[New Study] Econ/Finance majors are the most exposed to AI
A new study "The impact of AI on UK jobs and training" (UK Gov, Nov 2023) revealed the most exposed subjects using a scoring system.
Top 5 Most Exposed Subjects:
- Economics (Level 7+ studies): 1.26
- Accounting and finance (Level 4/5 studies): 1.22
- Mathematics and statistics (Level 7+ studies): 1.15
- Accounting and finance (Level 6 studies): 1.07
- Law and legal services (Level 7+ studies): 1.04
Top 5 Least Exposed Subjects:
- Sport, leisure and recreation (Level 4/5): 0.11
- Leisure, Travel and Tourism (Level 4/5): 0.10
- Engineering (Level 4/5): 0.08
- Sociology and social policy (Level 4/5): 0.05
- Animal care and veterinary science (Level 6): -0.03
Study Main Conclusions:
- The finance & insurance sector is more exposed to AI than any other sector.
- Employees with more advanced qualifications are typically in jobs more exposed to AI. For example, employees with a level 6 qualification (equivalent to a degree) are more likely to work in a job with higher exposure to AI than employees with a level 3 qualification (equivalent to A-Levels).
- Employees with qualifications in accounting and finance through Further Education or apprenticeships, and economics and mathematics through Higher Education are typically in jobs more exposed to AI. Employees with qualifications at level 3 or below in building and construction, manufacturing technologies, and transportation operations and maintenance are in jobs that are least exposed to AI.
What's your take on it? What can we do as Finance majors?
Generative AI is not based on databases, but LLM (or Human Intelligence).
Having said that, employees in accounting and finance will be responsible to give inputs to generate an AI system.
I don´t think that all enterprises will be dependent on specific AI companies, but every enterprise will bring out its own AI (i.e.: GS has own AI, MS has own AI, JPM has own AI etc.). And to generate own AI, analysts will do their current jobs as always.
So.. Can we just ignore this study? No. I could imagine that the back office will be more exposed to AI than -let´s say- the front office.
Anyway, thanks for the interesting study.
Interesting point regarding AI decentralisation.
The only thing that worries me is that the study states the jobs most at risk are those held by the most educated people, which would include Front Office roles.
AI is particularly good at Data Analysis, which is a big part of Front Office jobs.
Last I checked Chat GPT could easily pass the Bar, but struggled with the CFA Exams.
Yep, I think lawyers are even more at risk because of LLMs
I keep thinking here, and you're not wrong. It gets weird though? How do I train somebody that 20+ year Brinson attributions of shifting portfolios are crap?
Also, there is a lifetime skill about fixing the 5yr problem, which doesn't fix 20 years later.
This is too broad for meaningful insights...accountants and lawyers are VERY replaceable with Gen AI moving forward.
Engineering is a mixed back. If we're talking civil or electrical, then it's pretty insulated. Computer engineering though is VERY replaceable. If you're a skilled sr. engineer today with 10+yrs of experience, you'll continue to do well. If you're a in school for comp sci / new grad / jr. engineer with 3yrs< exp today, you are in for a rude awakening. # of jobs will go down, comp when you switch for another role won't be as high, and job security will be structurally lower. I am very glad I didn't study comp sci knowing what we know now since I wouldn't be a sr. engineer by this point in my life...even then in 5yrs even that may be at risk with only the top 10-20% sr. engineers being truly insulated
Only high-paying, white collar job that is truly protected from AI downside risk is doctors. I unfortunately didn't have the affinity for it but sometimes I wish I pushed harder to become a doctor, though if things fall into play in my career this year as I hope then I won't have any regrets on that front I think...just have to wait and see
Agree.
The only sectors I identified as "fairly protected" are Medicine, Electrical/Mechanical Engineering, Hospitality, Arts & Design.
Senior people may enjoy the benefits of automation, while the big impact will be on new grads.
I wonder though on the arts? Problem is when the tools for creation becomes 10x better (which they are today vs. a year ago) the amount of content EXPLODES. The noise gets bigger and bigger and even spending massive ad dollars (as book publishers, music labels, etc) do is becoming insufficient over time as there's too much crap in the way and too many different channels people are on for discovery. We are just seeing the beginning on AI influencers, AI-developed content (alone or with a human co-piloting), etc...it's already hard enough being a creator, that's going to be 10x harder in a few years as these tools become better
If you're top 1% creator today then you're probably fine for the next 5yrs....but anything below that line will struggle IMO
MID LEVEL, which is where the problem lies. ChatGPT is not going to have an answer when I ring up our head ETF counsel and say "We benchmark to this FTSE index which is our starting universe, but everybody who's an external client benchmarks to EAFA or World ex-US. talking to our net bets vs them may be more worthwhile--Should we do that in the commentary?"
These things are happening because tech companies overhired during the pandemic, not because ChatGPT is writing code. Any company that says otherwise is just trying to make laying off employees sound cool with AI related technobabble.
I'm not a contrarian, and I do plan on at least trying to teach ChatGPT to write SQL queries to my company's databases. But my hypothesis is that because it's trained on short code snippets from Leetcode, stackoverflow, etc. it won't know how to put together larger programs and will only give you things you can easily google.
Am I the only one who's incredibly bearish on generative AI in particular?
To me, ChatGPT is an inferior alternative to Google search when it comes to information retrieval. You can Google something, and state-of-the-art NLP algorithms will retrieve you a list of results, or you can ChatGPT something and you have extremely similar NLP algorithms give a summary of those results. That summarization takes away the ability to see what was actually written and to fact check it.
I think non-generative AI is where the money is. Using statistical models to predict stock prices, hotel room prices, manufacturing outputs and basically any aspect of a business has much more value than any of the low effort junk that ChatGPT spits out. Not to mention generative AI has massive negatives, I think in a few years there's gonna be tons of AI generated garbage clogging up the internet, it's gonna be like spam emails but harder to distinguish.
What does “exposed” even mean ? More replaceable ? More fragile ? More susceptible? More capability to understand AI ? No idea what it’s talking about
'Exposed to' almost always refers to downside career / professional risk
Omnis et est quos placeat. Beatae dolorem fuga et quaerat architecto aliquam fugit. Aut totam sequi enim nesciunt ut nesciunt iste. Et nihil et omnis explicabo nihil nobis ut. Non eos id non excepturi atque omnis reiciendis.
Optio error sit cupiditate enim. Sed eveniet est et molestias numquam quae. Iusto molestiae alias iure qui voluptas. Nihil id provident quas dolorem rerum.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...