There’s A Baby Sea Lion On Highway 101, What Would A Self-Driving Car Do?

There was a bit of extra excitement in the San Francisco Bay Area traffic last year when a baby sea lion decided it was time to checkout rush hour activities and so wandered or waddled onto the hectic and perilous Highway 101 (a similar occurrence happened a few years ago in northward Mendocino County).

In case you aren’t familiar with California drivers and their predilections, during rush hour they are famous for a stern take-no-prisoners brashness when it comes to dealing with traffic conflagrations.

Dare anything or anybody to get in their way, which worsens an already worst-case clog that accompanies the dreadful rush hour traffic, you’ll see hotheaded road rage acts and oftentimes bizarre attempts to push or squeeze their way around traffic.

What happened to the doe-eyed innocent sea lion that was merely curious about all the hubbub on the busy highway?

In a moment of unplanned humanity, some drivers purposely blocked traffic so that the sea lion pup would not get run over, and others got out of their cars to shoo the intruder over to the side of the road.

This act gives hope for the world, it would seem, and perhaps bolsters our survival chances on this planet.

Apparently, one enterprising good Samaritan used a handkerchief, waving it seemingly like you might at the Pamplona bull run, hoping it would inspire the wayward mammal into getting out of harm’s way.

Ultimately, a California Highway Patrol (CHP) car showed-up and the officers did more than just try to herd the creature off the busy road, they opened the backdoor of their patrol car and the pup leaped right into the backseat, not needing a baton or handcuffs as inducements.

The sea lion then got a free ride to the local animal humane society, which though the journey didn’t also get the mammal the archetypal sirens and flashing lights treatment, at least the pup was able to get some rest and relaxation after what must have been a quite harrowing experience.

Just think of the stories that the sea lion could tell when eventually being returned to its colony or raft.

Perhaps the other sea lions wouldn’t even believe the story and assume that it is yet another tall tale by a young pup with a vivid imagination.

Here’s an interesting question for you: What would a self-driving driverless car have done?

Suppose that rather than the traffic being composed of human driven cars, instead it was a slew of autonomous self-driving cars. Would the self-driving cars realize the nature of the situation, or might they just keep on rolling along?

It’s instructive to consider such a use case and reveal insights about the design and development of true self-driving cars.

Edge Cases For Autonomous Cars

Some pundits might decry the situation as a once-in-blue-moon circumstance and therefore it is not of importance and nor considered of any notable consequence to study. This is no more than an edge or corner case, they say, meaning that it is rare and therefore not something to be dealing with now.

They would also say that we should be focusing instead on self-driving cars becoming proficient in everyday driving aspects, and later on come back to the oddball instances.

How many times will any of us ever experience a sea lion on a busy roadway?

Admittedly, likely never, but then again you can logically and readily generalize the act of the sea lion on the highway to some other kind of animal.

The news is replete with stories of animals encroaching upon our streets and highways in the midst of car traffic, including tales about waddling ducks, lumbering bears, slithering crocodiles, and of course everyday dogs and cats.

If you are willing to agree that the sea lion is emblematic of a much larger set of possibilities, this hopefully is persuasive that perhaps this instance is not akin to a meteor striking earth or some other one-in-a-million chance of happening.

The odds seem high enough, along with the sad and dire consequence to the meddlesome animal if it were to get run over, we can reasonably argue that emerging self-driving cars ought to be prepared to deal with this type of situation.

Idealized Futuristic Handling By Autonomous Cars

In theory, eventually, we’ll have self-driving cars that are interconnected with each other, doing so via V2V (vehicle-to-vehicle) electronic communications.

The value of having V2V is that the self-driving cars will be able to share roadway status and provide cautions and insights to each other. Maybe there is a sofa that fell onto the freeway and the first driverless car that spots it would quickly broadcast to any nearby self-driving cars to be wary of the debris in the roadway.

In the case of the sea lion, presumably whichever autonomous car was the first to spot the wanderer, the AI using V2V would beam out a warning to the rest of traffic. Once the warning was received, the other nearby self-driving cars could opt to slowdown as they approached the location of the sea lion.

Slowing down is an easy step to take.

Would the AI of an autonomous car also determine that it might make sense to stop the self-driving car and block other traffic, upon having gotten near to the animal?

You might say that there would be no need to do so, assuming that all other traffic has gotten the V2V messaging and cautiously is maneuvering around the sea lion.

But, it is possible that some of the other self-driving cars might not have gotten the electronic message, and if there are any human driven cars in the mix, perhaps not equipped with the V2V, the effort of a self-driving car acting to block the road might be sensible to contend with any such loopholes.

In a twist that might surprise you, would the human passengers inside any of these nearby self-driving cars be able to possibly override the self-driving car in terms of the AI deciding to bring the driverless car to a halt?

Let’s first consider the other side of that coin, namely what about a human passenger urging the autonomous car to come to a stop.

Consider that maybe a human passenger inside a self-driving car might see the sea lion, possibly doing so before the sensors of an autonomous car did, and the human passenger might urge the AI to stop the car and block the road.

This though is somewhat problematic, because suppose the human passenger uttered such a command and yet there wasn’t any bona fide reason for the driverless car to come to a stop in the middle of traffic? The human could be drunk, they could be mistaken, they could be plain crazy, etc.

Do we want AI systems that obediently and unquestionably abide by whatever the human passenger tells them to do?

This brings us back to my other question, the one dealing with the notion that a human passenger might command the AI to not stop the self-driving car, in spite of a sea lion having been detected by the self-driving car.

You might instantly be angry that any human could be so uncaring, but suppose they were urgently trying to get to a hospital for themselves or had some other high priority need.

How is the AI to decide which is right, whether its own programmed automation is “right” to stop for the sea lion, or whether the human passenger is “right” to have the self-driving car keep going, presumably attempting to avoid hitting the sea lion and yet not helping to protect or save the sea lion?

Do not be fooled by whatever exaggerated claims you might read about in the media and assume that the AI of self-driving cars will have human-like common sense reasoning, they won’t for a very long time (if ever, some say).

Conclusion

I’ve brought up this enchanting story of the meandering sea lion to illuminate an even larger topic overall.

The advent of self-driving cars means that we will have non-human drivers, AI systems, making human-like choices while undertaking the driving task.

Today, we assume that human drivers will take societal and legally appropriate actions when driving and will ultimately be held accountable when they don’t.

This is a serious and unresolved matter for self-driving cars that have no such built-in moral code and guidance. A plethora of ethical questions abound when it comes to the choices to be made when driving a car.

These kind of choices are often likened to the famous Trolley Problem, a philosophical thought experiment about having to make hard life-or-death choices. Currently, automakers and tech firms have no specific guidance on what self-driving cars are supposed to do in these situations.

As a society, we’ll need to find some amenable means to figure out the manner in which these AI self-driving cars will make key societal decisions that impact us all.

Just think about those baby sea lions.