The government is weighing a proposal to make drug testing mandatory for every pilot once a year, a sharp jump from the current rule requiring only 10% of the overall pilot network to be sampled.
Civil Aviation Minister Ram Mohan Naidu confirmed the shift under consideration, saying, “Right now, the rule says 10 per cent across the overall pilot network,” as the DGCA holds discussions with stakeholders on the operational implications of full coverage.
The trigger was a single incident: Air India flight AI2379 lost around 300 feet of altitude after a hydraulic failure on August 4, injuring 17 people, after which the pilot-in-command’s confirmatory test came back positive for marijuana.
That single data point cascaded quickly, Air India expanded internal screening to roughly 400 pilots, found two more failed preliminary tests, and the regulator is now reconsidering a testing framework that had stood at 10% sampling for years.
This is a familiar pattern in compliance-heavy industries: sampling-based testing regimes are often defended as statistically sufficient right up until one incident makes the residual risk visible and impossible to ignore, at which point the bar for what counts as ‘sufficient’ shifts overnight.
For founders building anything with a safety or compliance surface, whether it’s a logistics fleet, a healthcare product, or a fintech platform, the underlying lesson holds: a sampling rate that looks statistically defensible on a spreadsheet can still be the wrong call operationally, and it often only gets tested by the one case regulators, or customers, weren’t sampling for.
The incident triggered concerns after a subsequent confirmatory drug test of the pilot-in-command reportedly returned positive for marijuana.
Air India subsequently expanded drug screening across its pilot workforce, with two more pilots later found to have failed preliminary drug tests and being removed from flight duties as a precaution.
Photo by Chad Davis, Wikimedia Commons, CC BY 2.0