The Confident Readback

The clearance arrived while both pilots were heads-down, so attention was divided and the auditory input was not fully encoded. The similar callsign and parallel departure created expectation bias: the crew expected a clearance consistent with the other aircraft's route, and expectation shapes perception. The readback was fluent and confident, which is precisely the problem: a readback can be a repetition of what the pilot believes was heard rather than a check of what was said. The controller's 'correct' closed the loop but confirmed only the crew's version, not the actual clearance. This is a breakdown at the perception and encoding stage, compounded by a confirmation loop that added no independent verification.

The crew should have established a sterile flight deck for the clearance, stopped configuring, and had the pilot flying listen and cross-check the readback rather than both being heads-down. If any element is uncertain, the correct action is to ask for confirmation, not to accept the fluent version. The controller should avoid 'correct' as a blanket acknowledgement and instead read back the critical elements, especially level and heading. Where callsigns are similar, controllers should use full callsigns and crews should be alert to the risk. Both sides share responsibility for closing the loop on the actual content.

Confidence is not evidence of accuracy. Both pilots were certain, and their certainty came from a coherent but wrong reconstruction of the clearance. In high-workload, high-similarity environments, the mind fills gaps with plausible content, and the resulting memory feels as solid as a true one. The practical rule is to distrust fluent recall when the conditions for error were present: divided attention, similar callsigns, and parallel traffic. Crews should build in an independent check, such as the pilot monitoring reading the clearance from the datalink or the pilot flying confirming the level against the flight plan, rather than relying on memory.