Avionics
Archer AI To Transcribe ATC Clearances in ForeFlight
On-device transcription requires no connection in flight
Archer’s ZEE AI Foundation Model turns ATC instructions into readable text in ForeFlight Mobile.
Archer’s ZEE AI Foundation Model turns ATC instructions into readable text in ForeFlight Mobile.

Pilots using ForeFlight Mobile for the Apple iPad will be able to check an ATC clearance against on-screen text instead of memory under the first phase of an integration that brings eVTOL developer Archer Aviation’s Zee aviation foundation model to the electronic flight bag (EFB) app. The transcription is designed to run on the iPad and requires no connectivity.

Zee converts each radio call into text as it arrives, giving pilots a log they can return to later to verify what was said. Archer’s in-house AI team built the model and trained it on more than 100,000 hours of recorded ATC radio traffic. The company describes it as the first foundation model able to run on-device, without a connection, at the accuracy level operations demand.

“Every pilot knows the moment when you read back a clearance and wonder whether you heard it right,” said Archer founder and CEO Adam Goldstein. “We built Zee to change that.”

Archer and Jeppesen ForeFlight said breakdowns in VHF communications exchanges between pilots and controllers contribute to most of the incidents pilots file to the NASA Aviation Safety Reporting System. Further, they said, pilot readback error rates top 25% on transmissions carrying four or more instructions.

Jeppesen ForeFlight is now measuring how the model performs on actual radio recordings. The companies expect to share more details early next year.