Free sampler. Re-scored every six months.
Engineering scores between 3.4 and 6.3 out of 10 for AI exposure depending on the discipline, where 10 is most at risk. Civil and structural engineering with a PE license and site work scores 3.4. Desk-based design and analysis scores 6.3. Engineering is the instructive contrast to computer science — the same technical aptitude, pointed at the physical world, picks up three protections software does not have.
The short answer for parents: yes — and for a technically-minded student, the engineering-versus-computer-science comparison deserves far more thought than it usually gets. Software pays more at entry, where entry survives. Engineering holds licensure, site work, and legal accountability for things that must not fail.
Every career in this index is scored 1–10, where 10 is most exposed to AI. Same six factors, same weights, applied identically to an engineer and a paramedic.
| Engineering track | 2023 | 2025 | Now | 3-yr move | Band |
|---|---|---|---|---|---|
| Civil / structural (PE, site-based) | 3.4 | 3.7 | 4.0 | +0.6 | Low–Mod |
| Electrical power & grid | 3.3 | 3.6 | 3.9 | +0.6 | Low–Mod |
| Mechanical (industrial) | 3.9 | 4.1 | 4.4 | +0.5 | Moderate |
| Manufacturing / process | 4.4 | 4.7 | 4.9 | +0.5 | Moderate |
| Design & analysis (desk, CAD-centered) | 5.2 | 5.8 | 6.3 | +1.1 | Mod–High |
2023 and 2025 figures are reconstructed using current methodology, not archived from past editions.
The comparison that matters most: entry-level software development scores 8.1, senior software architecture 5.4. Licensed site engineering scores 4.0.
Same kind of mind. Same math. Very different exposure — and the reason is structural rather than technical.
| Software | Engineering | |
|---|---|---|
| Regulatory protection | None — no license exists | Strong — the PE stamp is a legal monopoly |
| Physical requirement | None — the work is the medium of the threat | Real — sites, commissioning, inspection |
| Accountability for catastrophic failure | Commercial | Legal and personal |
| Entry route | Purely market-driven, and narrowing sharply | Licensure requires supervised experience years |
| Entry-level score | 8.1 | 4.0 licensed / 6.3 desk |
Software has two of our six protection factors sitting at essentially zero for everyone in the profession. Engineering has neither of those gaps at the licensed end.
That is the single most decision-relevant comparison in this index for a student who is good at math and likes building things.
Not the ones who sign, visit sites, or make calls where failure is catastrophic. It is taking a substantial share of the desk layer.
Generative design tools now produce and compare many options against constraints of cost, material and structure. Each of those tasks was, until recently, a junior engineer's assignment.
How civil and structural engineering rates against each. Ratings are 0–10 on each factor's own terms.
So you can see what the analysis actually looks like.
How often does the job hit genuinely new, high-stakes situations? — rated 8.0
AI is excellent at patterns and weak at genuine novelty. So this factor measures a specific thing: how much of the work involves situations that do not match anything the system has seen, where being wrong is expensive.
Engineering rates high, but not for the reason most people assume. It is not that the calculations are hard — calculations are exactly what machines do well, and standard-case analysis is among the fastest-automating parts of the profession.
It is that every site is different, and the consequences of being wrong are structural. A bridge sits on particular ground with particular water, in a particular seismic and thermal environment, subject to a particular jurisdiction's code, built by a particular contractor with particular constraints. The combination has no precedent even where every individual element does.
That is why the licensed, site-based track holds at 4.0 while desk-based analysis has moved to 6.3. The analysis is generalizable. The judgment about whether the analysis applies here is not.
One honest caution, and it is the profession's real problem. The PE license requires supervised experience years, which keeps the on-ramp legally open — the same mechanism that protects nursing and medicine. But if those years are increasingly spent supervising generated output rather than doing the analysis by hand, the question becomes whether they still build the judgment the license is meant to certify. The rung exists. What it teaches is changing, and the profession has not really debated this yet.
Ranked by exposure, safest first:
Electrical engineering pointed at power and infrastructure may be the single most demand-secure choice in this entire index. The AI buildout is itself the demand — data centers, grid capacity, electrification.
This sampler tells you where engineering stands. The full profile tells you what to do about it.
Most families are weighing two or three careers seriously, and a few more they haven’t ruled out. Pick the ones you need.
Each includes a short version written directly to the student and the technical scoring appendix. Spring 2027 updates of whatever you buy are included.
Read one complete profile free → We publish computer science in full so you can judge the depth before buying anything.