■ MODERATE RISK ■ Sports & Entertainment
Not fully, and the industry keeps rediscovering why: audiences can smell fake danger. CGI and digital doubles absorb the lethal and the impossible, while practical action — and the performers who survive it — has become a marketing point.
“CGI is replacing some stunts. Your bruises are now 'authentic filmmaking.'”
Our AI replacement risk score — how we score jobs
Stunt work is a craft profession with a union card and an actuary's nightmare of a job description. Performers rehearse and execute fights, falls, burns, car work, wire gags, and horse work; many specialize. Days are mostly preparation — rigging tests, pad placement, fight choreography drilled until it's muscle memory, safety meetings — because the visible three seconds of mayhem rests on hours of engineering. Stunt coordinators and second-unit directors, the career's senior rungs, design sequences, budget them, and own the safety plan. Increasingly the job also includes performance-capture suits and cyber-scanning sessions, because every major production now digitizes its stunt team.
The digital substitution is real and selective. Face replacement lets a double truly vanish into the star; digital doubles handle falls no human survives; crowd duplication ended the era of hiring forty guys to tumble down a hill. AI has accelerated all of it — cheaper face swaps, smarter physics simulation, generative previs that designs action sequences before anyone rigs a wire. The union fights over scanning rights exist precisely because studios can now reuse a performer's digital likeness, and background-level stunt work is the first casualty.
Yet the market keeps voting for practical. Blockbuster franchises openly advertise real stunts as a differentiator, because simulated physics still reads subtly wrong at the edge of danger and audiences respond to authenticity. More structurally, someone must still perform the motion — mocap needs elite movers — and someone must coordinate safety on any set with physical action. The craft is migrating upward: fewer bodies hitting pads, more specialists designing, capturing, and coordinating action. Our risk score reflects genuine displacement at the bottom and durability at the skilled core.
Automatability: our editorial assessment of current and near-term AI capability
Displacement is already underway: crowd work and extreme gags went digital years ago, and AI-driven face replacement and simulation keep lowering the substitution cost through the late 2020s. The 2030s bring contract fights over likeness scanning to a head and shrink entry-level stunt days further. By ~2040 the profession is smaller and more technical — coordinators, riggers, mocap athletes, and marquee practical performers — but decisively not gone.
Partially and selectively. Digital doubles take the fatal-if-failed gags, face replacement erases the double's identity, and crowd-scale stunt work has largely gone digital. But major productions increasingly market practical stunts as authenticity, mocap still requires elite physical performers, and every physical sequence needs human coordination and rigging. The bottom of the call sheet shrinks; the skilled core persists.
It's the profession's live contractual battleground. Productions routinely scan stunt performers, and AI makes reusing those scans cheap — potentially replacing future workdays with a past scan. Union agreements have been adding consent and compensation requirements around digital replicas, and performers should treat scan-session terms as seriously as injury waivers before stepping into the rig.
Realistic but narrower at the entry level, since background and crowd stunt days — the traditional apprenticeship — are the most digitized. Entrants now need standout specialties (martial arts, driving, gymnastics), mocap-quality movement, and early attention to rigging and safety skills that lead to coordination. Live shows and theme parks offer steadier practical work than film alone.
Because audiences can tell. Simulated bodies at the edge of physics still read subtly wrong, and franchises built on visible practical action have turned real danger into a selling point. There's also economics: for many mid-scale gags, a skilled performer is cheaper and faster than photoreal simulation. Practical and digital have settled into a hybrid, not a handover.