The Dangerously Stupid Thinking Surrounding Self-Driving Cars and Autopilots

By Rob Enderle September 18, 2013

This week there were two interestingly similar, though unrelated, stories about self-flying planes and self-driving cars.   On the Today Show, it was said that the FAA came out with a report advocating  that pilots who use autopilots need to stay vigilant in case the autopilot is unable to fly the plane. Apparently a number of crashes have resulted from pilots being unable to take control from the autopilot during an emergency.   In the Huffington Post, there was a similar story regarding self-driving cars and noted that should a car get into trouble while the driver is napping or doing something else the result would be equally catastrophic. 

Both seemed to advise that the fix is to assure that the pilot or driver be ready to take control in an instant which kind of destroys the reason for either technology in the first place. I mean if you have to be instantly ready to fly and drive you might as well be flying or driving in the first place.   The right fix is to take the step to make the pilot and driver equal, because in both cases they, not the computer, are the weakest link. 

Let me explain.

The False Assumption

Underneath the assumption that there needs to be a human in the loop is the erroneous belief that people can do this jobs better than computers.   Generally when machines are used successfully to do any job, this belief is false.    If a human is needed to step in when an event occurs it either showcases that the task shouldn’t have been done by a computer in the first place or there was wrongheaded thinking in the planning and creation. 

Why the last is true is because computers don’t get tired or sleepy, they don’t do drugs or get drunk, they can make decisions in a fraction of the time a human can, and, properly configured, they can go from inactivity to activity in fractions of a second.   They are also generally more robust and far faster than a human in anything that both do well.   

In short, a computer would be far better as a back-up for a human than a human would be as a back-up to a computer.   

This suggests that rather than trying to think of ways to get pilots to stay engaged when they aren’t flying or drivers to stay engaged when they aren’t driving, the focus should instead on creating systems that can better back-up a human or driving robot should a problem result.   For instance, many drivers haven’t trained to drive in snow or on ice and while Pilots are supposed to be trained in all eventualities there will always be events that the pilot either hasn’t trained for or doesn’t remember forcing them to look up what to do while a computer can instantly identify the problem and find a related solution in a database.   Done properly a computer, will always be more prepared for a catastrophic problem than a non-flying or non-driving human will.  

Wrapping Up: Fixing the Problem

Given a computer will, on average, be the better then a human in an emergency, the better solution would be to either have the computer step in if the pilot or driver gets into trouble or have a redundant computer which can step in if the primary computer gets confused. Humans just aren’t well equipped to go from a passive to an active role instantly, while that is actually a strength of the computer.   Humans can creatively come up with a solution but they don’t do that well in a panic which is what typically results if the human has to, with little warning, go from a passive to an active role and deal with a crisis.  

So I think the FAA is wrong in their approach to the problem. I say, either eliminate the use of autopilots to keep the pilots engaged and provide a more capable computer backup (kind of a ‘push button if crashing’ scenario) or provide a redundant computer which is designed to come on-line in case the primary autopilot fails and automate the flight. Same with self-driving cars, design them so a human doesn’t have to jump in unprepared to address a problem.   We’re just not good at it.  

Edited by Stefania Viscusi

President and Principal Analyst, Enderle Group

Related Articles

Bloomberg BETA: Models Are Key to Machine Intelligence

By: Paula Bernier    4/19/2018

James Cham, partner at seed fund Bloomberg BETA, was at Cisco Collaboration Summit today talking about the importance of models to the future of machi…

Read More

Get Smart About Influencer Attribution in a Blockchain World

By: Maurice Nagle    4/16/2018

The retail value chain is in for a blockchain-enabled overhaul, with smarter relationships, delivering enhanced transparency across an environment of …

Read More

Facebook Flip-Flopping on GDPR

By: Maurice Nagle    4/12/2018

With GDPR on the horizon, Zuckerberg in Congress testifying and Facebook users questioning loyalty, change is coming. What that change will look like,…

Read More

The Next Phase of Flash Storage and the Mid-Sized Business

By: Joanna Fanuko    4/11/2018

Organizations amass profuse amounts of data these days, ranging from website traffic metrics to online customer surveys. Collectively, AI, IoT and eve…

Read More

Satellite Imaging - Petabytes of Developer, Business Opportunities

By: Doug Mohney    4/11/2018

Hollywood has programmed society into believing satellite imaging as a magic, all-seeing tool, but the real trick is in analysis. Numerous firms are f…

Read More