While we are all watching the skies for aircraft passing overhead, FlightRadar24 to see where they are going, and (sadly quite often) the media for news of incidents and big events in aviation, the pilots are being watched too, and more closely than you probably think.

The big old ‘black box’

Let’s start by looking at the data investigators review after major incidents and accidents. The ‘black box’, as it is colloquially known (although it is orange for ease of location,) is the primary flight data recorder (FDR). These are mandated on aircraft with 10 or more seats, over a certain weight and manufactured after October 11 1991. Many others now have them retrofitted. So any commercial aircraft you fly on (in most countries) will have to be equipped with one of these.

These have to record and retain a minimum of 25 hours of continuous flight operation, and the recording covers a whole range of parameters including time, pressure altitude, indicated airspeed, heading, vertical acceleration, pitch attitude, roll attitude, manual radio transmitter keying, thrust of each engine, auto pilot engagement status… You know what, there are around 91 or so parameters that these generally store. Here’s a link to a list if you want the full lowdown.

Aside from these, there is also the cockpit voice recorder (CVR). These used to just have to retain the last 2 hours, but it was increased in February 2026 to 25 hours as well (for future manufactured aircraft). Currently, most probably still only have the 2 hour recorders installed. These automatically start recording when an aircraft is powered up, and they record both incoming and outgoing transmissions, as well as general cockpit chitchat including briefings, conversations, calls to the cabin etc. They also preserve other sounds such as engine noise, stall warnings, the sound of the gear going up and down and any random loud-ish noises that might occur.

Yes, the pilots complaining about their pay or gossiping about colleagues is caught on the CVR. However, there are fairly strict privacy rules attached to which prevent airlines from downloading and listening in to simply eavesdrop on their crew. They are only downloaded for investigative purposes involving serious incidents and accidents. This is very important because, as we know, human factors often play a large part in events, and these provide a different level to understating how things happened which basic flight parameters often cannot.

Then we have all the surveillance systems and the data they provide – things like the ADS-B trajectory information which is utilised on FR24 for example. All of this is combined with historical environmental data, reviews of company manuals, SOPs, training information etc etc which helps investigators both see what was happening but also try to make sense of the situation, and understand what factors may have led up to it.

The endless data stream

But for pilots, this isn’t the only time their actions are reviewed. And no, I am not talking about simulator checks or line checks. Pilots, or rather the data related to what they are doing is actually reviewed all the time, because airlines have the ability to download flight data for every single flight. Not only do they download it, but they check it continuously as well – using algorithms which detect when parameters are exceeded.

This might sound a little ‘big brother’ but the monitoring and review of flight data falls under airline safety management systems – SMS, and it is a critical component in training and development which helps prevent accidents and incidents (as well as solve the cause behind them when they do occur).

The way it is used is fairly prescribed, and again – it is all about improving safety in an airline, but for pilots this does mean they are constantly monitored, and because of that, they need to report when things don’t go to plan.

A good safety culture should ensure that none of this used punitively, unless gross negligence is suspected. In general, this means pilots who report occurrences – when they know they have exceeded something, or broken a specific flight procedure unintentionally – will not be punished. Instead the company will try to learn from the error or mistake by reviewing what led to it, ad by considering what might be done to prevent similar occurrences in the future. Most of this happens within the safety department, but if something is serious enough, or a trend emerges for a pilot, then training might also be involved.

A good pilot is a ‘perfect’ pilot, right?

You might be thinking “no good pilot should ever mess up though?” but we are talking a huge number of potential parameters here – from larger and more serious ones like being unstable below 1000 feet, over-speeding the flaps, exceeding 250 knots below 5000 feet, to more minor things like not carrying out the after start checklist before taxi because of distraction, or landing a few feet outside the touchdown zone on a 4000 metre runway because of an unexpected tailwind gust at five feet…

Note I’ve put some reasons why things might happen in there – things can occur outside a pilot’s control, because no matter how much you can try to prevent and mitigate, the environment and other aircraft might have a different plan! An aircraft routing over the Himalayas (known for its turbulence and mountain waves) and can exceed Vmo/Mmo – in what is termed an overspeed – fairly easily if the wind shifts too quickly for the automation to counter it.

In this situation, the pilot might be able to use the speed brake to try and prevent or reduce the overspeed, but they may not be able to either if it occurs too rapidly. This isn’t a ‘pilot error’ but needs reporting since depending on the severity of the overspeed, some engineering checks may be required. Likewise a TCAS event may be due to another aircraft, or a pitch above the recommended degree on landing may occur due to wind shifts.

A big part of the pilot role is mitigating threats and errors, but also trapping and correcting them when they do occur, and FDM is about looking to see why something happened, to learn from it in the future. So when pilots make omissions or mistakes, or when things occur outside the ‘norm’, the pilots report this (through an air safety report – ASR), and in that report they generally try to provide an understanding to the safety department as to why it happened, how they corrected it and possibly can make suggestions to ways they would avoid it again the future. This is all important to understanding and preventing those future occurrences.

FDM = SMS

So Flight Data Monitoring, or Flight Quality Assurance is one of the backbones of safety and training in airlines. An FDM agent, often a pilot themself because it helps understand the data, receives reports showing exceedances or breaches to SOPs. This information is generally de-identified so all they see is the data and any contextual information for the flight which might be related to it. If they can easily assess what happened and why, and there is not major safety issue, these are likely just filed away for statistical purposes.

When it is more serious, or requires more discussion to understand it, it might be sent for review by safety management and they might then contact the crew involved for further information. In instances where more training is required, then the training department will be involved as well, as I mentioned.

That is for events pilots catch and report. Of course, they may not see everything, and in some instances, where their capacity is down or situational awareness degraded, it can be really beneficial for pilots to be shown the data so they can see what happened. Often, this can be put into ‘animations’ which show a simplified version of the flight deck instruments so the pilots and safety department can watch in real time all the various inputs occurring and it is much easier to take it all in watching it in this way. Pilots often cannot recall every detail accurately if an event was high pressure, time critical, or complex because our brains just cannot do that.

The benefit of the all seeing eye

The data downloaded is not always linked to pilot error of course. It is all the parameters from the aircraft and might show a technical fault as much as a pilot one. The maintenance department also receives information from the aircraft, helping them detect issues which the crew might not see, and enabling them to plan for unscheduled maintenance more proactively if something occurs during flight. In complex failure situations, this data can also be reviewed by the manufacturer to assist in trouble shooting a solution for the situation.

An A380 has a very complex fuel system which automatically shifts fuel about for CG control, wing loading and also things like fuel temperature control. One auto function it has is to pump fuel forwards into the engine feed tanks when within around 200nm of the destination. Jettison is to available from these tanks, which normally isn’t an issue. However, if (as happened for one) they are diverting (and in the case I am thinking of, it was for a medical), this can become problematic. Although the A380 can land up to MTOW, it requires the CG to be within certain limits for landing – otherwise there may not be enough pitch authority for the flare.

Long story short, that’s what happened and the reason I included this here is because it was a good example of how the data downloaded provided more information than the pilots could see, and this helped Airbus and the airline find a solution for them fairly quickly.

Other ‘non pilot error’ monitoring benefits include things like fuel consumption, engine efficiency, even on time performance (when doors close, when park brake is released), to whether they are compliant with airport and airspace requirements such as noise abatement, ATC procedures etc.

We already mentioned the parameters it can monitor. The basic operational event set looks at hard limit exceedances – so things like Vmo/Mmo (the maximum speeds for the aircraft), Vle/Vlo or Vfe/Vfo (the maximum speeds for landing gear or flap operation), approach speeds (stable approach criteria), takeoff and landing pitch rates (things that can lead to tail strike), bank angles, rates of descent, abnormal configurations, and activation of warning systems such as GPWS, TCAS, stall etc.

Beyond this basic set which tends to cover those hard do not exceed parameters, there are extended sets which look at more nuanced parameters. Engine parameters, taxi speeds, acceleration during takeoff roll, how long an aircraft holds for prior to landing (useful for fuel burn statistics), how many minutes between landing and engine shutdown, altitude deviations, speed at rotation, landing force, fuel usage, whether a more efficient flap setting was used for landing, how long between 50 feet and touchdown, between flare and touchdown (good for spotting long landing trends) etc.

Some airlines provide summaries of general data to their crew. For example, they may receive guidance on what percentage of flights they are achieving fuel savings on, stats on how often they carrying out an engine shutdown during taxi, landing with idle reverse, or a lower flap setting. They might also receive information on their on time performance. The aim isn’t to pressure pilots, but to provide awareness and to get them thinking more about when these might be appropriate to do, and when they might not.

Always watched…

An airline with a robust, and ‘just’ safety culture use FDM data to review their safety and operational standards in order to enhance procedures and training, and to build awareness and understanding of their performance in crew. It is a good thing… but yes, sometimes it can pop into your head without warning that all eyes are on you, and whatever you do will be seen!