Mathematicians

High pay & lowrisk Trending
Minimal Risk
Low High

Explore safer careers (2)

Lower estimated automation risk

Computer and Information Research Scientists
11% automation risk | Minimal Risk
Pays better Higher growth
7.9 pts lower View career
Why it fits

Fits mathematicians using algorithms, theory, modeling, proofs, and computational research.

Mathematical Science Teachers, Postsecondary
13% automation risk | Minimal Risk
Higher growth More jobs
5.6 pts lower View career
Why it fits

Uses advanced mathematics, proofs, problem sets, explanation, mentoring, and curriculum design.

Alternative careers

Related career paths that build on similar skills and experience

Biostatisticians
19% automation risk | Minimal Risk
Higher growth More jobs
Similar risk View career
Why it fits

Applies statistics, probability, study design, inference, and scientific reporting to life-science data.

Statisticians
22% automation risk | Low Risk
Higher growth More jobs
View career
Why it fits

Directly reuses mathematical theory, modeling, proofs, inference, uncertainty, and quantitative communication.

Compare this job with others
or search for another job
Share your results with friends and family.

Occupation snapshot

What does this snowflake show?
The Snowflake is a visual summary of the five badges: Automation Risk (calculated), Risk (polled), Growth, Wages and Volume. It gives you an instant snapshot of an occupations profile. The colour of the Snowflake relates to its size. The better the occupation scores in relation to others, the larger and greener the Snowflake becomes.
JOB SCORE
5.4/10
What's this?
Job Score (higher is better):

We rate jobs using four factors. These are:

- Chance of being automated
- Job growth
- Wages
- Volume of available positions

These are some key things to think about when job hunting.

Risk & user votes

Calculated automation risk

19% (Minimal Risk)

Minimal Risk (0-20%): This occupation appears difficult to replace end-to-end with current or near-future automation, including AI software and robotics. Roles in this range usually depend on human judgement, creativity, care, leadership, specialist expertise, or adapting to messy real-world situations. AI and machines may still change parts of the work, but the occupation is likely to remain a distinct human role.

More information on what this score is, and how it is calculated is available here.

Human strengths important in this job

These are human abilities and work contexts that are important in this occupation. They may help explain why parts of the role are harder to replace end-to-end, but they are not the only inputs into the automation score.

Thinking creatively

Very important
Why this matters
Coming up with original ideas and designs—creating new concepts, products, systems, or artistic work. This kind of open-ended invention and taste-based judgment is harder to automate end-to-end than routine, rule-based tasks.
Jobs that also use this strength

Decision-making and problem solving

Very important
Why this matters
Analyze information, weigh tradeoffs, and choose the best solution—especially when situations are ambiguous, high-stakes, or have real-world consequences.
Jobs that also use this strength

Active learning

Very important
Why this matters
Keeps learning from new information and applying it to make better decisions now and in the future, especially when situations change.
Jobs that also use this strength

Developing objectives and strategies

Quite important
Why this matters
Sets long-term goals and chooses strategies and actions to reach them, weighing tradeoffs and adapting plans as conditions change.
Jobs that also use this strength

Communicating with people outside the organization

Quite important
Why this matters
Represents the organization to customers, the public, or government—handling questions, concerns, and relationship-building through conversations, writing, calls, or email.
Jobs that also use this strength
Show 1 more strength

Education and training expertise

Quite important
Why this matters
Designing and delivering instruction—adapting lessons to different learners and measuring whether training actually works.
Jobs that also use this strength

What users think

Based on 662 votes

46% chance of full automation within the next two decades

Our visitors have voted they are unsure if this occupation will be automated. However, employees may be able to find reassurance in the automated risk level we have generated, which shows 19% chance of automation.

What do you think the risk of automation is?

What is the likelihood that Mathematicians will be replaced by robots or artificial intelligence within the next 20 years?

Sentiment

Based on user votes over time

View sentiment trend

How opinions have changed over time

Pay & outlook

Wages

Very high paid relative to other professions

In 2024, the median annual wage for Mathematicians was $121,680 ($58 per hour).

The median annual wage for Mathematicians was 145.8% higher than the national median annual wage, which stood at $49,500.

View wage trend

Wages over time

* Data from the Bureau of Labor Statistics

Growth

Slow growth relative to other professions.

The number of 'Mathematicians' job openings is expected to decline 0.7% by 2034

View employment trend

Total employment, and estimated job openings

* Data from the Bureau of Labor Statistics for the period between 2023 and 2033
Updated projections are due 09-2025.

Volume

Significantly lower range of job opportunities compared to other professions

As of 2024 there were 2,220 people employed as 'Mathematicians' within the United States.

This represents around < 0.001% of the employed workforce across the country

Put another way, around 1 in 69 thousand people are employed as 'Mathematicians'.

People also viewed

Computer Programmers Lawyers Web Developers Actors Electrical Engineers

What people are saying (28)

Leave a comment
Nigg (Low)
21 May 2026 19:18
Nah, a mathematician's job also requirers the creative combination of previous research in order to leap forward.AI will only boost the efficiency of this process .For example, manual painstaking tasks such as literature reviews will be accelerated by AI
singe rapide (Moderate)
08 Apr 2025 21:25
AI is getting better by the day at solving problems it hasn't seen before. But proving mathematical theorems isn't just about formal manipulation, mathematicians rely on their senses, physical experience, intuition etc. I think AI will bypass these creative difficulties to produce interesting math eventually, but it will take a while.
FERNANDO  GARCIA CORTES (Uncertain)
09 Dec 2024 16:21
I suppose it will depend on how able the IA is to reply to the process of thinking a mathematician has when he/she makes a demonstration.
1 (Low)
28 Nov 2024 23:06
Because is hard and needs creativity but also computers can use programs
Guest
29 Jul 2024 08:08
Aaaand Google DeepMind invented AI (AlphaGeometry 2 and AlphaProof) that solves IMO on silver medal. That escalated quickly...
Nigg
21 May 2026 19:25
Umm..ok I am in no doubt that AI can do math.The issue is that it cannot do math like a human.Sure, it can absorb data from IMO math exercises but and learn from it.But I am betting that if you prompted on the said AI (and any AI really)"solve riemann hypothesis, make no mistakes" ,it absolutely cannot do it
Guest
24 May 2026 18:39
Well neither can we as humans. We can make progress sure, but any progress in these types of problems hardly come from solving the problem itself. They come from someone solving some different problem which then gives us a new set of tools to work with.
Kevin Gonzalez Gordillo
07 Jul 2024 02:05
Lately, I have been studying Gödel's Incompleteness theorems, and I was thinking, could an AI, or even an AGI, discover that in a consistent system of arithmetic, there are true statements that cannot be proven (incomplete system)? I mean the level of self-awareness and nuance that it would take to discover such a theorem and prove it true would be astounding since not even the best mathematicians of Gödel's times were convinced that was the case. So at this level of logic, I would really doubt an AI would take the jobs of people responsible for opening the heart of mathematics and showing the gaping hole of its fundamental flaw; for any mathematician reading this: Let's keep proving the true, the false, and the unprovable truths, and let's "embrace the unknown at the heart of any quest for truth" -(TED-Ed).
papo (No chance)
14 Jun 2024 01:03
Mathematics is a science that is still in discovery, AI will not be able to understand this context
skyss0fly (Moderate)
27 Apr 2024 03:59
Robots have to utilise logic in their programming using arithmetic skills and therefore use mathematics to perform tasks and as such can compute mathematical tasks quite easily
Chickenlover69 (No chance)
05 Jul 2024 08:40
Being a mathematician is researching math itself, like science. It needs creativity and human intelligence. For example, I certainly don't think AI can discover or invent complex numbers such as i, i = sqrt(-1). But I think you misunderstood the job of a mathematician as you probably thought that they just solve regular math questions that we already know how to solve, no it's not. Math is part of science and will always will be. Science will never be able to be taken over by AI. Remember that.
Anonymous (Uncertain)
11 Dec 2023 05:40
Logic can be programmed, so they may be able to eventually prove theorems
Europe2048 (Moderate)
15 Sep 2023 12:57
Maybe WolframAlpha is taking over?
Me (No chance)
24 Aug 2023 06:50
Mathematical insights rely on human originality.
Anish (Highly likely)
08 Jul 2023 02:47
It will simply outclass us in every field.
Probabilist
23 May 2023 17:39
Mathematicians generally find new ideas to solve problems, they also find new questions, whereas AI is just able to copy what already exists and to adapt it to classical situations. But maybe mathematicians will be able to create themself a new AI which will be able to replace them !!
Andre (Highly likely)
17 Mar 2023 23:39
People saying this is unlikely are silly. The progress in the field of computer generated proofs has been faster than the progress in AI art. I don't expect any of the Millennium Prize problems to be solved by humans, and I expect them all solved by 2030.
Jason
10 Jul 2024 03:50
I find it equally silly when you say that none of the Millennium Prize problems will be solved by Humans and all of them will be solved by 2030 (I consider myself an amateur mathematician at best, but from what I understand, the Poincare conjecture has been solved, and the solver, Grigori Perelman I think he's human?) I've hanged out in some mathematics communities and from what I've seen ai has proved some competitive math questions, I guess the four color theorem requires computer assistance, and I've seen Terrence Tao say that Ai could be used to help give Ideas to mathematicians? I've tried searching up ai replacing mathematicians and i only get results of lean which from what i know really only checks if a theorem is right or not. Overall I find the statement that Ai will complete all of the Millennium problems by 2030 to be ridiculous Ai does not seem to be anywhere close to proving some of the most fundamental century old problems in essentially 6 years
Carlos (No chance)
11 Jun 2022 21:36
I believe that even general AI will require a lot of time to reach a level where it can be compared on a full basis to a human in regards to maths.
Saurabh Sarkar (No chance)
06 Nov 2021 12:11
AI was created from Math. Its development is dependent on the development of new Math. It will be next to impossible for AI to develop new functional math.
James (Uncertain)
28 May 2021 08:49
Because, AI and supercomputer can overtake the common knowledge
Jo
21 Apr 2021 10:09
If you've ever done college level math and the number of ways you've been told to "Do it by inspection" and the unconventional method that every solution takes, Math is very unlikely to become automated.
AK (No chance)
29 Mar 2021 13:18
Mathematics can go so far, there are literally so infinite possibilities. That may sound extreme but there are so many number systems out there that mathematicians are uncovering everyday as a team and they still can not find all of them. A computer may be faster but since there are so many possibilities both mathematicians and computers must work TOGETHER.
Alex (Low)
15 Dec 2020 14:32
Mathematics requires quite a bit of creativity, which AI is not capable of.
Leave a reply about this occupation
0/8000

Job description

Conduct research in fundamental mathematics or in application of mathematical techniques to science, management, and other fields. Solve problems in various fields using mathematical methods.

O*NET-SOC code: 15-2021.00