What people actually search when they worry about AI and their career

Original data

We ran a keyword study in July 2026 to work out what to write about. The output was more interesting than the plan it produced, because the pattern in the numbers says something about how people are currently thinking about AI and work.

Here is the data, then what we think it means, then the reasons to be careful with it.

The numbers

All figures below come from a single DataForSEO run in July 2026, reporting estimated monthly Google search volume for the United States unless marked otherwise.

Search term	Est. monthly searches (US)	Competition
ai career	3,600	not recorded
ai internship	2,400	LOW
will ai take my job	1,900	not recorded
ai proof career	1,900	index 8 of 100
AI研修 (Japan)	1,900	HIGH
AI人材 (Japan)	880	not recorded

Two things jump out of that table before any interpretation: the size of the anxiety terms relative to the aspiration terms, and the competition column.

Reading one: anxiety and action are searched at the same volume

"Will ai take my job" and "ai proof career" each return about 1,900 monthly searches. That is a near-perfect split between the question and the response to it.

This is worth sitting with, because it contradicts the usual assumption that people are mostly doom-scrolling. The same order of magnitude of people are typing a phrase that only makes sense if they have already accepted the premise and moved to the practical question. Someone searching "ai proof career" is not asking whether the problem is real. They are asking what to do about it.

The interesting search is not the one asking whether AI will take the job. It is the one that has stopped asking.

The aspiration term above both of them, "ai career" at 3,600, suggests the same thing from the other direction: more people are searching for a way in than for reassurance.

Reading two: the competition column is the actual finding

Search volume tells you demand exists. The competition index tells you whether anyone has built for it.

"Ai proof career" came back at 8 out of 100. That is close to nobody. For a term with 1,900 monthly searches expressing a clear, urgent, actionable intent, an index of 8 means the query is being served by content written for something else.

"Ai internship" at 2,400 monthly searches with LOW competition says something similar. The term is specific, it has commercial intent, and the field is not crowded.

The gap between "people are searching this a lot" and "the results are mostly incidental" is where the useful work sits. It is also why our first blog post was a plain definition of what an AI internship is: the query had volume, and the existing answers to it were vague.

Reading three: Japan is a different shape

The Japanese figures invert the pattern. AI研修 draws about 1,900 monthly searches with HIGH competition, while AI人材 draws about 880.

That reads as a more mature training market with more established players buying the obvious term, and a smaller overall volume. The practical consequence is that a Japanese-language landing page competing on AI研修 is fighting a crowded auction, while long-form content answering specific questions has more room. We plan the Japanese content that way as a result.

What this data is not

Being specific about the limits, because keyword estimates get quoted as though they were measurements.

These are estimates, not counts. DataForSEO reports modelled monthly search volume derived from Google Ads data. It is a planning input. It is not a measurement of how many humans typed a phrase.

It is one run, one month. July 2026. Search volumes move seasonally and with news cycles, and a term tied to AI anxiety is more exposed to news cycles than most. A single snapshot cannot show a trend.

Competition indices measure advertisers, not content. A LOW competition score means few people are bidding on the term in paid search. It is a reasonable proxy for how contested a space is, but it is a proxy, and an organic result page can be crowded while the ad auction is quiet.

US volumes do not describe the world. The English figures are US-scoped. The Japanese figures are Japan-scoped. Neither generalises.

We have an interest in this. We run an AI internship programme, and we published this because the data supported a content plan we were already inclined toward. Read it accordingly, and check the terms yourself if a decision depends on them.

Why we published it

Most keyword research stays internal because it reads as a competitive asset. In this case the finding is not a secret worth keeping. It is that a large number of people are asking a practical question about their own working lives, and the answers available to them are thin.

That is a description of an underserved audience, not a moat. We would rather more people wrote for it properly.

---

→ Start with the definitional piece: what is an AI internship. Or go straight to the free four-week programme.