Lost flights, exploding rockets and death: when data goes wrong | Tech News

Holidaymakers have had a nightmare week, and for those still trying to get home after thousands of flights were cancelled, it won’t help much knowing the carnage they’re caught up in was caused by some rogue data.

National Air Traffic Services reported the chaos erupted because a flight plan its systems could not read caused an emergency shutdown in the system, grounding planes and leaving passengers stranded abroad.

And let’s face it, ‘data’ as the cause of any major disruption doesn’t scream sexy headline or future feature film.

Nevertheless, our lives are ruled by data, and when it goes wrong, it can go very wrong.

‘In the last two years, we’ve produced around 90% of the total amount of digital data since the advent of computer storage,’ says Amin Fard, MD of data management specialist Lyon. ‘Our reliance on data from here on will become ever more critical. 

‘Anything that alters the truthfulness of data can bring organisation to a halt. This can take shape in many formats – cybersecurity breaches that encrypt data, viruses or malware embodied within organisational data, human error leading to the deletion of data, human error leading to distribution of incorrect data, and more. 

‘A prime example is when back in March 2019 a BA flight landed in Edinburgh instead of Düsseldorf by mistake due the wrong flight plan being submitted. This is what some called an honest mistake that could have been prevented with varying levels of data validation.’

Human error and data can form a potent mix. Earlier this month the details of hundreds of PSNI (Police Service of Northern Ireland) officers and staff were accidentally leaked during a routine FOI request.

The consequences of this mistake are still unknown, as is the cost of this week’s flight meltdown.

‘The losses incurred as a result of the flight data issues are expected to be astronomical,’ says Mr Fard. ‘Although we haven’t fully understood the intricate details of how it all happened, there are steps which we can take to prevent this from happening in the future especially if incorrect data feeds were the culprit.’

The fallout from incorrect data in the past ranges from a frustrating spreadsheet to death. Bad data or rogue coding may itself be only a few bytes, but its impact can be global.

Here are more examples of when bad data had massive consequences.

1. Between 1985 and 1987, five patients died after a coding glitch in a Therac-25 radiation machine caused it to administer massive doses of X-rays. A sixth patient who had treatment around her collar bone suffered severe radiation burns, completely lost the use of her shoulder and arm, and had to have a breast removed. 

2. Sometimes you can put an exact cost on a software glitch. On August 1, 2012, rogue software sent the financial services firm Knight Capital Group on a $7 billion spending spree in less than an hour, leaving the firm in receipt of shares it couldn’t afford and facing bankruptcy. Goldman Sachs stepped in, buying the stocks for just $440 million.

3. Earlier this year, the much-loved Ariane 5 rocket was retired by the European Space Agency. It had a bumpy start however, after a software bug caused a test vehicle to veer off course self-destruct 37 seconds after launch. Carrying four Cluster satellites, the failure cost more than $370 million. 

4. In 2021, Tesla was forced to recall almost 12,000 vehicles after a glitch in its beta software caused cars to suddenly slam on the brakes after mistakenly identifying collision threats – risking rear end collisions with the traffic behind.

5. Another space-based data snafu, this time a little further from home. Nasa’s Mars Climate Orbiter was sent to study the Red Planet from above and relay communications for the Mars Polar lander and Deep Space Probes. Unfortunately, the spacecraft worked in imperial measurements, whereas the team on the ground controlling it used metric. Data carnage ensued, and the orbiter hurtled into the Martian atmosphere, being torn apart as it went.

6. Of course, data precedes computer technology by hundreds of thousands of years. It is simply ‘facts and statistics collected together for reference or analysis’, something early humans will have used in their own way, such as learning which plants are poisonous, or where certain animals will be at a certain time to hunt. 

As such, bad data has also been around for millennia. More recently, bad data led to Russia selling Alaska to the US for just $7 million in 1867 – seemingly unaware of its vast gold and oil reserves – and, of course, the sinking of the Titanic in 1912. 

MORE : UK air traffic chaos ’caused by French airline’s wrong travel plan’

MORE : EasyJet promises to fly stranded Brits back to the UK on repatriation flights

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.