2025: Aviation Deaths are Down 27.3%, But Things Didn't Feel Safer

As the final reports for 2025 are filed at NTSB headquarters, a contradiction emerges from the data.

At first glance, the numbers for the year just passed are a triumph of engineering and oversight: total recorded incidents dropped by nearly 10%, and fatal accidents decreased by over 11%. Most significantly, the number of lives lost plummeted from 763 in 2024 to 555 in 2025: a staggering 27.3% reduction in lethality.

However, beneath these encouraging drops in the death toll lies a statistic that should give every regulator pause: injuries are up.

While the number of deaths decreased, the number of injuries rose by 4.2%. In the clinical language of aviation safety, this represents "injury conversion," where plane crashes seem to be more survivable.

This progress shows that although our machines are getting better at protecting us, our systems are still failing to keep us out of harm's way in the first place.

This table was aggregated from the NTSB's incidents/accidents webpage:

The most striking conclusion from 2025 is that we've entered an era where the aircraft airframe has become a fortress. We see this in the "injury-conversion" successes that defined the year.

Consider the crash-landing of Delta Connection Flight 4819 in Toronto earlier this year. The hull was destroyed, a scene that a decade ago would have almost certainly resulted in a "black box" investigation for a mass-casualty event. Instead, the cabin stayed intact long enough for the crew to execute a textbook evacuation. The result? A surge in the "injury" column, but a victory in the "survivable" column. This is the fruit of decades of research into fire-retardant materials and seat-track integrity.

However, 2025 also exposed the fraying edges of our air traffic infrastructure. The Potomac River mid-air collision in January, involving a commercial CRJ700 and a military Black Hawk, was a reminder that all the safety engineering in the world can't save a plane from a collision it never saw coming.

With 67 lives lost in that single evening, it became the grim anchor of the 2025 death toll. The investigation pointed to a variety of ATC staffing shortages and communication dead zones in busy corridors.

It is an opinion widely shared in the industry that we are over-relying on pilots to "see and avoid" because our ground-based systems are reaching their breaking point.

June 2025 brought another sobering milestone: the first fatal total loss of a Boeing 787 Dreamliner. Air India Flight 171’s tragic crash in Ahmedabad ended the aircraft's decade-long streak of perfect safety. While investigators are still parsing the data, the event sent a shockwave through an industry that had perhaps become too comfortable with the idea of "ultra-safe" modern jets.

The 1,517 incidents of 2025 tell us that we are still tempting fate too often. We have successfully mitigated the outcomes of accidents (better fire suppression, training, software, etc), but we have not yet mastered the prevention of the human errors that lead to them.

The drop in fatalities is a credit to the engineers at these aircraft manufacturers, but the persistence of mid-air close calls and runway incursions is a failure of the regulatory bodies that manage our controllers.

As we enter 2026, we can't expect the machines to do all the work. We must address the ATC fatigue, the pilot training gaps, and the aging radar systems that allowed 555 people to never make it home this year. We are getting better at surviving crashes; now we just need to work harder to prevent them from happening in the first place.