I teach the people this argument is about, and that's most of what I bring to it. Every September a room fills up with students who'll be looking for a first marketing job in a year or two, and every year the public conversation about whether that job will exist gets louder without getting more settled.
The strongest evidence for the pessimistic case arrived this month. Stanford's Digital Economy Lab, working from ADP payroll records, finds employment for workers aged 22 to 25 in the most AI-exposed occupations now sits about 19 percent below where it would be had it tracked same-age workers in less-exposed jobs. That gap was 15 percent in their July 2025 version. Between November 2022 and June 2026, the two most-exposed quintiles lost roughly 11 percent of their young workers while the three least-exposed grew about 10 percent. Experienced workers show no comparable gap. The authors are careful to say this isn't economy-wide displacement, and I'd repeat that carefully too.
The strongest evidence the other way came at the end of June. Ramp's economics lab, with Revelio Labs, looked at more than 21,000 US firms and found that heavy AI adopters grew entry-level headcount 12 percent over the two years after adoption, and raised the entry-level share of their workforce. The authors say plainly that their adopters were already growing faster before they adopted, and were three times as likely to be venture-backed. So it isn't clean. It does suggest that firms buying a lot of AI are not, as a rule, the ones cutting juniors.
Marketing has its own numbers between those two. The American Marketing Association's 2026 careers report, from 1,412 professionals, puts marketing employment 27 percent below pre-pandemic levels. Content marketer roles fell 11 percent from 2024 to 2025. SEO specialist roles fell 15 percent, the steepest decline of any role they track. Influencer marketing grew 18 percent. The share of marketing postings mentioning AI doubled in 2025.
What I keep noticing is which roles are losing ground. They're the ones where the work was mostly production. Write the post, build the page. The roles holding are the ones built on a relationship or a judgment call, which are harder to specify and therefore harder to hand over.
That has an uncomfortable implication for teaching. Production work is how people used to learn. You wrote forty bad headlines and by the fortieth you understood something about headlines that nobody could have told you directly. If a machine writes the forty, the learning has to come from somewhere else, and I don't think we've worked out where.
My own attempt is to make students do the editing rather than the generating. Bring me the machine's forty, tell me which three are any good, and tell me why. It's a different skill and harder to fake. It also assumes a kind of taste the student hasn't built yet, so I'm not certain the sequence works. I'm running it anyway, because the alternative is pretending the tools aren't in the room.
On the Canadian side there's very little to go on. Statistics Canada put youth unemployment at 12.6 percent in July, down 1.9 points year over year, which is a better number than most people expect. But there's no Canadian equivalent of the AMA report and no occupation-level series that breaks out marketing. Anyone telling you what's happening to junior marketing jobs in this country is inferring it from American data. I'm doing the same thing, and it seems worth saying out loud.
What I tell students is narrower than what the headlines promise. The first job will likely require you to be useful sooner than my generation had to be. The way to be useful sooner is to be the person who can tell whether the output is right, which means knowing the subject rather than the tool. That's less satisfying than a prediction. It's also the part I'm confident about.
Sources
- Canaries in the Coal Mine, August 2026 update (Stanford Digital Economy Lab): digitaleconomy.stanford.edu
- Heavy AI adopters hire more (Ramp Economics Lab with Revelio Labs): ramp.com
- 2026 State of Marketing Careers (American Marketing Association): ama.org
- Labour Force Survey, July 2026 (Statistics Canada): statcan.gc.ca