The NTSB, in issuing its annual list of highest-priority safety improvements in the transportation sector, has included something that it points out has been suggested to the FAA for years—requiring cockpit video recorders, which they refer to as cockpit “image” recorders, in commercial airliners. The NTSB cited crashes in Texas (a Boeing 767, Atlas Air flight, which crashed near Houston two years ago), and Ethiopia, presumably Ethiopia Airlines Flight 302, a 737 MAX that crashed shortly after takeoff from Addis Ababa in 2019. They could have referenced a few others, as well.

As we all know, commercial aircraft are required to have both flight data recorders (FDRs)—which capture what’s happening with the airplane and its systems every fraction of a second—and cockpit voice recorders (CVRs), which, just as the name suggests, record what’s going on, noise-wise, in the cockpit at all times. In both cases, the technology is designed to record to media that is protected by incredibly strong shells, so that even after a crash, the data is most likely somewhere for investigators to find. And in most cases, they do.

In fact, in many investigations, FDRs and CVRs provide the missing clues investigators need in order to figure out what happened and why. Which engine failed? Did a particular flight control stop working, and how is that failure related to some other issue, such as an engine hiccup or as sensor malfunction? Causes of mishaps in sophisticated aircraft are sometimes notoriously difficult for investigators to suss out, and data recorders have been a godsend.

DFDR from Flight 294
The recovered digital flight data recorder from West Air Sweden Flight 294, a Canadair Regional Jet that crashed in mysterious circumstances in Sweden in 2016. Photo Swedish Accident Investigation Committee via Wikipedia Commons.

Voice recorders, not so much. The problem isn’t that they don’t work. They work pretty much as they’re supposed to. The problem is that there is seldom anything particularly useful. The reason is clear to any pilot with much experience: When things get dicey in the cockpit, things get quiet. There are two big exceptions to this, one dull and the other chilling. One thing a well crew does when confronted with an emergency is run the checklists. This is almost, but not always, a good thing. Anything that goes beyond a memory item, that is, a checklist item you need to remember and be able to perform completely by rote, takes time and can divert attention from new problems that might crop up. In rare instances, they can delay an airplane getting back on the ground, when it needs to do that— like right now—the checklists be damned.

The most noteworthy and devastating case of this was the crash of Swissair 111, a McDonnell Douglas MD-11, which crashed short of Halifax after delaying its landing to dump fuel to get down to its maximum landing weight. It’s possible that had the pilots diverted directly to Halifax instead of requesting a diversion (which lasted more than 10 minutes) to allow it more time to dump fuel, the plane might have landed safely, or at least with some survivors. As it was, the cockpit became completely engulfed in flames, the pilots, if they were still alive, lost all control of the plane, and all 229 passengers and crew perished when it crashed into the ocean.

Whatever a hardened video recorder might have preserved is not something that anyone would ever want to see. And it’s hard to imagine what useful information might have emerged. It’s hard to even imagine.

The most common useful data that investigators do get from CVRs is engine, wind and flap noises, from which investigators can derive flight data that might be missing.

But there are cases when a voice recorder is critical to understanding what happened in a crash, especially when the flight crew are longer around to share that information.


The most horrifying and dramatic of these is pilot suicide, which investigators have cited as the probable cause of a few crashes that are beyond tragic. While such crashes are rare, investigators determined that pilot suicide was the probable cause of a few terrible airline disasters, including EgyptAir 990, which in 1999 crashed into the Atlantic, south of Nantucket, killing all 217 aboard, and GermanWings Flight 9525, which crashed into mountainous terrain in the French Alps, killing all 150 aboard. (This, of course, omits the pilot murder suicide terror attacks of September 11, 2001).

And pilot suicide is a likely cause of another high-profile loss of an aircraft, Malaysia Air Flight 370, which crashed in the South Indian Ocean in 2014 after, investigators believe, one of the pilots (most likely the captain) disabled tracking systems and commandeered the flight, eventually leading to crash into a remote tract of ocean after many hours of flying out of radar contact. In this case, depending on how the requirement were implemented, if it ever is, a video recorder might not have helped, even if it ever could have been found. While the data from recorders invariably survives the crash, finding the recorders has proven more difficult. Malaysia Air 370s is still missing (and would likely have no data were it ever found), and it took searchers nearly two years and tens of millions of dollars to locate the recorders from Air France 447, which crashed in the remote Atlantic Ocean in the middle of the night on June 1, 2009, on a trip from Rio de Janeiro to Paris. That accident was found to have been caused by iced-up airspeed sensors, leading to the pilots inadvertently deep-stalling the Airbus A3330 into the sea, killing all 228 aboard. Would a video recorder have helped with either of these losses? With Malaysia, possibly, if it had enough recording time and improved discoverability.

With the Air France crash, there’s little doubt that investigators would have been able to explain the cause of the crash very shortly after having recovered the video, if such a recorder had existed. So, there are cases in which recorders can solve mysteries.

But is it worth it? The NTSB would like the mandate to include all commercial airliners, both new and existing. The cost of such a fleetwide mandate would be staggeringly high. Pilots have a legitimate privacy concern, too. It’s far easier to keep new data-gathering devices out of the cockpit than it is it get them out once they’re installed. And pilots fear that the devices would be used to enforce petty infractions against them, which could result in disciplinary action, along with the possible suspension or revocation of their certificates, realistic fears and all.


And that’s not to mention the very real fears that such video recordings of a crew’s last moments would become part of the public record, which is a terrible thought. I think of the crash of the Sikorsky S-76 carrying basketball superstar Kobe Bryant, seven other passengers—including his 13-year-old daughter Gianna—and the pilot, all of whom were killed when the pilot lost control of the helicopter in foggy weather and crashed into terrain, killing all nine aboard.

Those final moments of terror are not anything that anyone should ever see, even if, in rare instances, the recording could lead investigators to an answer. But answers are what they want. After all, it is their job, and because in a few rare instances recorders might provide answers that otherwise would forever remain a mystery, there’s a very real chance that the NTSB’s recommendation at some point before long turns into a new FAA mandate.