in , , , ,

Solving Big City Murder Cases And Other Crimes Via AI And Self-Driving Cars

The nationwide murder rate is heading in the wrong direction.

It is decidedly on the rise, rather than on the decline.

According to the latest news reports, the number of murders in the U.S. rose last year, increasing by about 40% in the country’s ten largest police districts over the prior year. Resources to investigate those murders are being stretched tenuously thin. Meanwhile, family and friends of the murdered victims become overtly frustrated and extremely exasperated over the delays in solving these homicides. The public at large also remains on edge as to whether a murderer is still loose within their community and ought to have already been captured and detained.

One potential new gizmo in the detective toolkit for undertaking murder investigations is something only now starting to appear. It is an innovation of technology that has already been predicted as a wide benefit for society in many ways, though few have considered the potential role in solving crimes. This new technology has still a long way to go before being fully ready for active use, and will undoubtedly take many years to be widely deployed, yet nonetheless has a multitude of uses for the betterment of us all.

What is that new technology?

Surprisingly, for some, the means of aiding the catching of criminals such as murderers could be spurred via the advent of true self-driving cars.

Yes, that’s right, the emergence of self-driving cars has all the makings of being a kind of modern-day gumshoe that can be a devout crime fighter. Note that self-driving cars aren’t going to leap tall buildings and nor do they need to drive faster than a speeding bullet. There are other ways in which a plethora of self-driving cars will be part of the law enforcement kit and caboodle.

The future of a world with self-driving cars is supposed to reduce the annual number of car crashes and car accidents, both large and small, dramatically ratcheting down the count of those yearly 40,000 fatalities and 2.3 million injuries that occur in the United States alone. Furthermore, the hope is that self-driving cars will spark a mobility-for-all era. Those that are currently mobility constrained or disadvantaged will finally have ready access to mobility at a presumed low cost and plentiful availability.

Seldom considered is the fact that self-driving cars can also be a vehicle toward serving justice, truth, and the American way (well, got a few puns in there, but anyway, the point is hopefully well-taken).

Here is the question worthy of rapt attention: Will the advent of AI-based true self-driving cars aid in solving murders and other crimes, acting on behalf of the law, and in pursuit of justice?

Let’s unpack the matter and see.

Understanding The Levels Of Self-Driving Cars

As a clarification, true self-driving cars are ones that the AI drives the car entirely on its own and there isn’t any human assistance during the driving task.

These driverless vehicles are considered a Level 4 and Level 5 (see my explanation at this link here), while a car that requires a human driver to co-share the driving effort is usually considered at a Level 2 or Level 3. The cars that co-share the driving task are described as being semi-autonomous, and typically contain a variety of automated add-on’s that are referred to as ADAS (Advanced Driver-Assistance Systems).

There is not yet a true self-driving car at Level 5, which we don’t yet even know if this will be possible to achieve, and nor how long it will take to get there.

Meanwhile, the Level 4 efforts are gradually trying to get some traction by undergoing very narrow and selective public roadway trials, though there is controversy over whether this testing should be allowed per se (we are all life-or-death guinea pigs in an experiment taking place on our highways and byways, some contend, see my coverage at this link here).

Since semi-autonomous cars require a human driver, the adoption of those types of cars won’t be markedly different than driving conventional vehicles, so there’s not much new per se to cover about them on this topic (though, as you’ll see in a moment, the points next made are generally applicable).

For semi-autonomous cars, it is important that the public needs to be forewarned about a disturbing aspect that’s been arising lately, namely that despite those human drivers that keep posting videos of themselves falling asleep at the wheel of a Level 2 or Level 3 car, we all need to avoid being misled into believing that the driver can take away their attention from the driving task while driving a semi-autonomous car.

You are the responsible party for the driving actions of the vehicle, regardless of how much automation might be tossed into a Level 2 or Level 3.

Self-Driving Cars As The Detective On The Beat

For Level 4 and Level 5 true self-driving vehicles, there won’t be a human driver involved in the driving task.

All occupants will be passengers.

The AI is doing the driving.

Your first thought might be that other than the nicety of not having to yourself be at the wheel of a car anymore, there doesn’t seem to be much in the way of crime-fighting skills as part of an AI-based driving system. In that respect, you would be correct that the AI itself is not going to be wearing a badge and will not be the local on-the-beat detective per se.

Here’s how self-driving cars are going to become a roadway warrior in the war on crime.

Imagine that a murder has just taken place on a street corner in the downtown area. A lone gunman opted to shoot someone. The murderer rushes off.

Normally, by the time that the authorities arrive, the perpetrator is long gone. One of the first actions to undertake involves trying to find witnesses that can attest to what happened. What did the murderer look like? Which way did the murderer go? Did anyone recognize the brazen killer? And so on.

Shift gears and envision a future whereby there are self-driving cars aplenty on our byways and highways.

First, keep in mind that true self-driving cars are chockful of specialized sensory devices, such as video cameras, radar, LIDAR, ultrasonic units, thermal imaging, and the like. These highly tuned sensors are used as the eyes and ears of the vehicle. The AI collects data about the driving environment surrounding the car via these sensors and plots out the means to navigate and operate the driving controls to safely proceed.

There is a somewhat unexpected byproduct to all that sensory equipment which bodes for both great advantages and sadly also great consternation. Besides the sensory data being used to drive the vehicle, there is also the possibility of using that collected data for other purposes. I’ve referred in my columns to this ancillary usage as the “roving eye” and discussed the merits and drawbacks accordingly (see this link here).

Here’s what this data allows.

When a self-driving car drives down a city block, the sensory devices are voluminously capturing everything within their designated range. The video cameras are collecting visual sightings of the street, people on the sidewalk, buildings adjacent to the street, and so on. All of the sensors are doing their thing, soaking up whatever they scan. This data gets stored on-board the self-driving car and can be kept inside the in-car computers and various electronic storage capabilities.

At some point, the self-driving car can perform an OTA (Over-The-Air) electronic communication, typically connecting with the cloud of the automaker or self-driving tech firm that devised the AI system. The most widely known use of the OTA involves pushing down patches or updates to the AI driving system of the vehicle. This can allow for example the capability to remotely update aspects of the vehicle such as the braking facility or perhaps used to tune the AI to improve making tricky left turns.

As they say, what comes down can also go up (not everyone says it that way, but you know what I mean).

The data that has been collected by the self-driving car during its travels can be uploaded into the cloud during the OTA process. If so desired, this could include every nook and cranny of whatever the sensors captured during the driving journey. Assume that the self-driving car was driving around for several hours. This implies that there are hours of a roving eye that has been like a souped-up video camera, one that has been moving around and not just planted in one spot. Also, besides the video data, there is the radar, LIDAR, ultrasonic, thermal images, and a slew of other assorted stuff.

Suppose a self-driving car drove down your neighborhood street, perhaps doing so on its way to pick-up someone that had requested a ridesharing lift. You were outside of your house, standing on the front lawn, perhaps playing kickball with your son and daughter. Without you directly realizing that you were being viewed, the self-driving car managed to record a video of you and your kids, doing so as the vehicle simply drove down the street.

This video gets uploaded into the cloud.

You might be thinking that you don’t care and cannot envision why you should.

Realize that at this point, the OTA has uploaded the video and other sensory data, including the GPS positioning indications, as part of the upload. Inspection of the cloud-based data would tie together that at that street address there was an adult and two children. If you then correlate that data with other demographic info, it sure might be interesting to a firm selling toys to know that possibly someone living at that particular house has two small children and they like to play kickball.

This data could be a goldmine for the automaker or self-driving tech firm that is collecting it from the self-driving cars.

Amp this up.

Over a period of a week, each time a self-driving car goes down your block, this data gets uploaded. One day you were out mowing your lawn. Another day you were planting pretty flowers. By examining the uploaded data over time, a fuller semblance of what you do, how you live your life, begins to become more apparent.

That’s based on just one self-driving car that perchance comes down your street on an occasional basis.

Imagine that there are hundreds, perhaps thousands, or eventually millions of self-driving cars on our streets (there are about 250 million conventional cars in the U.S. today, and the expectation is that by-and-large they will inevitably be replaced with self-driving cars). The odds are that throughout the day and night, there are going to be self-driving cars all around us, and relatively continuously cruising down your street and other streets.

By stitching together this data from those zillions of self-driving cars, it would be possible to piece together an overall semblance of what takes place daily in our world.

The good news is that this is how crime-fighting comes to the fore.

Return to the earlier scenario about the murder that took place on a street corner in the downtown area.

Rather than trying to find a happenstance of a video camera that was perched on a nearby building, an astute place to look for any captured data about the incident would be in the uploaded data from self-driving cars.

Assuming that self-driving cars are going to be always on the go, rolling around 24×7, which makes sense since they are potential money makers and only able to make money once they are transporting someone or something, the odds are high that the murder scene came within the purview of a self-driving car during the murderous act.

It could be that several self-driving cars wandered past that murder scene and caught the whole act from multiple angles. If you piece it all together, this could offer an uncanny portrayal of the murder from start to finish, and from a multitude of perspectives. One perspective might only show the back of the murderer, while another self-driving car got a glimpse of the front and the face of the killer. Think of those video editing special effects these days that can rotate in a 3D fashion around a fictional portrayal of a fight or other action and think about how this can be done for real-life incidents.

Okay, let’s switch into a crime-fighter mode.

We would check the uploaded data to stitch together the crime scene at the moment of the murder. This could likely showcase how the murder took place, along with providing an exact timestamp of when it happened.

Essentially, we have an electronic eyewitness to the murder.

Next, we could inspect the cloud data to see if we can trace where the murderer came from, before reaching the street corner and committing the crime. This might help figure out where the killer lives or maybe was at a restaurant or other place that they tend to frequent and might be known by others.

Remember that the murderer fled the scene.

We can inspect the collected cloud data to see where the murderer went. If he ran to another block and went into an apartment complex, there is a solid chance that this was also seen or detected via a multitude of self-driving cars that were perchance roaming around or making trips in that area.

It might also be useful to use the data to trace where the victim came from and how they arrived at the point of the murderous act.

In this instance, we are assuming that the victim fell right there, upon being killed, but suppose the murderer dragged the body or tried to do some kind of cover-up. Again, the data from a multitude of passing self-driving cars might reveal this aspect of the crime.

So far, we have been using the data solely to figure out the crime and sought to take a look at the killer and the victim.

Let’s expand our approach.

In the collected video, there were other people on the sidewalk and they seemed to have witnessed the killing. A person inside a store nearby was looking straight out the store window and appeared to see the murder take place.

These are all potential witnesses that can further help in solving the crime, including finding the murderer.

We can invoke facial recognition software to see if we can figure out who those people are. It is conceivable they are already posted somewhere on the Internet via social media or in other databases. This provides us with their names and contact information.

You could go to the scene and try to talk with these witnesses. Another possibility would be to contact them and ask if they would be willing to come to the police station to share what they know about the murder and what they saw happen.

Some of them might be willing and eager to help, but might not readily have transportation available to come to the police station. No problem. You send them a self-driving car that picks them up and provides them a lift to the police station, and will likewise give them a ride home afterward.

You decide that it would make sense to go to the crime scene and then go throughout the nearby area to do a canvas. Using a self-driving car, you head over to the scene. There is no need for you to be driving the vehicle and therefore can use the time while inside the self-driving car to be further analyzing the case. Driving is otherwise a distraction and instead, you can devote your full attention to the murder case.

Here’s the final tipping point on solving the murder.

From the data collected, you have an image of the killer that is relatively unobscured. Attempts to use facial recognition did not identify who the person is and apparently, they have not been previously spotted or otherwise posted their image anywhere online.

No problem.

All those self-driving cars that are crisscrossing the city on a moment-to-moment basis are now deputized as part of your murder investigation. You merely indicate that at any time, if any of those zillions of continually on-the-go vehicles spot that person, the face of the apparent murderer, the AI driving system should immediately send an alert to 911.

If that person meanders anywhere in the city, they are going to get nabbed.

For the moment, we can have pleasant thoughts that those that heinously wish to commit murders are going to have self-driving cars on their case. Those roving Sherlock Holmes will make it harder for murderers and other criminals to escape the hands of justice.


There are a bunch of caveats to keep in mind on this gumshoe self-driving car scenario.

Ponder these thorny questions:

·        Will the automakers and self-driving tech firms allow law enforcement to have access to the data being collected?

·        Will laws be enacted to force the data to be made available to law enforcement?

·        Will the data on-board the vehicles be uploaded, or will the data instead be kept only inside the self-driving car and not uploaded?

·        Will only some of the data be uploaded, selectively, diminishing the opportunity to use it in the fashion depicted herein?

·        Will the data be deleted immediately upon use for the act of driving the self-driving car and otherwise not be accessible beyond the moment that it was needed?

·        Etc.

Realize too that there are going to be a multitude of self-driving car brands and models.

This suggests that trying to tie together the disparate data and formats of data is going to be somewhat arduous. Yes, that’s a bit of a hurdle, but in the end, if the data is made available across these differing clouds, it will be technologically feasible to do the needed conversions and matchings. Of course, the computational costs to do so might be rather prohibitive, so we’ll need to see if this is the kind of usage that will get the financial backing needed to bring it to fruition.

As a final thought, we certainly should look beyond the assumed altruistic use of this data for exclusively undertaking the virtuous notion of fighting crime. You probably got chills up your spine since this facility has a dual-edged sword to it. Though it can be used as a crime fighter, there is an equally looming possibility of using it to invade our privacy and pursue those that are innocents.

We are in the midst of bringing to fruition self-driving cars, ostensibly a great asset that can benefit humanity and that will simultaneously provide a means to undercut humanity. Darned if this always seems to be the case with most innovations.

I realize this all seems like some bizarre future world of unimaginable nature, but do realize that everything described herein can be done with today’s technology.

That being said, as mentioned at the start, there aren’t enough self-driving cars on the roadways today to make this especially workable. On the other hand, once self-driving cars start to become viable, you can anticipate that a huge and frenetic race of sorts will take place to deploy as many of them as possible, as rapidly as possible (under the presumption that they are money makers on wheels, plus these vehicles offer the societal benefits of lessened lives lost and might fulfill the mobility-for-all potential).

Is this advent of self-driving cars and their use as amateur sleuths going to really happen?

Please take a moment and carefully put on your detective’s cap.

As a great fictional detective once emphasized, when you have eliminated the impossible, whatever remains, however improbable, must be the truth.

What do you think?

How to be less awkward on Zoom calls: 11 tips

Google Cloud Launches Vertex AI, A Machine Learning Platform