Frankfurt: AI at the security lane and aircraft stand
Frankfurt has placed AI-assisted prohibited-item detection into regular security screening and is expanding camera-based turnaround monitoring with Lufthansa. Their operating claims are substantial; measured airport results remain much thinner.
Frankfurt Airport Terminal 1 departures entrance, September 2023. Photograph: Quintin Soloviev / CC BY-SA 4.0 (cropped to fit). Shows the airport, not an AI implementation.
The picture at Frankfurt
Two decisions that cannot wait
The checkpoint and the aircraft stand have little in common until something goes wrong. A suspicious item needs prompt inspection; a missed ground-handling event can compromise the next departure. Frankfurt is using image analysis to provide earlier or more consistent signals in both settings.
The maturity is not identical. Fraport says its APIDS security software has been in regular operation on more than 50 CT scanners since January 2026. The seer turnaround system started at selected Lufthansa stands, with a larger rollout announced. A separate predictive-arrival data feed offers another way to plan resources, but its current performance is less visible.
All three are cases of machine analysis supporting human operational decisions. None should be described as eliminating the judgement of security staff, handlers or controllers.
Selected implementations
What the operational systems actually do
01 / Deployed
A second examination of the security image
The operational problem
A security officer has to make a reliable judgement from a complex CT image without letting the checkpoint become a bottleneck.
What the sources report
Fraport said that APIDS automated prohibited-item detection entered regular service on more than 50 Smiths Detection CT scanners across Frankfurt's terminals on 22 January 2026. AI software flags suspicious objects; security personnel retain the final decision.
Why it matters
This is a defined production use at an airport checkpoint, rather than a proposed concept. The worthwhile operating measures would include missed detections, false alarms, secondary checks and passenger throughput.
What remains unproven
Fraport has not published detection performance or a measured change in lane throughput attributable to APIDS.
Giving the stand the same timeline as the control room
The operational problem
Ground handlers, airport operations and the airline can disagree about when a turnaround task began or finished when they rely on different event records.
What the sources report
Fraport, Lufthansa and zeroG introduced seer, which uses cameras to timestamp ground-handling events. Fraport said five stands were covered in July 2025, with 20 planned by that year's third quarter and broader airport rollout envisaged.
Why it matters
A shared event record may help teams identify a missed handover and act sooner. The point is reliable information shared across organisations, not a claim that prediction alone removes delays.
What remains unproven
A later supplier-side figure of 55 stands has not been independently confirmed by Fraport or Lufthansa; a measured punctuality impact has not been published.
The ground operation needs to know when aircraft will land, not simply their scheduled arrival times.
What the sources report
Fraport announced in 2019 that it would use FlightAware's predicted landing times through the Firehose feed. FlightAware says the estimates draw on machine-learning models and can inform resource planning.
Why it matters
An accurate landing estimate would let teams position resources ahead of aircraft arrival. Unlike the 2026 checkpoint deployment, this record relies heavily on older reporting.
What remains unproven
Neither current forecast accuracy at Frankfurt nor a measured effect on turnaround performance is established.
The wider register also contains a large contract for passenger security scanners using AI-based analysis and a SITA self-service kiosk programme. Contracted equipment should not be treated as fully delivered, and a biometric-capable kiosk is not proof that AI identification is used.
Fraport contracted Rohde & Schwarz in November 2024 to install 100 QPS201 security scanners at Frankfurt Airport, for existing checkpoint lanes and for checkpoints still to open, with delivery over three years. Fraport's release of 6 November 2024 says the scanner uses millimetre-wave technology and AI-based algorithms to detect prohibited items while keeping false alarms low; the same release's subheading refers to "over 100" scanners.
Image analysis is almost entirely automated, Rohde & Schwarz says, so security staff only check the areas that have actually triggered an alarm. On 30 April 2026 the company said it had installed 17 QPS201 units and two QPS Walk2000 walk-through scanners in Frankfurt's Terminal 3, which opened on 22 April 2026.
Neither company has said whether the Terminal 3 scanners count towards the 100. The record stays at contracted until the wider delivery is shown to be complete, and no throughput or alarm figures have been published.
SITA TS6 self-service kiosks and baggage messaging at Frankfurt Terminals 1 and 2 · Deployed
SITA announced on 18 November 2021 that Frankfurt Airport had selected 87 of its TS6 kiosks for Terminals 1 and 2, running on the SITA Flex common-use platform, together with SITA's baggage messaging service. SITA called it its largest implementation in Europe and expected installation to finish later that year.
The kiosks let passengers check in and print bag tags for later self-service bag drop. SITA describes them as biometric-enabled, but the sources do not show biometric identification in use at Frankfurt's kiosks.
Dr Pierre-Dominique Prümm, Fraport's executive director for aviation and infrastructure, was quoted in the release on offering passengers innovative, safer and smarter ways to travel. No later source confirms that installation was completed, so the record rests on the 2021 announcement, and neither company describes the kiosks as AI.
Evidence supports a reported claim; it does not automatically establish a measured operating benefit.
From a detected signal to a defensible intervention
At the checkpoint, the security officer still decides whether an item warrants action. On the stand, staff must use the event timeline to manage an aircraft turn. Forecast landing times only help if they change where teams and equipment are positioned.
Those are measurable operational decisions. Public results for them would be more informative than another announcement of scanner counts or camera coverage.
Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.
Aerostack register · 5 tracked deployments
Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.
.png?width=1600)