Data Scientist

Data science is one of the jobs AI can reach most easily, because almost all of it happens on a computer and most of it is working with information. A 2025 Microsoft Research study measured how much of each job's tasks AI chatbots can help with, using 200,000 real conversations. Data scientists came out 19th of 785 jobs, in the top 3% and above software developers. Anthropic's March 2026 study, which combines what AI could do in theory with how Claude is actually used at work, puts data scientists 18th of 756 jobs in its published data. Both companies sell AI, so read these as their measurements, not neutral ones. The routine work (cleaning and checking data, writing code, writing up results) is the kind of computer work these studies find AI helps with most, and it is the work new data scientists have usually been hired to do while they learn.

The clearest warning sign is in job ads. The job site Indeed tracks postings by job group. In mid-September 2026, postings for data and analytics jobs were 38% below their February 2020 level, the weakest of the 47 groups it tracks, while postings overall were 3% above. Since ChatGPT came out at the end of November 2022 they have fallen 55%, against about 30% for all postings. Two cautions: the slide began earlier in 2022 as tech hiring cooled, so not all of it is AI, and job ads are not hires. They also cover every level, not only entry-level jobs.

The government sees it differently, and you should see its number too. BLS projects data scientist jobs to grow 35% from 2025 to 2035, much faster than the average for all jobs, with about 24,800 openings a year. BLS expects companies to need more data scientists partly because they are adding AI to their own work. That growth is the reason this is not a 5. We still rate it a 4, higher than the BLS number alone suggests, because the work that makes the job possible to enter is the work AI is taking, and the job ads show that way in getting narrower. The weak spot in our case: we have not found hiring data for new data science graduates specifically. Anthropic found only tentative signs that hiring of 22 to 25 year olds into the most AI-exposed jobs has slowed, and no rise in unemployment.

What would change this score. If data and analytics job postings climb back toward their pre-pandemic level while AI tools keep spreading, or data on new graduates shows them being hired into data science as easily as before, this should come down to 3. If BLS's own count of data scientists starts to fall, it goes to 5. Inside the field, the safer ground is the work AI is weaker at: asking the right question, knowing a business deeply, and judging whether a result is actually true.