What is changing
The U.S. Bureau of Labor Statistics projects computer and information technology jobs as a group to grow faster than average from 2025 to 2035. Within that group, routine work is shrinking: computer programmers are projected to decline 7% and computer support specialists 3%. Software developers are projected to grow 10%, information security analysts 21% and data scientists 35%.
The engineering work itself is moving. Writing code from scratch matters less, and connecting AI into the way a company works matters more: linking AI agents to the systems a company already runs, redesigning the workflow so AI handles the routine steps and people handle the exceptions, and building it to grow without breaking.
What AI changes, and what stays yours
| Role | What AI changes | What stays yours |
|---|---|---|
| Software developer | First drafts of code, tests, documentation, explaining old code | The design, the review, what goes into production and why |
| Quality assurance analyst or tester | Test cases, test data, bug report drafts, running routine tests | Judging what matters to users, finding the bug nobody thought to test for |
| Help desk or support specialist | Password resets, routine questions, ticket sorting, knowledge articles | The hard problem, the frustrated user, spotting the pattern behind the tickets |
| Network or systems administrator | Log reviews, scripts, change plans, status reports | Every change to live systems, access decisions, the call during an outage |
| Data analyst or data scientist | Cleaning data, first-pass analysis, charts, queries | Asking the right question, judging what the numbers mean, telling the story honestly |
| Information security analyst | Sorting alerts, summarizing threats, drafting reports | Deciding what is a real attack, the response, the escalation |
| Web developer | Page layouts, code, testing across devices, accessibility checks | What the client needs, the quality of the finished site |
| Drafter or product designer | Drawing revisions, design variations, documentation | Fit, function, safety and the client's intent |
Writing routine code from a clear specification, answering password and printer tickets, building simple websites by hand, redrawing the same part with small changes and writing first-draft documentation are shrinking quickly. Move toward the work around them: understanding what users need, reviewing and testing what AI writes, keeping systems safe and running, and explaining technology to the people who use it. Become the person who teaches the chatbot and reviews the AI's code, as Kwame did.
The Decision Line
AI can write the code, the script and the change plan. A person decides what goes into production, who gets access to which systems, what data is collected and kept, and how to respond when something goes wrong, because those decisions carry the company's security and its customers' trust. Passwords, keys and customer data never go into a tool your employer has not approved.
Before a change to a live system, ask your AI assistant to list everything the change could affect, the steps to undo it and the questions a careful reviewer would ask. You still approve the change, with the risks laid out in front of you.
Try a request for your role
Pick the role closest to yours. Copy the request into any AI assistant you are allowed to use, fill in the parts in brackets with your own material, and leave out anything private.
Code, credentials, logs and customer data go only into tools your employer has approved. Check every fact, number and date before you use the answer. Steps that put things on your calendar may need a paid plan or a linked calendar; if you do not have one, add them yourself from the list. The check-in line keeps you in control: Before you start, ask me any questions you have. Then tell me in a few lines how you plan to do this, and wait for my OK.
Where this work can grow
For most people in technology, the clearest step is to take on the review, testing and teaching that AI tools now need, in the job you already have. Beyond that, three directions are growing fast.
Information security analyst
Projected to grow 21% from 2025 to 2035. It suits help desk staff and system administrators who already know how systems fail and how people get fooled.
Data analyst and data scientist
Data scientists are projected to grow 35%. Analysts who can ask good questions of data, and explain the answers in plain words, are needed in every field.
AI integration, automation and quality
Companies need engineers who can connect AI agents to the systems they already run, and people who can test what the agents produce and watch for when something goes wrong. These roles go by many titles and have no single job count yet, so build toward them while you earn.
Sources
- U.S. Bureau of Labor Statistics, Computer and Information Technology Occupations: bls.gov
- U.S. Bureau of Labor Statistics, Computer Programmers: bls.gov
- U.S. Bureau of Labor Statistics, Software Developers, Quality Assurance Analysts, and Testers: bls.gov
- U.S. Bureau of Labor Statistics, Computer Support Specialists: bls.gov
- U.S. Bureau of Labor Statistics, Information Security Analysts: bls.gov
- U.S. Bureau of Labor Statistics, Data Scientists: bls.gov
- Stack Overflow, 2025 Developer Survey, AI section: survey.stackoverflow.co