In a "futuristic" society (think Ghost in the Shell or Human Revolution), where you have AI driving vehicles that carry goods in a reliable way, is there any reason to have a human traveling with the cargo? Is there any reason to have a cabin for that matter?

The "hauler" unit is capable of sending a distress signal if anything goes wrong along the road, which in this case means something out of the control of the AI happened.

The AI is connected to a Network, but is perfectly capable of finishing the travel in case of a disconnection. Technicians can remotely control the unit at any time. They use airless tires so no punctures.The vehicle is electric, and no recharge is required for the distances they are meant to travel.The cargo container can only be opened at the intended destination, until then is practically sealed.

The setting for this is that I need cargo disappearing from containers in a mysterious way, no witnesses, so I need the human out of the formula, but I want to have a good reason for it to not be there. This type of transport is commonplace and cheap as can be. This is just for city to city transport, local delivery trucks would still require people. Note that this is for regular stuff, it is to be understood valuable cargo would still go escorted.

  • $\begingroup$ You know thia is already a reality, right? Self driving carsare ever more common. And Amazon does deliveries with flying drones. $\endgroup$ Commented Dec 29, 2017 at 4:10
  • $\begingroup$ Am I missing something here? You want to know if it's believable to not have a human in the vehicle if it can drive itself? Obviously yes. Having goods in the car doesn't change this. $\endgroup$ Commented Dec 29, 2017 at 4:23
  • 1
    $\begingroup$ "The cargo container can only be opened at the intended destination" and your computer is unhackable. And credit card data can't be stolen. Yea, right. $\endgroup$
    – Mołot
    Commented Dec 29, 2017 at 9:12
  • 1
    $\begingroup$ @CandiedOrange The type of cargo does matters, there's a big difference, between carrying tourist vs tomatoes. $\endgroup$
    – Inferry
    Commented Dec 29, 2017 at 18:23
  • 1
    $\begingroup$ @Inferry Then explain that difference in your question. $\endgroup$ Commented Dec 29, 2017 at 18:25

6 Answers 6


There are already autonomous vehicles that have no human driver. As per vice.com:

By the end of 2013, there were 48 fully automated public metro systems in use in 32 countries, according to the International A​ssociation of Public Transport (UITP), an advocacy group that promotes public transit.

Those would seem to be passenger trains. Cargo should be even easier to justify.

Presumably this will happen with trucks too, although trains have the stronger use case.

The primary argument for a human backup is that the human might catch circumstances that the AI doesn't. But as the AI gets better, the human is more likely to introduce error than correct one. At some point, pure AI is safer than human-backed AI. And frankly, we may have already crossed that point.

It's noteworthy that the only self-driving car fatality was caused by a human repeatedly passing control back to the AI. In a pure AI, the AI would have pulled over to the side of the road and rebooted or called for repair.

You also may want to note that Positive Train Control is going to be required for all trains. PTC requires that all trains can be stopped automatically, without human intervention. It's a small step from there to allowing trains to go without human intervention. And as I previously said, the technology for allowing trains to go human-free already exists.

  • $\begingroup$ Thank's for providing those references, they will be helpful for what I want. $\endgroup$
    – Inferry
    Commented Dec 29, 2017 at 19:42

I would say that a necessity of having a human supervisor of the cargo depends on three factors:

  1. frequency of thefts;
  2. AI's vulnerabilities to hacking;
  3. business' ability to compensate for lost goods.

You want a human supervisor if goods are stolen often,

A high number of successful attempts to steal your goods indicates that something is wrong and the AI is not capable of doing its job. A human supervisor can be appointed temporarily or permanently to act as a failsafe mechanism or for debugging purposes.

if the AI can be fooled or hacked on the road,

A developed fully-robotised and autonomous transportation system will most likely be accompanied by a special infrastructure of sensors, navigational beacons, guiding transmitters, etc. These can be installed on the roads and on self-driving cars. If hackers can access this infrastructure and use it to fool the cargo unit, a human supervisor might be necessary to prevent a theft.

if business cannot pass the cost of lost goods onto consumers.

Some businesses can charge more for their products without losing customers, but it might not be the case if competition is high and consumers have a lot of choices. In a competitive business environment, companies will be doing everything to protect their goods, including adding a human supervisor if it reduces the number of thefts.

Additional political considerations

Depending on your country's labour and safety laws, any self-driving vehicle may have to have a human driver. It can be a remnant of earlier distrust to AIs or it can be a policy protecting jobs.

If you want to get rid of human supervision, make thefts insignificant in number and damage, allow oligopolies, and don't let socialists and progressivists to rule. The free market will take care of the rest.

  • $\begingroup$ Your answer is a really good, I already have most of these considerations, and I was really looking for something I might have oversight. $\endgroup$
    – Inferry
    Commented Dec 29, 2017 at 19:56

Would you need an expensive driver? Nope. Could you use a cheap handler? Maybe....

If you're dealing with a mature autonomous driving technology then (with the possible exception of repair diagnostics) you would never need a driver. Why would you? The machine can coordinate with routing much faster and much more predictably than a human could. It would never be distracted by the babe/hunk on the side of the road, would never need to stop for coffee breaks, and could be depended on to drive as quickly and efficiently as possible (vs. humans who tend to time streetlights to get all the reds they can because they're paid by the hour).

But a handler. You should have seen the stairs to my house before we remodeled. Raw rock, uneven, set poorly, unpredictable step heights (and widths), loose stones, bad after decades of use... when it comes to dealing with the far more unpredicitable situation of actually unloading the vehicle and making the delivery at a home, nothing beats the versatility of a human being. (Not to mention that winning smile when you deliver a nice elderly woman her package!)

And since it's unskilled labor, it's minimum wage.

Home delivery might not be the only human-useful delivery issue. Navigating crowds of people in a busy business district to deliver to some start-up on the fourth floor of a renovated slumlord building... narrow halls, steep stairs, people with robophobics who have no compunction against swinging an aluminium baseball bat at the 'bot... all would give way to the sharply-dressed minimum-wage worker (mostly).

And, lest we forget, humans are cheap to fix. Robots aren't. Never send a multi-million dollar robot when a disposable human can do the job. The economics of today the future are ruthless.

Pretty much any other kind of delivery would be better off using an automated skidsteer.

As for would you need a cab? No. But you would need access to very basic controls in the cargo cabin in case of accident where the automation is lost and somebody needs to move the vehicle. But a handful of pushbuttons for "move forward" and "turn left" for such adjustments is nothing more than a panel.


Why bring a human if you aren't going to use them

If you are a company shipping goods, why would you pay a human to go along with an autonomous vehicle? If you need loading and unloading, it makes sense to staff the source and destination locations. But sending someone with the payload is just a waste of money.

If you really want to send someone along, why not just have that person drive? If you are paying a human anyways, and your human is capable of driving, then you can save money on your vehicle by not having it be autonomous.

The whole point of autonomous vehicles is to stop either a. people from wasting their time driving or b. companies from having to pay human drivers. There would be no humans with an autonomous transportation vehicle.

  • $\begingroup$ I agree generally, but not with "you can save money on your vehicle by not having it be autonomous". When unmanned cars become generally usable (whether that's 10 or 100 years from now), they will be better drivers than humans. The insurance discount will dwarf the tech cost. Plus you can tell the human to do other stuff during the trip. $\endgroup$ Commented Dec 29, 2017 at 10:13

Usually you have a human in self driving cars because the AI isn't capable of handling all situations and needs a human to take over if it gets stuck. Humans are usually responsible for what the cars does, since it's a very iffy situation with AI and how they interact in all those philosophical problems where you have to murder someone (the 1/5 people train problem always pops up).

In a futuristic society, having fully automated cars driving around is 100% viable and I'm pretty sure this is also where we are heading today. The distress signal is a little interesting but pretty pointless. By the time someone responds to the signal, anything that could go seriously wrong (robbery, crash) would've already occurred and wouldn't exactly be fixable.

As for the cabin, you would likely still have some variation of it for aerodynamics and to store the engine, batteries and computer running your vehicle.

The stuff below are just extra stuff I would like to point out.

Even if there was a human in the vehicle, they would probably end up being asleep or distracted if the AI is reliable. Doing something as tedious as monitoring an AI day in and out gets pretty boring. It would be possible to take the cargo while a human driver is in the cabin without them noticing, which would add to the mysteriousness of the theft.

Your vehicle can be remotely controlled and is connected to a network. This means it can be hacked and the hacker can remotely control it. You could trick the car into thinking it's at its destination (you could even drown out a GPS signal and replace it with a fake signal). With extensive enough resources, you could even have the driver think they delivered it to the correct location.

  • $\begingroup$ The signal is more of a notice that something happened rather than a plea for help. Also, don't spoil the plot pls :P $\endgroup$
    – Inferry
    Commented Dec 29, 2017 at 18:35

In the early days of autonomous driving there will always be a human with the vehicle to monitor for errors by the computer. But, in a society as advanced as in Ghost in the Shell, the AI is so advanced that the number of accidents is probably incredibly low, making it too costly to bother with a person on the vehicle at all.

With no need for a cab area, cargo vehicles would most likely look like shipping containers on wheels with an antenna on top. The entire engine/drive system would be underneath, allowing the maximum amount of space for cargo. You can still play with the design however on the points that aerodynamics are important to efficiency, and companies know that their vehicles are billboards so they want them to stand out.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .