■ HIGH RISK ■ Public Service & Government
The document-checking, face-matching core of primary inspection is automating at every major border. Humans keep the interviews, the fraud instincts, and the hard calls — a smaller, more investigative version of the job.
“Biometric scanning is faster and less judgmental. Mostly.”
Our AI replacement risk score — how we score jobs
An immigration inspector at a port of entry runs primary inspection: verifying passports and visas, querying watchlist databases, asking the purpose-of-travel questions, reading nervous behavior, and deciding in under a minute whether to admit, question further, or send to secondary. Secondary inspection is the deeper version — extended interviews, document forensics, database cross-checks, and the paperwork of refusals, paroles, and expedited removals. It's high-volume decision-making where the base rate is 'fine' and the cost of the rare miss is a headline.
Automation has been eating the front of that funnel for a decade. E-gates with facial recognition now clear arriving travelers at hundreds of airports worldwide; biometric entry-exit systems match faces to passport chips more reliably than a tired human at hour seven of a shift; and risk-scoring runs on every traveler before the plane lands, using itinerary and database signals to pre-sort who deserves attention. Automated kiosks took the routine declarations years ago. The design goal, openly stated in border-modernization programs, is for machines to clear the compliant majority so officers concentrate on flagged cases. That's efficient — and it means fewer officers per million travelers, which is what our 65 score reflects. Volume growth in travel offsets some of it, but the ratio moves one way.
The resistant core is adversarial judgment. Document fraud evolves specifically to beat automated checks; interviews that unravel a rehearsed story, asylum-claim referrals, human-trafficking indicators, and the discretion the law explicitly grants officers all require people — legally and practically. Secondary inspection is growing more complex even as primary automates. And the political dimension cuts both ways: border staffing is a perennial budget priority, and no government wants headlines blaming an algorithm for admitting the wrong person. Expect fewer booth seats, more investigative and oversight roles, and a job that starts where the machines give up.
Automatability: our editorial assessment of current and near-term AI capability
E-gates and biometric matching are standard at major hubs today and expanding through every border-modernization budget. Through 2030, expect primary-inspection booths to keep converting into automated lanes with fewer officers overseeing more travelers, while secondary, fraud, and enforcement roles hold or grow. Government hiring cycles and travel-volume growth cushion the workforce; the underlying officers-per-traveler ratio declines steadily anyway.
At the routine end, visibly yes — facial-recognition gates now clear compliant travelers at major airports with officers supervising banks of lanes rather than staffing booths. But every flagged mismatch, suspect document, and complicated case still lands with a human, and secondary inspection is getting harder, not easier. The officer corps shifts from processing everyone to investigating the exceptions.
Safer than the automation trend suggests, because politics funds border staffing and travel volumes keep growing. Still, the long-run ratio of officers to travelers is falling as biometric lanes expand, and routine primary-inspection work is the part disappearing. Candidates should expect and pursue the investigative side — interviews, fraud, enforcement — where the career runway is longest.
No credible system does, and several tried. Automated deception-detection pilots have drawn heavy scientific criticism, and no deployed technology reliably reads intent. What AI does well is narrower: matching faces, spotting document anomalies, and risk-scoring travel patterns. The interview — where a trained officer pulls a thread until a story unravels — remains a human craft, and legally must stay one in most systems.
Move toward what the machines escalate: secondary interviewing, forged-document examination, trafficking and smuggling indicators, and asylum processing. Fluency in the biometric systems themselves is also valuable — agencies need officers who understand the tools' failure modes well enough to catch what they miss. Booth work is the shrinking tier; investigation and oversight are the growing ones.