Will AI replace software engineers?

Software developers today: the official numbers

From the Bureau of Labor Statistics Occupational Outlook Handbook (updated August 27, 2026). BLS reports developers together with QA analysts and testers, and computer programmers separately:

	Software developers, QA analysts and testers	Computer programmers
Jobs in 2025	1,905,400	110,800
Median pay, May 2025	$134,040 a year (developers alone: $135,980)	$100,390 a year
Projected change, 2025 to 2035	+10% (185,400 jobs)	-7% (-8,100 jobs)
Openings a year, on average	About 106,100	About 4,400

All jobs together are projected to grow 3.5%.

For developers, BLS expects strong demand "due to the continued expansion of software development for artificial intelligence (AI), Internet of Things (IoT), robotics, and other automation applications," plus more spending on security software and more products that run on software.

For programmers, BLS names AI directly: "To save time and money, many companies are expected to leverage technologies, including artificial intelligence (AI), to automate repetitive programming tasks." It adds that some higher skilled programming tasks will likely shift to software developers.

A projection is an estimate from today's trends: the official best guess, not a promise.

What AI is actually being used for in software work

How much of the job shows up in AI use. Anthropic, which makes the AI model Claude, published in March 2026 a measure of how much of each job's work people actually do with Claude, which it calls observed exposure. It counts tasks AI could speed up that show real work use, with full weight when AI does the task and half when it helps, weighted by time spent.

Occupation	Observed exposure
Computer programmers	75%, the highest of any job
Software QA analysts and testers	52%
Web developers	48%
Software developers	29%

Across all computer and maths jobs, Anthropic found that Claude "covers just 33%" of tasks, against 94% that are theoretically possible. Coding also dominates use: computer and mathematical tasks were 35% of Claude.ai conversations in February 2026.

The difference between programmers and developers is the useful part. Writing and fixing code is heavily used. Working out what to build, with whom, and how it should behave is used much less.

What developers report. In Stack Overflow's 2025 Developer Survey (about 49,000 respondents), 84% were using or planning to use AI tools, and 51% of professional developers used them daily. Trust is low: 46% distrust the accuracy of AI output, against 33% who trust it.

The entry-level evidence

This is where the data is least comfortable, and it applies to software directly.

• Stanford's Digital Economy Lab found that "by September 2025, employment for software developers aged 22-25 declined nearly 20% compared to its peak in late 2022," while older developers held steady or grew.
• Its August 2026 update puts employment of 22 to 25 year olds across highly AI-exposed jobs about 19% below where it would be had it kept pace with less exposed peers. "Experienced workers show no comparable gap." The authors say these are "descriptive patterns, not causal estimates."
• Anthropic found young workers less likely to be hired into exposed jobs: a 14% drop in the job finding rate compared with 2022. It also found "no systematic increase in unemployment for highly exposed workers since late 2022."

So far the effect shows up as fewer junior hires, not mass layoffs of working engineers. Usage data measures how people use AI tools, not job losses, and none of these studies can say what happens next.

Which engineering tasks are exposed, and which are not

From the software developer task list on O*NET, the US Department of Labor's job database:

More exposed: code, data and documents

• Modifying existing software to fix errors or improve performance
• Storing, retrieving and manipulating data for analysis
• Analysing requirements and recommending changes to a system
• Writing reports and status updates on a project

Less exposed: design, standards and people

• Designing systems and weighing the trade-offs, with the scientific and mathematical reasoning behind them
• Setting performance standards, and owning testing and validation
• Working with analysts, engineers and customers to pin down limits and requirements
• Coordinating installation and making sure the system runs to spec
• Supervising and assigning work to other engineers

Is software engineering safe from AI?

It is one of the most exposed fields and also one of the fields projected to grow. Both are true. The pressure so far lands on junior roles and on pure coding work. Projections are estimates, and nobody knows the long-run answer for certain. To weigh your own role task by task, see will AI take my job.

What to lead with on your resume

Lead with decisions and outcomes, not lines of code.

Instead of	Lead with
Wrote features in React and Node	Designed and shipped the checkout rewrite, cutting failed payments from 2.1% to 0.8%
Fixed bugs	Owned on-call for the billing service and cut paging alerts by half in a quarter
Experience with AI coding tools	Used AI coding assistants to migrate 40 services, with tests and review on every change
Participated in code reviews	Set the review and testing standard for a team of 6, adopted across 3 teams

If you are early in your career, show the parts that read as judgement: a project where you chose the approach, tested it and explained why. Use your own true numbers. Then run the resume check on this page, which is free and needs no account: it reads your actual resume and shows how your experience comes across to an employer's software and to a recruiter. It does not predict anything about AI.