Search marketing (SEO, SEA, acquisition on search engines) is among the occupations the public would most readily hand over to AI, according to a Harvard Business School working paper reported on 7 September 2026 by Search Engine Journal. On a 1 to 7 scale measuring moral objection to automating a job, the search marketing strategist scores 2.31, which places it 938th out of 940 occupations tested. The study also shows what protects a job: the perceived competence of the human on a given task, far more than any principle.
The facts
- In the table of 940 occupations of the working paper "Performance or Principle" by Simon Friis and James W. Riley (Harvard Business School, no. 26-017, October 2025), the occupation "search marketing strategist" receives a moral objection score of 2.31 out of 7, ranked 938th, just ahead of file clerks (2.17, last place). At the other end, clergy (5.91) and childcare workers (5.86) remain the occupations the public most refuses to see automated (working paper on SSRN).
- The survey questioned 2,357 US adults, recruited on the Prolific panel with quotas matched to the census (age, gender, ethnicity, political affiliation), for 23,570 ratings in total. With current AI capabilities, the public supports fully automating 30% of occupations; when AI is described as outperforming a human at lower cost, that support rises to 58%. Only 12% of occupations remain categorically off-limits whatever the machine's performance: caregiving, therapy, and moral or spiritual authority (working paper on SSRN).
- AI as a tool that assists human work, without replacing it, draws 94% approval with current capabilities and nearly 96% with a more advanced AI, according to the presentation of the study published on 5 February 2026 by HBS Working Knowledge. The authors point to one limitation: respondents were not asked about the automation of their own occupation.
What protects a job from automation?
James Riley and Simon Friis distinguish two forms of resistance to a technology: one that rests on performance and fades as the tool improves, and one that rests on principle and persists whatever the machine can do. For the large majority of the 940 occupations tested, search marketing included, resistance belongs to the first category: the public judges the ability of AI to match a human on a given task, it does not defend the idea that the task must remain human. A study by Elisabeth Paulson (Harvard Business School) and Kirk Bansak (UC Berkeley), published in December 2024 in PNAS Nexus with 9,000 participants, points the same way for bank loan and pretrial release decisions: at equal performance, the preference for a human stays modest (4.3 points for the loan, 7.6 points for release), and 54% to 56% of respondents convinced that the algorithm does better choose it (PNAS Nexus).
Riley gives leaders four recommendations in HBS Working Knowledge: concentrate short-term deployments where the public does not object and where AI is already solid; build social norms into long-term planning, beyond technical constraints, with talent retention and customer reactions in mind; stay transparent about the actual use of AI; and know the lines some customers refuse to see crossed, because an automated task can devalue a service in their eyes, regardless of its performance.
Greg Jarboe, author of the Search Engine Journal analysis, draws three rules for search marketing practitioners. Put a real, checkable name behind any work AI touches, with a profile and a track record anyone can consult, rather than a generic editorial team byline. Publish a performance record alongside the method, so that competence is observed instead of claimed. Reserve full automation for repetitive tasks with nothing at stake, such as link audits or draft meta descriptions, and keep an identified human on everything that touches a reader's trust or a client's money. The question thus moves from the diploma to the proof.
AIxH's view
That is the role we take on for every assignment of our SEO agency in Luxembourg and every campaign of our digital advertising agency in Luxembourg: AI helps us go further on creative work, content and analysis, and a named consultant remains responsible for understanding the client's challenges, budget and regulatory constraints, and for the arbitration between channels. That reading of the context protects the relationship far better than an automated dashboard. Dependence on a single algorithm remains the risk a human keeps under control. If your AI-driven management is moving ahead without that supervision, an audit quickly measures the gap.