Daughter's career choices and AI

In My Humble Opinion
23

An awful lot of actuaries are also data scientists these days, and most of the rest work with data scientists, at least in the property-casualty field (also called general insurance), which is where i worked and what i know best.

You can certainly combine them. For a similar combination; I have a friend who started out doing something with computers. Then he decided to become a nurse. He was a clinical nurse, directly caring for patients, and as others have pointed out, the hours sucked. In particular, he had a lot of trouble taking vacation when he wanted to. (If five people do statistics on Monday, and all take the rest of the week off, you may start the next week in about the same place as if each had worked a day. That’s just not true of caring for patients.) After a couple of years, he ended up supporting the software that a nursing team used, with more regular and flexible hours.

But speaking about actuarial work and ai… At this point, i think no one knows how that will play out.

I started my actuarial career just as computers got substantially cheaper. I started with a terminal to a big computer across the country that needed a human to manually load data when i ran a job, went through a PC on every desk, moved to PCs saving data to a local server, and now i have a PC that talks to “cloud storage”, which has a lot in common with what i started with, except for being vastly faster, more powerful, and cheaper.

And when spreadsheet apps started catching on, there were questions about how that would affect actuarial work. Back then, actuaries spent a lot of time filling out spreadsheets, which were giant sheets of paper covered in hand-written calculations.

There are many more actuaries employed today than there were back then. There were, indeed, some dislocations and reductions in, for instance, jobs for entry level pension actuaries (caused by computerizing the basic work) but it turns out that an actuary wielding a spreadsheet (or more complex computer programs) is a lot more valuable than an actuary with just a desktop calculator.

At least for now, actuaries face more competition from other brands of data scientists than from AI. My guess is that the jobs will change again due to AI, but there will continue to be a role for human beings to ask useful questions and interpret the answers. But who knows.

24

This. With the added note that a lot of students discover in the course of their nursing degree that it isn’t what they want to do after all, or that it isn’t a good match for their aptitudes. (This is true for most degree programs, to some extent – the majority of undergraduates change majors at some point – but it seems to be particularly common with nursing, at least at my institution.) It’s a good idea to go in with an open mind, and to remember that most required courses can be applied toward other degree programs as well.

25

One thing she could do is get a degree which overlaps her two interests, like Biostatistics. That’s a degree which uses statistics to analyze biological systems. The degree would give her a basis in both biology and statistics. That would allow her to branch into many different careers related to nursing and/or stats.

One thing to consider is that any career may go away sometime in the near future. AI may take away stats jobs. Immigration might take away nursing jobs. So rather than focusing on a specific career, consider a degree path which can lead to that career as well as many others.

26

Yeah, I tend to think of actuaries as building predictive models based upon established distributions and validated assumptions, and data scientists as pulling trends out of data and using them to iteratively update ‘priors’ to make ad hoc predictions but I’m sure the reality is much more blended and with a lot of emphasis to moving toward Bayesian modeling given the computing power and tools to support the less rigorous data science approach. I have some level of concern about actuarial models built by ‘AI’ because they are so much a black box versus being able to explicitly see how a traditional actuarial model is built, and I don’t know how you can ever rigorously validate a model without that insight but frankly I’m already seeing this in other types of modeling like guidance and control systems (which scares the hell out of me). Regardless, we will still need actuaries to evaluate models and decide what data to feed into them because no model makes makes reliable predictions given bad assumptions or garbage inputs.

Stranger

27

My sister was a research nurse at Thomas Jefferson University Hospital in Philadelphia, where she ran clinical drug trials for many years. I’m sure statistics were her constant companion. Back when I was a broke undergrad, she signed me up for a number of those studies. The pay was great… even if I did grow a third ear. :joy:

When she retired from Jefferson, she spent a few years as a grade school nurse at a school in her town. She found both roles rewarding—just in very different ways: one involved data and drug protocols; the other involved Band-Aids, and the occasional glitter emergency.

Nursing offers many options.

28

Hi, there, OP here:

thx for the really thoughtful answers.

I didn’t want to introduce too many variables in my OP … but family is located in Chile, which I would describe as smallish 2nd world country, amongst many large 3rd world countries. For you ‘muricans, thinking about the state of Alabama (similar per capita gdp), without option to leave the state should set the scene, mentally.

There are no majors/minors that you can hop between in the local system. A score system gets you into a given career at a given Uni (or not) … If you want to switch after a year, you lost a year of your life, probably 10k in tuition and are back to square one - so is best avoided.

But my (hopefully unbiased) diagonal reading of the answers so far is more o less:

  • nursery is potentially way more future-proof than statistics, on which’s future the jury still is out, but a statistics degree might (or might not) have a shelf life shorter than a can-of-tuna as AI progresses.
  • there is a way wider spectrum of jobs in nursing than the tv-trope “bedside” nurse. Obv. way less if you are limited to “Alabama” (as opposed to the whole of the US), but there are options. (I “think”, nurses in the public system are not very well paid here).
29

Alabama is large enough to have lots of options in nursing, and a fair number in statistics, too. Especially if “Alabama" were it’s own nation, and didn’t lean on other states for some technical support.

Can your daughter pursue nursing and also study statistics?

30

realistically, not … pay 2 studies, handle 2 completely unrelated timetables/coursworks, the number of U that offer Stat. is very limited (2-3) … and those campuses are nowhere near the campuses of nursing. I rather envision study nursing and then specialize in diplomas/courses towards the part of the spectrum that she is interested in.

Don’t get me wrong, Chile is also 20 mio people milling around and gettin ill/medical treatment, It just seems that there are fewer of those special (and more sexy) roles that you have in a 330+ mio country … so hence my “isolate Alabama” mental crutch.

31

If geography is a factor, I think it’d be a lot harder to get a nursing job abroad (would need recertifications at a minimum?) than a statistics or AI one. If she wants to stay in Chile and work there for a local company for her whole life, that’s fine. But statistics might give her more mobility and transferability to other countries, or for remote work in her own country (and possibly higher wages)?

The high wages of some US nurses is also a byproduct of both our insurance system and regional supply and demand. There’s clusters of small areas where healthcare demand is high, doctors and hospitals are present, but there aren’t enough nurses and those who are there are subject to high costs of living. Those conditions can give rise to higher wages than you might see elsewhere in the world, or in “Alabama”. Not sure how it is in Chile, but you might want to check?

On the other hand, if she’s perfectly happy working in Chile, yeah, I imagine nursing would be more futureproof. Statistics seems like a dead man walking unless she can directly that into AI, and even then, it’s quite the hype bubble right now and it might not last too long. Hard to say.

Can she do something more specialized, combining both healthcare and quantitative stuff, like biomedical engineering?

32

My best guess is that a statistician wielding AI is going to be extremely powerful. I don’t know how many jobs there will be for those people, but the ones that exist will be pretty decent. Someone (someone human) will be needed to point the AI in the right direction and separate the wheat from the chaff.

33

I didn’t mean take two entirely separate college programs at the same time. That sounds crazy. I meant either find something that blends them, or study some stats on her own, in addition to the formal nursing program. I dunno about Chile, but in the US, nursing is heavily regulated and you need the right degree and stuff, but stats is less so, and you can often prove competence and find employment via other means.

34

Yeah, I’ll again reiterate that the notion of ‘AI’ obsoleting expertise is just flatly wrong. ‘AI’, in various forms, is a useful and potentially powerful tool that can be wielded usefully by experts (and dangerously by ignorant people who believe it makes them an expert) but it is not a substitute for the ability to apply knowledge in real world applications. Even in the case of AGI, someone has to ‘prompt’ the system to formulate the problem and provide useful results, and then verify and apply those results. If being a statistician or data scientist merely meant mastering the calculations and methods of statistics and data analysis, they would already be obsolete because we have had computer tools that do that much better and faster than any human for many decades. Doing those jobs means understanding the correct use of the applicable method and interpreting results in a critical and useful way.

Stranger

35

Yes, that was the point of my story about actuaries and the proliferation of spreadsheets and other computer calculating power. A few jobs did go away, but many other jobs were created.

36

And in fact, those tools give an expert far more ability to do far more powerful analysis than by hand. ‘AI’, if properly used, can also increase the ability to tease out trends and do more complex analyses, but not if you just ask ChatGPT vague questions and don’t know enough to suss out erroneous or useless responses.

Stranger

37

To your point, ultimately AI, ML, or data science in general has to solve some business problem. At least a problem other than “not wanting to pay wages to humans”.

I’d worry less about “AI gonna take our jobs” and more about what the daughter actually wants to do day to day at worl:

Nurse - Take care of sick patients, clean bedpans, take blood, and so on.

Statistics - Work as an actuary or some other sort of analyst, data scientist, investment banking quant, etc.

I disagree about statistics necessarily being solitary, focused work. Yes, you might spend more time at your computer crunching numbers. But in my experience, there is still a lot of human interaction. Perhaps more than an introverted number cruncher might like.

38

A statistician (or someone working in a job that relies heavily on statistical analysis) will certainly spend some time working on a computer to build models, sift through and classify data, produce visualizations that demonstrate a causal claim or tell a narrative, et cetera but they also have to interact with the real world to understand the nature of the data, explain their results in a way others can understand, and often overseeing or performing the collection of data. One of the biggest traditional fields that relies upon statistics and data science are the quantitative social sciences (sociology, psychology, anthropology, archeology, linguistics and communication science, political and cultural science, et cetera) all of which typically involve a lot of human interaction. You can certainly find jobs in statistics which are sit-behind-a-desk-and-calculate jobs but those are actually the most likely to be marginalized by AI and more advanced tools; the jobs that require a combination of analytical skills and human interaction are the least likely to be automated completely away.

Stranger

39

This is not an accurate nor fair description of what a nurse does. It sounds like an uninformed stereotype.

from a US nurse of 44+ years…

40

I fully agreed, from the brother of two recently retired nurses and the Uncle of a Nursing Professor.

41

My Mom was in nursing for her entire career and has two Ivy League masters and doctorate degrees in nursing related fields. I’ve also been to a hospital or doctors office from time to time. I know there’s more to nursing than just taking care of patients. But it’s still a fundamentally part of nursing as I understand it.

42

Health care is going to be a more reliable career choice than a field like statistics. For one thing, medical care involves a mix of cognitive skills and manual labor skills, which robots will eventually replace but not for a while. However purely cognitive tasks are going to be automated far sooner. Even if cognitive tasks aren’t fully automated, AI will help one employee do the work that used to take 3-4 employees. At a company I once worked at, they laid off a huge % of their technical writing team recently due to AI. The AI can’t replace all the technical writers, but it can write the bulk of a document, then the employee proofreads and edits the AI written document. So they cut a huge number of technical writers. Granted, in the long run, maybe the demand for cognitive labor will go up and people will get hired back, but even with that cognitive careers are going to see huge boom/bust cycles with long bouts of unemployment and uncertainty.

Also with health care, there is government regulation. We could have self driving cars in 2025, but government regulation exists because we want the technology to be near perfect before it gets released widely. The same thing will happen with medicine. Even if AI becomes extremely good at medicine, there will be pushback against full automation even when AI is capable of full automation in health care.

As far as nursing, has she looked at any career in health care other than nursing? There are a lot of careers in the medical field, and nursing is just one of many. There is a massive nursing shortage, but that isn’t because we lack people trained as nurses. There are million(s) of people in the US who are trained as nurses who choose not to work as nurses. However the job is so miserable for endless reasons that a lot of people leave the field after they get their training.