Which job will AI change the least, against all expectations?

[Dan]

Most conversations about AI and jobs focus on which roles will be replaced or transformed. We have covered plenty of those stories here, from desk jobs to coding to customer service. But I am curious about the opposite question. Which jobs will AI change surprisingly little, even though people expect it to?

Some answers seem obvious. Plumbers, electricians, carpenters and gardeners work in physical, unpredictable environments where robots still struggle. Care work relies heavily on human contact and trust. Hairdressers, tattoo artists and chefs combine craft with personal relationships

But I am more interested in jobs that people assume AI will transform but probably will not. Maybe teaching, where the relationship with pupils matters more than delivering information. Perhaps sales, where trust and persuasion count for more than speed. Or roles like judges and referees, where accountability and legitimacy matter as much as accuracy

The Irish Times column we discussed argued that most office work involves far more human coordination than tech leaders assume. That suggests even some desk jobs could be more resilient than predicted. It is easy to overestimate how much of a job is the visible part. Office work may turn out to be stickier than predicted

Which job do you think will change least, and why? Bonus points for surprising answers