh applying for credit while they were shopping in Germany, and she said, Well, the credit was.
• 22:00
  SPEAKER_03: Decided on the basis of essentially German FICO.
• 22:02
  SPEAKER_03: She said, but instead I'm going to take something called your digital footprint, which Acadia has five things: time of day, Mac or PC, desktop phone or tablet, domain name of your email, and whether your name is in your email.
• 22:21
  SPEAKER_03: And then she ran a computer optimization program and found that if you weighted those things the right way, you would be more accurate than German FICO.
• 22:31
  SPEAKER_03: Right?
• 22:31
  SPEAKER_03: A couple hints on this shopping at two in the morning is probably bad.
• 22:38
  SPEAKER_03: Gmail is a better credit risk than Hotmail.
• 22:44
  SPEAKER_03: And shopping on your phone isn't necessarily a great sign.
• 22:50
  SPEAKER_03: That being said, in the US, most, if not all, of those five things could be considered illegal under creating bias.
• 23:01
  SPEAKER_03: Hint uh uh minorities are more likely to be able to access their internet only through their phone.
• 23:10
  SPEAKER_03: Apple users skew far whiter than PC users.
• 23:16
  SPEAKER_03: Uh and people who have their name in their email tend to have more what I'd call traditional names.
• 23:22
  SPEAKER_03: Uh there's a very good paper by an economist from the University of Chicago, Marianne Bertrand, who took two identical resumes and gave one a very white sounding name and one a very African American sounding name, and sent them out online and just tracked responses.
• 23:41
  SPEAKER_03: And the responses were so more heavily white that it took the African American having like another level of education, so a college degree versus high school or a master's degree versus college to overcome the bias from name.
• 23:56
  SPEAKER_03: So on the one hand, you may say, well, if your name's in your email or not, that shouldn't be biased.
• 24:01
  SPEAKER_03: People can control their own email.
• 24:03
  SPEAKER_03: The other point says people are lived experience, and the lived experience of having a very uh uh common or white name is very different than having a uh commonly identified minority name, and those people then respond differently to kind of hide their name a little bit because they've had this experience applying for jobs, applying for credit, other things.
• 24:26
  SPEAKER_03: Uh so it's unclear in the US context what is and isn't allowable under the law.
• 24:34
  SPEAKER_03: Uh right now, I would guess during the Trump administration, whether it's allowable or not, there's you're not going to find too much enforcement on some of these discriminatory issues.
• 24:43
  SPEAKER_03: The Trump administration has objected to the major tool of enforcement of anti-discrimination known as disparate impact, a theory that the Supreme Court has said is okay, but that doesn't mean much if the current administration isn't going to enforce it.
• 24:58
  SPEAKER_03: But you know, the answer to your question is people could use a lot of information, some of which is probably allowable, some of which isn't, and different providers of credit are using different information.
• 25:11
  SPEAKER_02: So then, right, how is your credit score consistent if these companies are not even using the same type of data?
• 25:22
  SPEAKER_03: Right.
• 25:23
  SPEAKER_03: So your score, your FICO score is your score from FICO.
• 25:27
  SPEAKER_03: But what most many lenders will do is they'll take that as one input and then they'll add their own speech special sauce.
• 25:34
  SPEAKER_03: And that's part of the magic of capitalism, uh, which is we all have that we pay for some level of common data, and then we all think we can do it a little better than our competitor, right?
• 25:45
  SPEAKER_03: By better, meaning I can give you a little bit of a better rate because I've determined that you're less risky than people who look similar uh to you by the same credit score.
• 25:56
  SPEAKER_03: Uh for a consumer's point of view, it's very hard to know how to make yourself look less risky.
• 26:01
  SPEAKER_03: And that's by design, right?
• 26:03
  SPEAKER_03: If they if you knew how to improve your appearance, people would look better even when they really aren't, right?
• 26:09
  SPEAKER_03: You'd you'd kind of fake it a little bit.
• 26:12
  SPEAKER_03: Uh, if you knew that your email address changed your credit score, people would create a new email address just for for that purpose.
• 26:21
  SPEAKER_03: Uh so it's it's it's confusing by design.
• 26:25
  SPEAKER_02: You've written about inferring from someone's bank records that they might be going through a divorce and using that to deny credit.
• 26:32
  SPEAKER_02: What is the problem with that?
• 26:40
  SPEAKER_03: So the problem with that is it's illegal.
• 26:43
  SPEAKER_03: So, as we talked about a little bit earlier, way back before computers, when there was a lot more gender discrimination in society, and there's still a bunch of issues, but that doesn't mean it hasn't gotten better.
• 26:56
  SPEAKER_03: Uh uh, people were not giving credit to working women the way they were to married women.
• 27:03
  SPEAKER_03: So Congress wrote into law your marital status is illegal to use in allocating credit.
• 27:09
  SPEAKER_03: Period.
• 27:10
  SPEAKER_03: End of story.
• 27:11
  SPEAKER_03: Now, how you figure that out is actually more complicated, particularly when you apply for credit individually.
• 27:16
  SPEAKER_03: There's a famous case where the Apple card, all right, Apple's attempt to get into credit card lending with Goldman Sachs ran into trouble, uh, pretty high profile because some uh very rich uh uh influencers on uh social media were getting denied because the credit card was underwriting based on their income, not their their joint income.
• 27:38
  SPEAKER_03: And you may be famous, doesn't mean that you're necessarily wealthy, but these people were married to rich people, and they were saying, well, wait a second, how am I getting denied?
• 27:48
  SPEAKER_03: So there's an interesting question whether you should use how how you should do all of that.
• 27:52
  SPEAKER_03: But the very simple point is your marital status is just point blank not allowed.
• 27:57
  SPEAKER_03: End of story.
• 27:58
  SPEAKER_03: That being said, the two highest predictors of whether or not you're going to go bankrupt are getting divorced and a healthcare crisis.
• 28:10
  SPEAKER_03: Those are the top two drivers of bankruptcy in America.
• 28:12
  SPEAKER_03: It is not people who are like, oh, I'm gonna borrow a bunch of money and not repay and have a party.
• 28:17
  SPEAKER_03: That's not the common, you know, bankruptcies are usually caused by those two reasons.
• 28:24
  SPEAKER_03: Healthcare things we generally think of as being out of your control, right?
• 28:27
  SPEAKER_03: I mean, you know, you skateboard a lot, you may break your leg, but generally speaking, those are considered out of your control.
• 28:34
  SPEAKER_03: Divorce, less so, um, but uh uh so it's an interesting economic question.
• 28:40
  SPEAKER_03: Some people would say, well, wait a second, if we if you're more likely to get a divorce, you are a riskier borrower.
• 28:46
  SPEAKER_03: And Acadia, this is the kind of core threshold question, which is that all allocation of credit, people tend to support risk-based underwriting.
• 28:56
  SPEAKER_03: Riskier people should pay more.
• 28:58
  SPEAKER_03: But they also don't like discrimination.
• 29:00
  SPEAKER_03: If it's a risk factor beyond your control, then that's unfair.
• 29:05
  SPEAKER_03: And so you shouldn't pay more, which then means that there's gonna be a cross-subsidy.
• 29:10
  SPEAKER_03: That is, the people who don't have that riskier thing beyond their control are gonna pay more, and the people who have it beyond their control are gonna pay less.
• 29:18
  SPEAKER_03: And there's a big tension in society that's never been well resolved between those two things.
• 29:25
  SPEAKER_03: The situation about marital status and divorce is an example of that, and it's just an example where society ended up in one place, not another.
• 29:34
  SPEAKER_03: Uh, and that's because of the history that we discussed.
• 29:37
  SPEAKER_03: A different example would be driving.
• 29:39
  SPEAKER_03: I don't know if you've gotten your driver's license yet.
• 29:43
  SPEAKER_01: No.
• 29:46
  SPEAKER_03: Um, so I'm taking my eldest daughter for her permit uh tomorrow.
• 29:52
  SPEAKER_03: And what I can tell you is that when she does start driving and I start paying insurance, I will be paying less than my neighbor across the street who has a son her age.
• 30:02
  SPEAKER_03: Why?
• 30:04
  SPEAKER_03: Because boys are riskier drivers, particularly 16-year-old boys.
• 30:09
  SPEAKER_03: And in car insurance, which is regulated at the state level, not the federal level, society has allowed gender to be one of the larger factors for underwriting, particularly for young drivers.
• 30:23
  SPEAKER_03: By the time you're my age, you have a driving record, and the insurance company can see how many accidents you've gotten in and you know make a prediction based on that.
• 30:33
  SPEAKER_03: But when you're 16, your first time behind the wheel, they try to do a couple factors.
• 30:37
  SPEAKER_03: If you're an honored student, you'll get a little bit of a discount because people who drive, people get higher grades, tend to drive a little more safe.
• 30:44
  SPEAKER_03: But the number the number one factor is your gender.
• 30:47
  SPEAKER_03: Totally legal in car insurance.
• 30:49
  SPEAKER_03: In fact, big deal.
• 30:51
  SPEAKER_03: Beyond your control.
• 30:54
  SPEAKER_03: In credit card borrowing, men default more than women, even when you control for income and other factors, men are more likely to default.
• 31:02
  SPEAKER_03: Uh, but totally illegal to use gender in your credit card.
• 31:05
  SPEAKER_03: Why?
• 31:06
  SPEAKER_03: Because the federal government wrote that into law.
• 31:09
  SPEAKER_03: Uh, and again, so this is an example I use about how the same characteristic gender is not allowed in lending, but is allowed in insurance, two different financial transactions.
• 31:22
  SPEAKER_03: That's just the way the world is at the moment.
• 31:25
  SPEAKER_03: Whether you should have it in both or neither, interesting intellectual conversation.
• 31:30
  SPEAKER_03: I find it difficult to understand why it's in one and not the other, other than the history, the legal status, and other factors.
• 31:38
  SPEAKER_03: Abstractly, it's hard to justify using gender for one financial product but not another.
• 31:45
  SPEAKER_02: So, with the use of these black box AI and algorithms, how do we know whether these companies are using this illegal data or not?
• 31:57
  SPEAKER_03: Very good question.
• 31:58
  SPEAKER_03: So the companies, when they turn you down, are required to tell you why.
• 32:03
  SPEAKER_03: And this was put in there in order to try to catch folks, because if they were turning down a lot of people who shared the same characteristics, you could catch discrimination.
• 32:14
  SPEAKER_03: I tend to think that's not terribly effective in a modern society because they can turn you down for 10 different reasons, and they can pick one, and it's you know not clear you can do anything about it.
• 32:25
  SPEAKER_03: Not a long enough credit history.
• 32:27
  SPEAKER_03: Well, I'm 18.
• 32:29
  SPEAKER_03: How can I have a longer credit history?
• 32:31
  SPEAKER_03: Uh, you know, I can, you know, we all age at the same rate.
• 32:35
  SPEAKER_03: Uh, so I don't think often there are scandals, things come out, there are lawsuits, uh, banks are regulated.
• 32:45
  SPEAKER_03: We set up a consumer financial protection bureau in the Dodd-Frank Act of 2010 to regulate non-banks providing financial services so that there'd be a cop on the beat who would look through folks lending.
• 32:58
  SPEAKER_03: In fact, that CFPB sued and uh for returned billions of dollars to consumers who'd been harmed.
• 33:05
  SPEAKER_03: Unfortunately, that agency has largely been shuttered by the Trump administration that ideologically doesn't believe in a consumer financial regulator.
• 33:14
  SPEAKER_03: Um, so you know, it can be difficult.
• 33:18
  SPEAKER_03: We have a system where lawyers are able to make a lot of money if they find uh problems in this, some of which is sent back to the people who are harmed, meant to be a bit of a check on industry.
• 33:30
  SPEAKER_03: Bank regulators and others have uh you know are able to go through the banks who are providing this to enforce the rules.
• 33:39
  SPEAKER_03: Uh, and the non-banks were supposed to be being enforced by the Consumer Financial Protection Agency.
• 33:45
  SPEAKER_03: So that's how the system is supposed to work.
• 33:48
  SPEAKER_02: If a listener gets denied a loan today, can you walk them through briefly what their options are?
• 33:55
  SPEAKER_03: Well, they should be told why they're denied the loan.
• 33:58
  SPEAKER_03: So that then they they should get an answer from the lender on that.
• 34:02
  SPEAKER_03: Uh, and then the option is to try to get a loan from a different person or try to get a different type of loan.
• 34:09
  SPEAKER_03: So there are loans that are secured and unsecured.
• 34:13
  SPEAKER_03: So unsecured is just you borrowed me money and I give you a promise.
• 34:16
  SPEAKER_03: That's kind of a credit card.
• 34:18
  SPEAKER_03: Then there's secured loans.
• 34:19
  SPEAKER_03: A car loan is secured by the car.
• 34:22
  SPEAKER_03: There are secured credit card loans where I've sent you a bit of money ahead of time before.
• 34:27
  SPEAKER_03: Uh, and so there are different types of loans.
• 34:30
  SPEAKER_03: Then you can also have a co-signer, that is a parent or guardian, who says, Well, if you don't pay back, I'll pay.
• 34:37
  SPEAKER_03: And so that person adds more security.
• 34:39
  SPEAKER_03: So those are the common kind of alternative options.
• 34:47
  SPEAKER_02: What's one risk AI in the financial system you think is underestimated?
• 34:52
  SPEAKER_02: And what should young people do to prepare for it?
• 34:57
  SPEAKER_03: So when you ask the question about the financial system, it's kind of, I think of things like the whole system.
• 35:04
  SPEAKER_03: Having lived through a financial crisis, what can make the whole system go up or go down or collapse?
• 35:10
  SPEAKER_03: I don't think young people can prepare themselves for that type of systemic risk, right?
• 35:16
  SPEAKER_03: Like you can't prepare yourself for COVID coming and shutting down schools.
• 35:20
  SPEAKER_03: You can deal with the consequences, you can mitigate it, but that's kind of beyond your control.
• 35:25
  SPEAKER_03: In fact, financial crises are beyond the control of most people.
• 35:29
  SPEAKER_03: That's why government is supposed to set up special dedicated things to stop them from happening.
• 35:36
  SPEAKER_03: Um, one of the biggest risks of AI is some sort of system that can crack security.
• 35:44
  SPEAKER_03: So our financial system is built on trust.
• 35:48
  SPEAKER_03: You give the bank money, you trust the bank keeps it.
• 35:51
  SPEAKER_03: You get an insurance contract, you trust that when there's an accident, the insurer pays you back.
• 35:57
  SPEAKER_03: That trust involves some level of security, right?
• 36:01
  SPEAKER_03: I have an account, I am me.
• 36:04
  SPEAKER_03: There's not going to be fraud.
• 36:06
  SPEAKER_03: One of the bigger risks in AI is the ability to crack security, and that's not something people can do.
• 36:13
  SPEAKER_03: You can try to have a little bit better security, have stronger passwords, keep your information private.
• 36:20
  SPEAKER_03: I think a lot of younger people are often a little loosey-goosey with some of their privacy things.
• 36:25
  SPEAKER_03: But ultimately, if there's some quantum hacked breakthrough, it is what it is.
• 36:31
  SPEAKER_03: Um, and I I'm not sure you can do much, but I do spend a little bit of time worrying about what we'd call cyber risk or cyber threats, uh, that an artificial intelligence can hack a system.
• 36:43
  SPEAKER_03: And uh I think that's a big deal.
• 36:47
  SPEAKER_03: I don't think many people can do too too too much to protect yourself uh other than having what they call good cyber hygiene.
• 36:57
  SPEAKER_02: If you're speaking directly to a student right now about the single most important financial decision they'll make in the near future, what do you think it is and what do they need to know?
• 37:14
  SPEAKER_03: It's a very good question.
• 37:15
  SPEAKER_03: I'm an economist, not an investor, so I hesitate.
• 37:18
  SPEAKER_03: Let me give a little disclaimer like you know, the the uh uh I'm not sure I'm the best financial advice giver for a person in their life.
• 37:27
  SPEAKER_03: But I think I would say the following things to them.
• 37:29
  SPEAKER_03: Number one, make sure you have a bank or credit union account.
• 37:32
  SPEAKER_03: Those are the safest places to store your money.
• 37:35
  SPEAKER_03: The federal government will insure you for the first$250,000.
• 37:39
  SPEAKER_03: Banks and credit unions are part of this financial system where you can grow and establish your credit history, your credit report.
• 37:47
  SPEAKER_03: They will help you provide credit.
• 37:50
  SPEAKER_03: Uh uh, and so please have an account.
• 37:53
  SPEAKER_03: If you need a parent to have one, please make sure your parent gives you your own account.
• 37:58
  SPEAKER_03: If you're in an unfortunate situation where you can't trust your parent, see if there's another legal guardian or somebody else who can help you who you can trust.
• 38:07
  SPEAKER_03: That I think is the most important thing you can you can do there.
• 38:11
  SPEAKER_03: Number two, um, try to save a little bit, even if it's only one or two percent of everything you earn.
• 38:19
  SPEAKER_03: A culture and history of savings is incredibly important, particularly when you're young.
• 38:25
  SPEAKER_03: Life is a lot easier when you always have a little bit of money in your bank.
• 38:29
  SPEAKER_03: We have a financial system where if you run out of money, it's very expensive to get a little bit of money.
• 38:36
  SPEAKER_03: And if you have money, it's much cheaper to get a little borrow a little bit of money.
• 38:41
  SPEAKER_03: So always building a little cushion in life.
• 38:45
  SPEAKER_03: Even if you don't use it, you say, Well, wait a second, why do I always have this hundred dollars in my bank account?
• 38:50
  SPEAKER_03: There's a concert I could go to, you know, there's this or that, and then I'll rebuild it.
• 38:55
  SPEAKER_03: Well, it takes a lot of work to get that first hundred dollars saved in your bank account.
• 38:59
  SPEAKER_03: Uh uh, try to leave life with a little bit of a cushion.
• 39:03
  SPEAKER_03: When there's a true emergency, you'll know and you'll know when to use it.
• 39:08
  SPEAKER_03: Uh, but having that little culture that says, hey, wait a second, I got this money.
• 39:13
  SPEAKER_03: Uh, to the extent these new bank accounts are they're called, you know, uh uh savings uh investment accounts are coming out for people under the age of 18 and and are getting funded a little bit, trying to invest when you're young.
• 39:26
  SPEAKER_03: It's a lot easier to invest when you're young.
• 39:29
  SPEAKER_03: Interest compounding interest is one of the great phenomenon of life.
• 39:34
  SPEAKER_03: And so saving a little bit when you're young will make a huge difference as you get older and want to buy that first car, want to buy that first house, want to have that first child, uh, all of which are important milestones that can give you a lot of uh purpose and fulfillment in life.
• 39:52
  SPEAKER_02: Are there any ways that AI could help those who've been historically locked out?
• 39:59
  SPEAKER_03: Absolutely.
• 40:01
  SPEAKER_03: Uh AI is a huge just like computers helped end a lot of discrimination that was going on at the bank branch, AI can help end a lot of discrimination that's going on through these algorithms, particularly for kids, kids whose parents left them not an inheritance of money, but an inheritance of debt and bad credit reports.
• 40:23
  SPEAKER_03: Uh I think AI could be tremendously transformative because the current system we have is not terribly accurate and not terribly fair.
• 40:31
  SPEAKER_03: Uh and I I hold a lot of promise that AI and advanced algorithms, and frankly, other types of data.
• 40:40
  SPEAKER_03: Whether or not you were excuse me, whether or not you were on time with a credit card payment five years ago is not nearly as important as whether or not you've been able to keep a few hundred bucks in your bank account every day for the last year.
• 40:57
  SPEAKER_03: But that's just not how the current reporting system works.
• 41:00
  SPEAKER_03: So there's a tremendous amount of opportunity to use new data uh to advance a lot of these problems that the current system is not going to solve.
• 41:09
  SPEAKER_03: The current system is not trying to create the most accurate credit report possible.
• 41:14
  SPEAKER_03: The current system is trying to create the cheapest credit report that it can sell to as many people as possible and make and make as big a profit.
• 41:22
  SPEAKER_02: So what might be standing in the way of that positive version?
• 41:30
  SPEAKER_03: So there's a lot.
• 41:32
  SPEAKER_03: Uh I don't know, Katie.
• 41:33
  SPEAKER_03: Did you ever play an instrument in the orchestra?
• 41:36
  SPEAKER_00: No.
• 41:37
  SPEAKER_03: Yeah, I did.
• 41:39
  SPEAKER_03: Uh and uh in fact, I played uh the second least common instrument in the orchestra called the oboe, which is a double-readed instrument.
• 41:48
  SPEAKER_03: Um and the oboe, if you ever go to an orchestra, uh and you'll hear that the orchestra starts tuning up right before they play, they tune to the oboe's A.
• 42:01
  SPEAKER_03: And do you know why they tune to the oboe's A?
• 42:05
  SPEAKER_00: No.
• 42:06
  SPEAKER_03: So it's not because the oboe's A is the most accurate A.
• 42:11
  SPEAKER_03: Uh the most accurate A is gonna be the concert master violinist.
• 42:16
  SPEAKER_03: The reason they tune to the oboe is the oboe is the hardest instrument to tune.
• 42:20
  SPEAKER_03: And so if the oboe is just gonna be a touch flat, the whole rest of the orchestra is gonna be just a touch flat.
• 42:28
  SPEAKER_03: Because then they'll sound really good.
• 42:30
  SPEAKER_03: Because unless you have the the goal of the orchestra is not individuals each seeking to replicate the perfect sound.
• 42:37
  SPEAKER_03: The goal of the orchestra is that everybody makes the same sound.
• 42:41
  SPEAKER_03: It's that conformity.
• 42:43
  SPEAKER_03: That's the goal of the financial system.
• 42:46
  SPEAKER_03: The goal of the financial system is to maximize profits for the system, which is best done when everybody individually maximizes profits in a way that works systematically with the other players in the system.
• 43:00
  SPEAKER_03: It is not accuracy.
• 43:02
  SPEAKER_03: It is to be in tune with each other, so this giant financial system of credit bureaus, credit reporting agencies, investors, savers, borrowers, blah, blah, blah, blah, blah, can all work together.
• 43:13
  SPEAKER_03: And so I've often said that the entire financial system is tuned to FICO, these credit scores, which are like an out-of-tune oboe.
• 43:24
  SPEAKER_03: And so, you know, what's standing in the way is it's very difficult to get everybody to retune.
• 43:30
  SPEAKER_03: And if a few people start doing things a little more accurately, the rest of the system isn't going to incorporate that.
• 43:36
  SPEAKER_03: And it's going to become difficult, they're not going to be able to go ahead and do that.
• 43:42
  SPEAKER_03: And so the first is the whole leap of the system is standing in their way, which is very difficult.
• 43:48
  SPEAKER_03: I built a better system, but nobody can plug in and use it.
• 43:53
  SPEAKER_03: The second point is that we have some out we have some difficult laws and regulations that has I described before, have good purposes, but are difficult to square with new data.
• 44:07
  SPEAKER_03: So is it discriminatory to use whether or not you're going to get a divorce?
• 44:13
  SPEAKER_03: What things are and aren't allowed?
• 44:15
  SPEAKER_03: What things are proxies for things that are and aren't allowed?
• 44:18
  SPEAKER_03: Racial discrimination pad shouldn't have it.
• 44:21
  SPEAKER_03: What's a proxy for it?
• 44:22
  SPEAKER_03: What isn't?
• 44:23
  SPEAKER_03: Very difficult.
• 44:25
  SPEAKER_03: I think if FICO were a new technology today, it would be outlawed.
• 44:28
  SPEAKER_03: It would fail all these discriminatory tests.
• 44:31
  SPEAKER_03: But because it's been built up over so many years, it's been grandfathered in, it's been allowed.
• 44:36
  SPEAKER_03: Just like how much money you have is allowed, even though there's a massive correlation between wealth and race, but it's considered allowed.
• 44:45
  SPEAKER_03: And so new information is held to a higher standard than existing information, which perpetuates the existing system.
• 44:53
  SPEAKER_03: So there's laws, there's regulations, there's rules, there's the system itself operating.
• 44:59
  SPEAKER_03: And then lastly, as if that's not enough to overcome, there's been a growing disagreement in society about what should and shouldn't be allowed.
• 45:08
  SPEAKER_03: And we have a very schizophrenic position on that, as I illustrated with a point about gender and car insurance versus credit.
• 45:17
  SPEAKER_03: And so we need to have a moment where they say common heads prevail.
• 45:22
  SPEAKER_03: We need to have a moment where we can have a more honest conversation across society and say, okay, look, we have to build a system.
• 45:29
  SPEAKER_03: The system's inherited.
• 45:30
  SPEAKER_03: What we inherited wasn't great, wasn't fair, wasn't equitable.
• 45:34
  SPEAKER_03: We have a tension between risk-based pricing, cross-subsidization, immutable characteristics choices, and you have to make some tough choices.
• 45:42
  SPEAKER_03: Nothing's going to be perfect.
• 45:44
  SPEAKER_03: Something's going to be led to either too much cross-subsidization or too much reliance on immutable characteristics.
• 45:54
  SPEAKER_03: And we have to make, we have to kind of hold hands and make some agreements about what is and isn't allowed.
• 45:58
  SPEAKER_03: Should you be able to use Mac versus PC?
• 46:01
  SPEAKER_03: On the one hand, that's very correlated with race.
• 46:04
  SPEAKER_03: On the other hand, I don't feel like Macs is going out, you know, trying to get just white people.
• 46:09
  SPEAKER_03: I I don't think their advertising campaign is dedicated.
• 46:12
  SPEAKER_03: I don't think it's a product with a legacy in history that way.
• 46:15
  SPEAKER_03: I could be wrong, other people can disagree, but we need to be able to have those kinds of conversations.
• 46:19
  SPEAKER_03: Why are Mac users bet better credit risks than PCs?
• 46:22
  SPEAKER_03: Is that a is that, you know, what is that really telling us?
• 46:30
  SPEAKER_02: For a student who wants to work on how these financial systems are governed, what might be an optimal path to get there?
• 46:52
  SPEAKER_03: The first thing is work hard in school.
• 46:54
  SPEAKER_03: Uh people judge based on your educational attainment, uh, and whatever schools you go to, do the best that you can.
• 47:03
  SPEAKER_03: Second thing is meet people.
• 47:05
  SPEAKER_03: When you're young, you have an opportunity to meet people, whether it's at public events, lectures, forum, you know, uh particularly uh uh at book signings.
• 47:17
  SPEAKER_03: You would be shocked, and Acadia, you wouldn't be shocked, but many of your listeners might be.
• 47:22
  SPEAKER_03: Just reach out to people.
• 47:23
  SPEAKER_03: People will respond.
• 47:24
  SPEAKER_03: You reached out to me, I didn't know you, but you know, I said this is an opportunity, uh, and I was impressed by the fact that you reached out to me.
• 47:32
  SPEAKER_03: And so people and networks make a massive difference in your life.
• 47:38
  SPEAKER_03: Uh, and so if you want to get into this world, you got uh you got there's an old story of a Chicago politician who uh ran a political machine, and a person came to him and said, I want to be part of the machine, and he said, Who sent you?
• 47:56
  SPEAKER_03: And the person said, No one sent me, I want in.
• 47:59
  SPEAKER_03: And he said, I don't want no one that no one sent me.
• 48:04
  SPEAKER_03: And so just getting to know people and networking.
• 48:08
  SPEAKER_03: Some people have a little natural advantage because of where they were uh uh came from, but you know, everybody can pick up an online profile and send a cold email, or you see somebody who's written a book that you thought was fascinating and is coming through your town.
• 48:24
  SPEAKER_02: What does this field look like for someone entering it in the next five years?
• 48:29
  SPEAKER_02: And how is AI changing what financial policy and regulation careers require?
• 48:37
  SPEAKER_03: Um, so I think AI is changing it in the sense that the more you know AI and the more you know how to use it, it will be very valuable.
• 48:46
  SPEAKER_03: People used to want somebody who could write computer code, now somebody wants somebody who can instruct an AI to write it.
• 48:52
  SPEAKER_03: So being able to be more proficient in AI, I think, will be a massive benefit.
• 48:57
  SPEAKER_03: Also, being able to understand the limitations of AI.
• 49:00
  SPEAKER_03: I think a lot of older people are like, this is good, but it's not perfect, so I'm a little scared and nervous, and I'm not quite sure how to incorporate it.
• 49:07
  SPEAKER_03: So I think being involved in it will make a big difference.
• 49:12
  SPEAKER_03: Um I think the field will be using AI more for research.
• 49:17
  SPEAKER_03: I think in some ways it will be harder.
• 49:18
  SPEAKER_03: The traditional place to break in in academia was as a research assistant, and I think AI is doing the jobs of research assistance in ways that will make people senior people like me need research assistance less.
• 49:33
  SPEAKER_03: Uh, so I think you'll have to it'll it'll make it a little harder to break in.
• 49:39
  SPEAKER_02: So before you leave, are there any books or podcasts you would recommend to a student who wants to learn more?
• 49:47
  SPEAKER_03: It's a very good question.
• 49:48
  SPEAKER_03: I listen to uh Peter Conti Brown's uh Substack, uh Wharton professor on financial regulation.
• 49:57
  SPEAKER_03: I listen to Rob Blackwell's Banking with Interest podcast.
• 50:01
  SPEAKER_03: Uh those are two that that that are frequently high on on my list.
• 50:06
  SPEAKER_03: Um there are a bunch of good books.
• 50:10
  SPEAKER_03: Uh often in financial regulation the books are a little different depending on what subtopic you want.
• 50:16
  SPEAKER_03: Uh, but but but I think those uh that podcast and that substack would be where where I would start.
• 50:23
  SPEAKER_02: Well, thank you so much for joining us today.
• 50:27
  SPEAKER_03: Thank you so much for having me, and congratulations for for focusing on this.
• 50:31
  SPEAKER_03: It's a wonderful thing that you're doing, and and I hope your listeners enjoy it very much.
• 50:42
  SPEAKER_02: This has been cause of curiosity in an AI world.
• 50:46
  SPEAKER_02: Exploring careers through the lens of AI.
• 50:49
  SPEAKER_02: Until then, stay curious.