Beside the president’s perversion of social media and the car crash of Uber, few topics have risen up the tech news ranks in 2017 quite like automation.
Gartner forecasts the robotic automation market will grow by 41% each year to 2020 (that’s $6.2bn a year by 2023, according to P&S Market Research), while automation technology in the home will be worth $21bn by 2020, says Zion Market Research.Automation is following the standard tech hype cycleClick To Tweet
Automation is following the standard tech hype cycle. There is almost religious zeal, a belief that smart algorithms, plus a smattering of artificial intelligence, can help businesses put processes on auto-pilot, saving vast quantities of person-hours and freeing executives to make every task repeatable.
The only ones who have to fear the automation revolution are the up to 38% of US workers and 30% of UK employees whose jobs may be lost to automated processes, according to a PwC report.
But the people can sleep a little easier. Because automation will not deliver on its smart promise if its underlying processes are dumb in the first place. You see, automation is simply the combination and automatic execution of existing digital workflows – and, as my new book Humans Vs Computers explains, human executives are notoriously bad at creating those workflows.
Just ask drivers in California. When the state introduced a surveillance camera system to automatically charge road tolls by detecting licence plates, it seemed like a good idea. But the state overlooked its own California Vehicle Code 4456, section C2, which allows a new car to be driven without licence plates for up to 90 days – a loophole which allowed new drivers to drive with impunity.
The researchers at IBM now know the risks, too. When they pre-loaded the Urban Dictionary, a repository of slang, in to their super-intelligent Watson artificial intelligence engine, they were surprised that Watson could no longer tell the difference between profanity and clean language, responding to one user with simply: “Bullshit”.
They shouldn’t have been surprised. Automation is no different from any other GIGO computer process – if you put Garbage In, you will get Garbage Out. And, whilst automation will frequently prove to be fallible in this way, it is the human navigators with their hands on the tiller who we can blame for charting the wrong course.
In the huge obsession with automation these days, people lose track of the fact that automating makes things faster, not better. Speeding up a beneficial process delivers value faster. But automating a bad process only makes it spiral out of control.
The dirty little secret of the software industry is that automation doesn’t make things better, only faster. Take for example, Microsoft’s chatbot, Tay. Tay was designed to become smarter as more users interacted with it, however, it instead quickly learned to spew anti-semitic tweets that human Twitter users fed into the programme. Microsoft eventually realised and shut it down, but the damage had already been done and at a very fast pace.
Every company examining digital automation today should stop, step away from the technology options and vendor websites, and rewind.
A prerequisite to any automated process should first be auditing and examining your current manual workflow – whether it be digital or analogue – for gaps and faults.
You need to look on automation not only as a positive force that can boost business efficiency but also a negative one that, if implemented without prior due diligence, could lead to damaging consequences.
And, before you get tempted, no, you can’t ask a computer conduct that audit. As the world will soon see, sending a bot to do a man’s job can cause more ill than good.