SAFE
The product is a human body competing in real time in front of people who care specifically that it is a human body — there is no screen-work core for AI to eat. Physical performance, in-game reads, and audience connection are the entire job; the only AI-touched pieces are film study, scouting reports, and training-load analytics, which are support functions handled by staff. The real career risks here are injury, roster math, and a national median that hides a brutal pay pyramid — not automation.
Tasks largely resist digitisation. Sprinting a 4.4 forty, absorbing contact, and executing a play call under a shot clock cannot be tokenized — the 19 rather than 20 reflects that opponent-tendency film review and periodized training plans, which used to be your own homework, now arrive pre-chewed from an analytics department.
Hands-on in uncontrolled environments. The job is performed in stadiums, on ice, in mud, in heat, against opponents actively trying to physically stop you — there is no version of this executed remotely, and injury risk is an occupational condition rather than an exception.
No licence, no signature requirement. There is no licence to compete; league eligibility, a contract, and a drug-testing consent form are the gate, and a team can waive you Tuesday morning with no credential to fall back on.
The human relationship is the product. Fans, sponsors, and jersey buyers pay for a specific named person's identity and story — a Curry stepback and a G-League stepback are the same physics with wildly different revenue — but the 18 accounts for the fact that in team sports the franchise brand carries a share of the loyalty you would take with you.
Meaningful discretion. You own live reads — when to shoot, when to concede the bunt, when to say the hamstring is done — but coaching staffs set the game plan, playbooks constrain the option set, and consequences of a bad call are a loss rather than someone's ruin, which caps this well below the licensed-professional tier.
Has AI actually changed your work?