Tomorrow everything will happen online. Of course for us, individuals, but also for many objects which connectedness will enable to produce and share data and, finally become intelligent what means being able to receive instructions and even make decisions and act upon them.
The major stake coming with this evolution is about security. Either about self-driving cars, digital fingerprints, contactless payment, the level of risk we accept today will be pushed to an unprecedented level.
These risk have in fact nothing new. Car or identity theft are nothing new. What’s going to change is the scale at which it will possibly happen. A “car-jacking professional” can make one or two victims a day, the guy hacking cars will be able to this at an industrial scale. Same for digital fingerprints.
As always when it comes to IT security, the question is not to know if a system is unbreakable but how long and what it will take to break it. Cars, planes, means of payment will be hacked with the consequences we can imagine. The real question is to know if we are ready to consider it as an acceptable and unavoidable issue or like an unacceptable risk.
But it’s not only only about hacking. Dysfunctions also matters. It reminds me of a conversation I had about delivery by drones. Imagine that it becomes mainstream : there is no doubt that one day or the other we’ll have drone accidents. What will be the impact if a drone, even light, carrying a parcel, even light, makes a 300m fall and crashes on the crowd ? Will be consider that this is an acceptable risk or will we refuse to have one or two people injured every week. This is a real issue and we’ve already been close to a catastrophe.
We’re a little bit like in the first half of the previous century. Innovation was at a high level and side damages were numerous until we reached a satisfying level of mastery. But two things are different from then :
– appetite for risk was high. Today no one would take the risks taken by the first aviators, by the ones who first tried to break the sound barrier, by the fist astronauts who put their lives in the hands of computers less powerful than today’s pocket calculators. Even in medicine, they started from so far they were ready to test drugs and vaccines with protocols no one would accept today. In short they learned by doing and it was normal. Today we live in risk averse society and the idea of hazards has disappeared from our system of thinking.
– the scale of risk is different. It was limited to those willing to take it. Today it will be about thousands or hundreds of thousands people at least.
In fact risk is not the matter. Trust in this world under construction is. Today it’s nearly possible to land a plane without any human action. So why don’t we do it ? Because even if the system was a thousand times safer tan with humans the first accident will call the entire industry into question. We will prefer the least imperfect human over the nearly perfect machine for long. Remember of the debate about killer robots : before, the issue would have been dealt with afterwards, today we ask for guarantees before robots become a mainstream reality.
On such critical matters with high network effects, if hazards are not considered normal (with their consequences) the first problem will harm an entire industry. Period.
There’s no doubt that what’s under experiment today will become mainstream, even at a scale we cannot image today. We’re incapable of imagining what 2050 will be. On the other hand something is sure : we’ll need time to unleash machines at scale. Because we will not want a simple hazard to stop progress by making it socially unacceptable.