Hurricane Harvey and Asimov’s Laws of Robotics

Texas is still tallying up the damage Hurricane Harvey inflicted on it, but it’s already clear that hundreds of thousands of vehicles were destroyed–overwhelmingly by flooding.

‘”We do know that approximately 100,000 claims have come in” as of Thursday, said Matt Stillwell, manager of governmental and regulatory communications at the Insurance Council of Texas, a trade association. He said the number was expected to climb as high as 500,000.

“It is looking to be a huge impact on the auto insurance market,” he said.

While homeowners’ insurance policies almost always exclude flood damage, comprehensive auto policies do cover flooding. The typical household in Houston has two cars, and Mayor Sylvester Turner urged residents to “hunker down” as Hurricane Harvey made landfall, hoping to avoid a replay of the tie-ups and crashes that killed about 100 people fleeing Hurricane Rita in 2005.

That means few people moved their cars out of harm’s way before the flooding started. Texas drivers are not required to have comprehensive auto insurance — the type that covers flood and other types of damage. People holding only the legally required insurance — liability coverage for damage done to other people’s cars — will not have valid claims, Mr. Stillwell said.’    SOURCE

(FYI, floodwaters are bad for cars because they destroy their computers and wiring and fill their cabins with mold. In general, once the water level rises halfway to a car’s roof, it’s totaled.)

This loss is going to cost car insurance companies billions of dollars in payouts, and since there’s no such thing as a free lunch, it will be paid for by raising everyone’s insurance premiums. Alternatively, it’s also possible that the costs of periodic disasters like Harvey are already baked into car insurance rates, so we’ve already paid for it. Either way, the destruction of so many vehicles hurts millions of people who carry car insurance.

I imagine that most of the owners of Harvey-flooded cars just didn’t realize that their vehicles were parked in places that would flood (you never know until it happens). By the time they saw what was happening, the water level might have been too high to safely walk out to the car and move it to higher ground. Surely there were also cases where owners had parked their cars out of view of their homes, leaving them completely unaware that their cars were flooding. Some people probably didn’t realize flooding could destroy their cars at all.

All that being the case, it occurred to me that autonomous cars programmed with something like “Asimov’s Third Law of Robotics” could have mitigated much of the economic loss.

  1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
  2. A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
  3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Put simply, an autonomous car could protect its own existence from floods by moving to higher ground, such as a hill or multi-level parking garage. Maybe FEMA would broadcast a signal to all cars in an area about to get hit by a hurricane to avoid low-lying areas, or maybe the cars would constantly watch their surroundings even when parked and vacant, and be able to recognize rising floodwaters and move (or to recognize other threats like vandals and burglars).

That sounds great until you think about how many humans would be abruptly deprived of personal transportation during a life-threatening natural disaster. Even if this didn’t lead to any loss of life (e.g. – everyone just waits in their living rooms until the waters recede and the cars come back), I think the feeling of helplessness and anxiety it would engender would be unacceptable. Moreover, according to Asimov’s Second Law, if an autonomous car’s owner has parked it in his driveway and left, doesn’t that count as a standing order from a human to stay in that location until told otherwise?

The solution is for autonomous cars to notify their owners of the impending flood risk and to ask for permission to move to safe ground. A tidy solution, right? I think it’s superior to our current way of doing things, but it opens another can of worms for any person who answers “No.” If you do that, force your car to stay parked in your driveway, and a flood destroys it, will your insurance company hold you liable? Should you be?

Let me pose another answer: Car insurance companies should investigate every such case the same way they investigate accident claims today. A representative would go to the scene to interview the car owner, to hear their side of the story, and to gather evidence. If it became clear that the owner’s poor judgement contributed to the loss of the vehicle, they would be held responsible. Claims adjusters would probably have a lot of time for this sort of thing since autonomous cars would make routine car accidents so much rarer.

Finishing up, how do we deal with Asimov’s First Law during natural disasters? Let’s assume a scenario even more distant in the future, where the average person not only has an autonomous car, but a personal assistant AI, robot butler, and other technologies that together constantly keep track of his location and what’s going on around him, and that are smart enough to talk to him. The machines would serve as the untiring voice of reason, and I think they’d persuade a greater percentage of people to leave town before something like a Hurricane Harvey hit (autonomous cars also won’t get in traffic jams, which will facilitate mass evacuations), and they’d coach the people who chose to remain into preparing better and not taking dangerous risks as the disaster unfolded.

The imperative to protect human life poses interesting questions when we consider situations where the humans refuse to listen to reason and endanger themselves. Consider the stubborn old man with mild dementia who ignores the mandatory evacuation order. Will his machines report him to the police to protect him even if it means possible imprisonment? What if there’s no mandatory evacuation order, but the old man’s poor health and the low-lying location of his house combine to create a high likelihood of death? After so many hours of arguing, will his mechanical butler just hog-tie him, toss him in the autonomous car and send him off for a little vacation? Or will the arguing just continue up until the moment the old man’s mouth is gurgling with floodwater and he goes under? I guess at least the butler would have a record of it all to show he did the best job he could.

It’s clear that more intelligent technologies at least have the potential to save human lives, to empower people, and to mitigate property damage and all forms of waste, but human stupidity and our desire for autonomy will remain powerful countervailing forces for the foreseeable future. One step at a time.