Tech jobs safe from AI? The 20 tech roles with the lowest AI-exposure scores
Short answer: no tech job is safe from AI. Every computer and math occupation in Microsoft's 2025 "Working with AI" study sits above the middle of its 785 occupations for generative-AI applicability. The lowest one in the field still lands past the 50th percentile. Writing code, querying data and explaining systems are exactly the work people already bring to AI chat tools.
That 45/100 is the average for the roles on this list. What's yours?
Experience moves this number a long way in both directions, and it is the only thing about you we cannot read off this page. Answer two questions and GPT-4o writes you the 3 moves that keep you on the safe side of this, the skills gap for each, and a 90-day plan — free, in about 30 seconds.
How long have you been working?
Your own risk score and 3 AI-resistant paths start writing on the next screen. Free, no card, about 30 seconds. Skip the questions.
New · every week
New remote roles in your field — with each one's AI-exposure score — every week
10–15 remote postings from the past week, picked from company career pages and the Himalayas and Jobicy boards. Each shows how exposed the occupation is to AI, from Microsoft's 2025 study of 785 jobs. You apply on the original posting.
Our risk score
5–67/100
Microsoft exposure
p55–p93
BLS median pay
$76,220–$140,300
So this list doesn't pretend. It ranks the tech roles from least to most exposed on Microsoft's measure and puts our own replacement-risk score beside each one. You can see where the two agree and where they don't. If you work in tech, the useful question isn't "which job is untouched". It's "which job keeps the parts AI is worst at": owning a production system, judging risk, negotiating requirements with people who don't agree.
Read the order as relative, not absolute. Actuaries and research scientists come out lowest because so much of their work is judgment about models rather than producing text or code. Security, networking and architecture follow, because accountability for live infrastructure doesn't move to a chatbot. Pure production roles, like web development, QA testing and business-intelligence reporting, sit at the exposed end.
How this list was built
Every occupation our database files under Technology, ranked by Microsoft's AI applicability percentile, lowest first. Catch-all "all other" groups are left out because Microsoft scores them as a single block. This is the one list we rank by Microsoft's number instead of filtering on it: no tech occupation is below the median, so a filter would leave nothing.
Each role shows two different numbers. Our risk score (0–100) is the one on its job page. AI applicability comes from Microsoft's 2025 study of Copilot use across 785 occupations. Lower means less exposure on both. Neither one is a promise about any single job.
The list
1.Actuaries
Risk: 45/100- AI applicability 0.16: the 55th percentile of 785 occupations
- BLS median pay: $130,000
Analyze statistical data, such as mortality, accident, sickness, disability, and retirement rates and construct probability tables to forecast risk and liability for payment of future benefits.
A task our analysis keeps with a person: Collaborate with programmers, underwriters, accounts, claims experts, and senior management to help companies develop plans for new lines of business or improvements to existing business.
2.Computer and Information Research Scientists
Risk: 50/100- AI applicability 0.17: the 59th percentile of 785 occupations
- BLS median pay: $140,300
Conduct research into fundamental computer and information science as theorists, designers, or inventors.
A task our analysis keeps with a person: Develop performance standards, and evaluate work in light of established standards.
Full breakdown for Computer and Information Research Scientists →
3.Information Security Analysts
Risk: 56/100- AI applicability 0.22: the 74th percentile of 785 occupations
- BLS median pay: $129,180
Plan, implement, upgrade, or monitor security measures for the protection of computer networks and information.
A task our analysis keeps with a person: Train users and promote security awareness to ensure system security and to improve server and network efficiency.
4.Telecommunications Engineering Specialists
Risk: 50/100- AI applicability 0.25: the 81st percentile of 785 occupations (score for its occupation group)
- BLS median pay: $134,050
Design or configure wired, wireless, and satellite communications systems for voice, video, and data services.
A task our analysis keeps with a person: Implement system renovation projects in collaboration with technical staff, engineering consultants, installers, and vendors.
Full breakdown for Telecommunications Engineering Specialists →
5.Network and Computer Systems Administrators
Risk: 51/100- AI applicability 0.25: the 81st percentile of 785 occupations
- BLS median pay: $99,130
Install, configure, and maintain an organization's local area network (LAN), wide area network (WAN), data communications network, operating systems, and physical and virtual servers.
A task our analysis keeps with a person: Analyze equipment performance records to determine the need for repair or replacement.
Full breakdown for Network and Computer Systems Administrators →
6.Computer Network Architects
Risk: 52/100- AI applicability 0.25: the 81st percentile of 785 occupations (score for its occupation group)
- BLS median pay: $134,050
Design and implement computer and information networks, such as local area networks (LAN), wide area networks (WAN), intranets, extranets, and other data communications networks.
A task our analysis keeps with a person: Coordinate network operations, maintenance, repairs, or upgrades.
7.Software Developers
Risk: 55/100- AI applicability 0.28: the 85th percentile of 785 occupations
- BLS median pay: $135,980
Research, design, and develop computer and network software or specialized utility programs.
A task our analysis keeps with a person: Analyze user needs and software requirements to determine feasibility of design within time and cost constraints.
8.Web and Digital Interface Designers
Risk: 5/100- AI applicability 0.29: the 87th percentile of 785 occupations (score for its occupation group)
- BLS median pay: $104,000
Design digital user interfaces or websites.
9.Video Game Designers
Risk: 45/100- AI applicability 0.29: the 87th percentile of 785 occupations (score for its occupation group)
- BLS median pay: $104,000
Design core features of video games.
A task our analysis keeps with a person: Solicit, obtain, and integrate feedback from design and technical staff into original game design.
10.Database Administrators
Risk: 53/100- AI applicability 0.30: the 87th percentile of 785 occupations
- BLS median pay: $104,620
Administer, test, and implement computer databases, applying knowledge of database management systems.
A task our analysis keeps with a person: Train users and answer questions.
Full breakdown for Database Administrators → · Will AI replace them?
11.Computer Network Support Specialists
Risk: 49/100- AI applicability 0.30: the 88th percentile of 785 occupations
- BLS median pay: $76,220
Analyze, test, troubleshoot, and evaluate existing network systems, such as local area networks (LAN), wide area networks (WAN), cloud networks, servers, and other data communications networks.
A task our analysis keeps with a person: Perform routine maintenance or standard repairs to networking components or equipment.
12.Operations Research Analysts
Risk: 45/100- AI applicability 0.31: the 89th percentile of 785 occupations
- BLS median pay: $88,940
Formulate and apply mathematical modeling and other optimizing methods to develop and interpret information that assists management with decisionmaking, policy formulation, or other managerial functions.
A task our analysis keeps with a person: Perform validation and testing of models to ensure adequacy, and reformulate models, as necessary.
13.Health Informatics Specialists
Risk: 42/100- AI applicability 0.31: the 90th percentile of 785 occupations (score for its occupation group)
- BLS median pay: $105,850
Apply knowledge of nursing and informatics to assist in the design, development, and ongoing modification of computerized health care systems.
A task our analysis keeps with a person: Providing empathy and emotional support
14.Computer Systems Analysts
Risk: 44/100- AI applicability 0.31: the 90th percentile of 785 occupations (score for its occupation group)
- BLS median pay: $105,850
Analyze science, engineering, business, and other data processing problems to develop and implement solutions to complex applications problems, system administration issues, or network concerns.
A task our analysis keeps with a person: Providing empathy and emotional support
15.Computer Programmers
Risk: 53/100- AI applicability 0.31: the 90th percentile of 785 occupations
- BLS median pay: $100,390
Create, modify, and test the code and scripts that allow computer applications to run.
A task our analysis keeps with a person: Perform or direct revision, repair, or expansion of existing programs to increase operating efficiency or adapt to new requirements.
16.Database Architects
Risk: 57/100- AI applicability 0.31: the 90th percentile of 785 occupations (score for its occupation group)
- BLS median pay: $139,500
Design strategies for enterprise databases, data warehouse systems, and multidimensional networks.
A task our analysis keeps with a person: Collaborate with system architects, software architects, design analysts, and others to understand business or industry requirements.
17.Data Warehousing Specialists
Risk: 59/100- AI applicability 0.31: the 90th percentile of 785 occupations (score for its occupation group)
- BLS median pay: $139,500
Design, model, or implement corporate data warehousing activities.
A task our analysis keeps with a person: Providing empathy and emotional support
18.Biostatisticians
Risk: 58/100- AI applicability 0.32: the 91st percentile of 785 occupations (score for its occupation group)
- BLS median pay: $105,650
Develop and apply biostatistical theory and methods to the study of life sciences.
A task our analysis keeps with a person: Design research studies in collaboration with physicians, life scientists, or other professionals.
19.Statisticians
Risk: 67/100- AI applicability 0.32: the 91st percentile of 785 occupations (score for its occupation group)
- BLS median pay: $105,650
Develop or apply mathematical or statistical theory and methods to collect, organize, interpret, and summarize numerical data to provide usable information.
A task our analysis keeps with a person: Develop and test experimental designs, sampling techniques, and analytical methods.
20.Software Quality Assurance Analysts and Testers
Risk: 58/100- AI applicability 0.33: the 93rd percentile of 785 occupations
- BLS median pay: $104,300
Develop and execute software tests to identify software problems and their causes.
A task our analysis keeps with a person: Design test plans, scenarios, scripts, or procedures.
Full breakdown for Software Quality Assurance Analysts and Testers →
Why our score and Microsoft's disagree on tech
Our risk score is 40% the Frey & Osborne computerisation probability, plus a skills and work-activities analysis of O*NET data. It asks how much of a job's task list could be automated. Microsoft's score comes from real Bing Copilot conversations. It asks how often people already use generative AI for the work activities a job is made of, and how well it goes.
A few specialties, like web and digital interface designers, score very low on ours because O*NET lists few routine tasks for them, while Microsoft puts their occupation group near the 87th percentile. When the two numbers split this far, trust the direction they share: tech work is heavily exposed. Plan for a job where AI does more of the typing, and you own more of the decisions.
Big tech roles that rank lower down
These are some of the most-searched tech jobs. They fall below the 20 above on Microsoft's measure, with the scores to show why:
- Data Scientists: risk 8/100 · AI applicability 0.36: the 98th percentile of 785 occupations (score for its occupation group) · BLS median $120,230
- Web Developers: risk 51/100 · AI applicability 0.35: the 97th percentile of 785 occupations · BLS median $92,650
- Computer User Support Specialists: risk 73/100 · AI applicability 0.33: the 95th percentile of 785 occupations · BLS median $61,860
Not on the list? Check your own job
We score 1,000+ occupations. Search yours to see both numbers.
New · every week
New remote roles in your field — with each one's AI-exposure score — every week
10–15 remote postings from the past week, picked from company career pages and the Himalayas and Jobicy boards. Each shows how exposed the occupation is to AI, from Microsoft's 2025 study of 785 jobs. You apply on the original posting.
Frequently asked questions
Which tech job has the lowest AI exposure?
On Microsoft's 2025 study, actuaries and computer and information research scientists have the lowest AI applicability in the technology group, around the 55th to 59th percentile of 785 occupations. That's still above the median occupation. No tech job is low-exposure.
Is software development safe from AI?
Not in the sense of being untouched. Software developers sit around the 85th percentile on Microsoft's AI applicability measure and score in the mid-50s out of 100 on our risk score. The work is changing fast: more review, design and ownership of systems, less writing every line by hand.
Why do your risk score and Microsoft's score disagree?
They measure different things. Ours estimates how much of a job's O*NET task list could be automated, weighting the Frey & Osborne probabilities. Microsoft's measures how much people already use generative AI for the job's work activities. Both are shown so you can judge for yourself.
Related
Sources and method
- Risk score: ReplacedByAI's 0–100 score from O*NET task, skill and work-activity data, weighted 40% on the Frey & Osborne computerisation probabilities. Full methodology.
- AI applicability: Kiran Tomlinson, Sonia Jaffe, Will Wang, Scott Counts and Siddharth Suri, “Working with AI: Measuring the Applicability of Generative AI to Occupations” (Microsoft, 2025). Data from microsoft/working-with-ai, licensed CC BY 4.0. Percentiles are our calculation, among its 785occupations. Joined on the 6-digit SOC code, so O*NET specialties share their group's score.
- Pay: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, national annual median wage, May 2025.
- Filters: Microsoft percentile 50 or lower and our risk score under 60 (not applied to tech; see above). “All other” catch-all occupations are excluded.