Believing in humanity again

Dated: 

2 February 2017
Breaking News: The Dutch government has decided to abandon "software solutions" in their forthcoming elections: http://www.standaard.be/cnt/dmf20170201_02707485   That's one small step back for machines, one giant leap for mankind.

Similar advancements are being made in connection with self-driving cars: Safety engineers are casting doubts about the viability of having self-driving cars roam our cities. Security experts — in computer science's flagship journal, Communications of the ACM (February 2017) — even contemplate the possibility that the Internet of Things (and especially: Moving Things) will become a failed technology. Are self-driving cars the zeppelins of the 21st century?

Regardless of whether these "technology pessimists" are right or wrong (or a bit of both), voices of concern are being raised and are starting to be heard by fellow engineers. This progress comes despite the large investments that have already been made to deploy fully-autonomous vehicles as soon as possible (if not already), so that our children allegedly will not have to learn how to drive.

.

That being said, as a historian of technology and a concerned safety engineer myself, I still foresee the following publication in a few decades from now:

"During the early years of self-driving cars, progress created among many the strong feeling that a working system was just around the corner. The illusion was created by the fact that a large number of problems were rather readily solved. It was not sufficiently realized that the problems solved were just the simplest ones whereas the few remaining problems were the harder ones, very hard indeed."

That is my slight modification (already presented here) of Yehoshua Bar-Hillel's 1960 report `The Present Status of Automatic Translation of Languages' in which he retrospectively scrutinized the entire field of machine translation. In a forthcoming talk (to be announced in a couple of weeks) I will provide arguments why the next chapter in the History of Failed Technologies will be one about the Internet of Moving Things and, specifically, network-connected cars that are designed to roam arbitrary roads. On a positive note, I hope to convey that some less ambitious i.e., more incremental projects, such as the deployment of podcars on railroads, are far more realistic in actually solving the mobility problems that we face today. My inspiration also comes from Peter Hellinckx's insightful talk, which he presented a couple of months ago in Antwerp.

A sociological reflection, in turn, amounts to asking the following question: Why do some entrepreneurs want to change the world in a drastic manner, given that small, incremental steps are more likely to succeed? Perhaps they are not aware of computer science's holy grail, which is to engineer systems that are both correct by construction (or completely reliable) and fully automated. In layman's terms: a seemingly insignificant software error in the digital system of a network-connected, automated car can, when detected by a malicious hacker, result in a crash of that car and in thousands of other crashes. (Having cars mechanically attached to railroads is one of many ways to at least reduce the impact of this hazard. Mechanically and drastically constraining the cars' maximum speed, is yet another.)

Safety engineers are paid and expected to envisage such calamities and I shall therefore refrain from refraining to do so in my talk — entitled: Self-Driving Cars are the Zeppelins of the 21st Century: Towards Writing the Next Chapter in the History of Failed Technologies with the main purpose of opening up a responsible debate about our future digital world. 

Tags: