👨‍🚀AI & The Future

Astronaut · The rarest job on Earth is practiced off it: roughly 700 people have ever held it, from Gagarin's 108 minutes to year-long tours on the space station.

This is the rare profession whose replacement is a seventy-year-old going concern: robotic spacecraft have outnumbered and out-traveled humans since the beginning, exploring every planet in the Solar System while people stayed within a few hundred kilometres of Earth — plus nine round trips to the Moon. Every budget cycle, someone argues the humans are the inefficiency.

The counter-argument is that the astronaut's product was never data alone. Crews exist to be there — as physiology experiments that cannot run on hardware, as repair capability no arm yet matches, as the reason the public funds the enterprise at all. Automation keeps eating the flying; it has yet to touch the presence.

26 / 100
Low

Share of the work a machine could do

Task by task, a meaningful slice of the job is already automated: ascent, docking and re-entry are machine-flown with humans as monitors, and robotic probes do planetary science at a fraction of crewed cost. What cannot be automated is the mission's actual product — human presence, human physiology as the experiment, and improvised repair — so the profession shrinks or grows with political ambition, not with algorithmic capability.

Scored from the tasks, not the job title. Lower is safer.

Jobs AI cannot take →

What machines cannot take

Presence is the product

96

A large share of human spaceflight's purpose — proving people can live off Earth, national prestige, inspiring the public that funds it — is definitionally unautomatable: an uncrewed program answers a different question.

The body is the experiment

88

Bone loss, vision changes, immune drift: the core research justifying the station is what space does to humans, and it cannot run without them. Every Mars plan depends on data only crews can generate.

Improvised repair

84

Apollo 13's rebuilt CO2 scrubber, the 1985 revival of the dead Salyut 7 station, five Hubble servicing missions — spaceflight's history turns on unscripted fixes no current robot could plan or perform.

Autonomy beyond the light lag

76

At Mars, radio takes up to twenty minutes each way; no ground team can joystick a crisis. Deep-space crews must carry mission control's judgment with them — more human decision authority, not less.

Command and crew judgment

68

Someone accountable must decide when to abort a spacewalk, deviate from the timeline or overrule the automation — Boeing's 2019 uncrewed Starliner flight went astray precisely where a crew might have intervened.

What they already take

Piloting: ascent, docking, re-entry

85

Soyuz's Kurs system has docked automatically since the 1980s, the Soviet Buran shuttle landed itself uncrewed in 1988, and Crew Dragon flies hands-off end to end. Manual flying is now the trained backup, not the job.

Planetary fieldwork

70

Rovers like Curiosity and Perseverance have done geology on Mars for over a decade, caching samples for return — slower per day than an Apollo geologist, but sustainable for years at a small fraction of crewed cost and risk.

External inspection and routine maintenance

55

Free-flying camera robots, arm-based surveys and planned humanoid systems (NASA flew Robonaut 2 to the ISS in 2011) keep absorbing survey and simple-swap work that once required a suit and eight hours of a crew's day.

Experiment tending and monitoring

50

A growing share of ISS payloads run autonomously or by ground command, with crews reduced to sample changes; station systems monitoring is largely a ground and software function already.

How the work is changing

From pilot to payload — and back

Automation turned the corps from stick-and-rudder pilots into scientist-operators over fifty years; deep-space missions now pull the other way, demanding crews who can run a spacecraft without real-time ground support for weeks.

Employers multiply

For sixty years every astronaut worked for a government; now Axiom employs retired NASA veterans as professional commanders, SpaceX's Polaris program fields private crews, and planned commercial stations advertise for crew — the profession's first genuine labor market.

The Moon becomes a workplace again

NASA's Artemis II crew — including Christina Koch and Victor Glover, the first woman and first Black astronaut assigned to a lunar mission — marks the return of expedition-class flying, with geology training and lunar surface operations reshaping selection priorities.

The eligible body widens

ESA selected parastronaut John McFall, a leg amputee, in 2022 and announced his medical clearance for a long-duration ISS mission in 2025; private crews have flown a cancer survivor with an internal prosthesis. The Right Stuff's physical template is being deliberately dismantled.

New jobs branching off

Commercial astronaut trainer

Preparing private crews and customers for orbital and suborbital flight — a role staffed heavily by former agency astronauts at companies like Axiom and SpaceX, and the first civilian market for astronaut skills.

Space flight surgeon

Physicians certifying, monitoring and rehabilitating crews, and setting the medical standards that decide who flies — a specialty growing as commercial passengers with ordinary medical histories reach orbit.

Robotic teleoperator

Controlling rovers and humanoid robots on planetary surfaces from orbit or Earth — ESA and NASA have run astronaut-in-orbit rover-driving experiments as a template for early Mars operations conducted from Martian orbit.

Private mission director

Designing, integrating and commanding commercial spaceflights — matching customers, vehicles, training and research payloads — a role invented in the 2020s as missions stopped being agency property.

Outlook

The honest projection is a split profession. The flying keeps automating: nobody hand-flies an ascent, and lunar landers will be more autonomous than Apollo's. What grows is the expedition side — long-duration crews on the Moon, eventually Mars — where light-lag and isolation demand more human judgment per person than any mission since 1972.

Headcount will stay tiny but stop shrinking. National corps are stable at a few dozen each, China's is expanding with its station, India is training its first Gaganyaan crews, and commercial stations would create the first demand for professional crew not set by parliaments. None of this makes the job common; it makes it slightly less singular.

The deepest risk to the profession was never robots — it is political will, the budget line that funds presence for its own sake. Automation has, if anything, protected astronauts: every task a machine absorbs makes the remaining human hours cheaper to justify. The last job to automate is being the reason the mission exists.

Keep exploring

More in Frontier & Field