Autonomous vehi­cles, or AVs, will be the most dis­rup­tive tech­nol­o­gy to hit soci­ety world­wide since the advent of the motorcar.

This pro­nounce­ment by the team of jour­nal­ist Karen Kel­ly and for­mer New York City traf­fic com­mis­sion­er Sam Schwartz is the sort of boil­er­plate futur­ism you’ll find writ­ten about any new tech­nol­o­gy. Like­wise, the very next state­ment could almost be chalked up to typ­i­cal hyper­bole: “Some futur­ists and pol­i­cy experts even talk about dri­ving being banned on some or all roads.”

What sets No One At The Wheel: Dri­ver­less Cars and the Road of the Future apart from that sort of replace­ment-lev­el schlock isn’t where it looks for­ward, then, but for how it looks back­ward to show how a sim­i­lar process already happened.

A cen­tu­ry ago, the orig­i­nal grand theft auto was let­ting the car indus­try steal the roads from pedes­tri­ans and non-motor­ized traf­fic. Soon, dri­ver­less indus­tries will be in a posi­tion to take the roads from the pub­lic entirely.

But only if we let them.

Many peo­ple like to make com­par­isons between horséd car­riages and how they irrev­o­ca­bly gave way to the horse­less ones, but this is the first book I’ve come across to make that com­par­i­son direct and exam­ined in detail.

In a dif­fer­ent book review, I argued that for all of the breath­less ado­ra­tion about the future of autonomous vehi­cles, we could achieve huge reduc­tions in traf­fic fatal­i­ties by uti­liz­ing proven tech­nol­o­gy, which coin­ci­den­tal­ly has proven cost-ben­e­fit analy­ses already per­formed for it.

Although not the main tar­get of that review, apartheid-era white South African and “rough-upbring­ing” sur­vivor Elon Musk is as close to the pla­ton­ic ide­al of that posi­tion as you could ask for. Argu­ing that AVs would save many more lives than they’d kill, Musk has called jour­nal­ists writ­ing crit­i­cal cov­er­age of AV tech­nol­o­gy akin to murderers.

“[When] you write an arti­cle that effec­tive­ly dis­suades peo­ple from using an autonomous car, you’re killing peo­ple,” Musk said in 2016.

Yet, Schwartz and Kel­ly point out that Musk is not near­ly inter­est­ed in push­ing mass (pub­lic) tran­sit options or even installing more roundabouts.

Round­about dri­ving | WSDOT

“Round­abouts, or traf­fic cir­cles, reduce fatal crash­es by 70 to 90 per­cent com­pared to stan­dard inter­sec­tions, and injuries in crash­es by 75 per­cent, accord­ing to the Insur­ance Insti­tute for High­way Safe­ty.”

But Musk does­n’t own a round­about-instal­la­tion com­pa­ny, so this method of sav­ing lives is not as moral­ly com­pelling to him.

Per­haps just as impor­tant­ly, we actu­al­ly have some idea just how much it costs to install a round­about, what it can accom­plish, and can’t.

The civ­il engi­neer Justin Rocz­ni­ak, known on YouTube as donoteat01, applied this crit­i­cism to Musk’s pro­posed under­ground Loop under the com­mon engi­neer­ing dis­tinc­tion between AM/FM, that is “actu­al machines” and, well, “fan­ci­ful magic”.

No One At The Wheel gives exam­ples of oth­er ways that so many promis­es of Autonomous Vehi­cles are already achiev­able if we changed our approach to road trav­el to some­thing akin to air travel.

And for air trav­el, no fail­ure lead­ing to a loss of life is accept­able. The rea­son we don’t lose 35,000 peo­ple per year fly­ing is that we know humans are fal­li­ble crea­tures and the sys­tem is designed around folks mess­ing up with­out the result being deadly.

Roundabouts—boringly—don’t involve engag­ing in elab­o­rate philo­soph­i­cal con­struc­tions about which peo­ple tied to rail­road tracks you’d have to kill in the case of a run­away trol­ley.

But they do make it much more dif­fi­cult for any “trol­leys” to go out of con­trol, which seems the bet­ter invest­ment, all-in-all.

So Schwartz and Kel­ly lay out a con­vinc­ing case for why tech­nol­o­gy will con­tin­ue to devel­op to put AVs on the road but that this is a smoke­screen for how the real ben­e­fits will accrue and who for.

Algo­rithms can’t union­ize, and new, pri­va­tized projects are by def­i­n­i­tion more prof­itable than main­tain­ing or improv­ing exist­ing pub­lic systems.

Whether or not AV trac­tor-trail­ers actu­al­ly make soci­ety safer is a moot ques­tion. They’ll get rid of a major fixed cost and point of fric­tion for busi­ness own­ers, and with enough mon­ey spent on the right cam­paign con­tri­bu­tions and good lob­by­ing, any­one killed by the vehi­cle’s bad pro­gram­ming will be the ones blamed for putting them­selves in danger.

Or, it won’t be the fault of the com­pa­ny for releas­ing a vehi­cle that can be eas­i­ly hacked and dri­ven into a crowd; it will be the fault of the hack­er, solely.

Again, the book real­ly shines because it looks at the past and how all of this already hap­pened as auto­mo­biles became legal­ly priv­i­leged over pedes­tri­ans and oth­er modes of trans­porta­tion in city streets.

It was not nat­ur­al, it was not easy, but as the viral­ly share­able illus­tra­tion by Swedish artist Karl Jilg evokes, we sur­ren­dered a lot of valu­able real estate to dan­ger­ous, pol­lu­tion-belch­ing machines that aren’t there most of the time, and don’t actu­al­ly get peo­ple places any quick­er once rush hour hits.

But it was what the auto­mo­bile com­pa­nies and asso­ci­at­ed indus­tries want­ed, and they pushed for it method­i­cal­ly with lots of resources behind them.

In the name of progress, car­riages no longer had to avoid peo­ple: peo­ple had to avoid the car­riages.

I don’t think Schwartz and Kel­ly are being inher­ent­ly alarmist or imag­in­ing a wild dystopia; the super­hero film Logan did a fine job show­ing the plau­si­ble ter­ror of shar­ing the road with AV trac­tor-trail­ers when it’s your fault for not hav­ing a new enough vehi­cle to dri­ve you better.

They also log­i­cal­ly extend the present into a future where the inevitable Uber­Lyft merg­er and its AV fleet are per­ma­nent zom­bie traf­fic because it’s cheap­er to leave them dri­ving aim­less­ly, avail­able to pick up folks, than it is to ever pay for park­ing. In all like­li­hood, the cars won’t even solve Sudokus as they do it.

Tech­nol­o­gy is nev­er just tech­nol­o­gy, and what seems inevitable with ben­e­fit of ret­ro­spect took tremen­dous cost and effort in the moment to be made just so and not dif­fer­ent.

And yet, the authors pro­vide you some cause for opti­mism. The Unit­ed States is not a democ­ra­cy, we know: we were set up to be a repub­lic, like the Romans.

Our pol­i­tics are not incapa­ble of pro­duc­ing pos­i­tive change, in any case, and we could real­ly see huge improve­ments in the qual­i­ty of life for every­day peo­ple if those are the goals we choose to actu­al­ly and con­sis­tent­ly advance.

Schwartz and Kel­ly argue that if the Autonomous Vehi­cles are pub­licly owned, inte­grat­ed into mass tran­sit, and uti­lized only in sit­u­a­tions where they make sense like long dis­tance trips and last-mile for peo­ple with dis­abil­i­ties, then it could be a boon to peo­ple’s time and comfort.

But we have to be vig­i­lant to see that solu­tions pro­posed based on new tech­nol­o­gy actu­al­ly can do what they promise and with­out un-men­tioned, exter­nal­ized costs.

If you were an ear­ly 20th-cen­tu­ry city dweller, get­ting rid of the dan­ger fatal kicks in the head by a horse and the smelly pol­lu­tion of manure on the streets no doubt would have made horse­less car­riages appear an attrac­tive, safe improvement.

Which is the dif­fer­ence between Actu­al Machines and, well, Mag­ic.​