Self-driving carshave been heralded as a new dawn for road safety. Enthusiasts promise a future with a dramatically lower number of auto accidents and related fatalities. A future where everyone is mobile, irrespective of health or age.
It seems far-fetched to imagine a city in which all vehicles are operated by
artificial intelligence(AI), and can zip around with perfect, automated efficiency. However, this is exactly what electric car makers like Tesla are working towards.
Aside from preventing traffic accidents, there is another opportunity for self-driving car technology to make our cities safer, which is not being talked about as much.
In the future, could self-driving cars, and the AI used to power them, help law enforcement to solve major crimes like murder? Some people certainly think so.
What is a self-driving car?
Forbescontributor Lance Eliot recently delved into the crime fighting potential of AI and self-driving cars. Prior to explaining his vision, he makes it clear that he is discussing _true _self-driving cars.
That is, fully autonomous vehicles which don’t need human intervention to safely drive around. These cars do not exist yet, or at least they are not yet available to buy.
Most new cars have some autonomous driving abilities, like lane-keeping assistance which can keep you centred, or automated parallel parking, but human drivers are still required.
There are other, more advanced automated vehicles which can operate by themselves. But they are still undergoing testing, and will require big improvements to the AI before becoming commercially available, let alone widespread.
The self-driving cars envisioned by Eliot will probably become a reality, but are still many years away.
The artificial intelligence behind crime-fighting technology
The crime fighting capabilities of self-driving cars will come down to the technology they use to get around. Namely, tons of cameras, radar, thermal imaging, and other sensors designed to stop the vehicle from crashing.
Some of this technology is already employed. For example, most new cars include backup cameras, and many also have built-in dashcams. They already help police investigators in some cases, like showing who is
at fault for traffic accidents.
With a fully AI powered car of the future, these features will be the eyes and ears of the vehicle, enabling it to safely navigate busy city streets.
All of the information captured by these cars would be uploaded by OTA (Over-The-Air) electronic communication to cloud storage, and used to further refine the AI technology. And it is precisely this kind of data which may be of use to police.
How can self-driving cars be used to solve crimes?
Eliot imagines a scene in which a murder takes place on a crowded sidewalk, and the killer runs away before he can be apprehended.
If a vehicle "sees" this happen, a sophisticated AI system could recognize the violent act and even call 911 before a human witness has time to react.
Today, detectives would canvas the area for witnesses or store front CCTV. In the future, they could tap into the data that has been captured by our cars’ sensory devices, almost as soon as the crime happens.
With an integrated system, it would be fairly easy for law enforcement to review a time-stamped map and find all vehicles which were on the street at the time of the murder.
Once they have identified the cars that were nearby, they could access the cloud which holds the data collected during the OTA process, and review the stored footage.
On a busy street, several different cars might have recorded the act, giving cops the benefit of multiple different angles. As Eliot writes, self-driving cars are "electronic eyewitnesses to the murder."
There is nowhere to hide in an automated world
The network of electronic police informants could also be expanded, and used to trace the murderer’s movements before and after the crime. Investigators would follow a patchwork of photos and videos captured by cars in the surrounding area.
If, for some reason, the criminal has managed to evade these roaming cameras during his escape, police can use facial recognition software to find out his name from social media.
Going one step further, a photo of the killer’s face could be uploaded to a crime fighting database, and if a camera spots him in the future, the AI system could immediately alert authorities to his whereabouts.
The suspect could never leave home again! Even a mask might not be enough to help him, as sophisticated facial recognition software can detect a match with much of the face covered.
As explained by Lance Eliot in Forbes, there are many caveats that could render this scenario complete fantasy.
For example, will lawmakers allow cops to access this kind of sensitive data. Will electric car makers allow it? Will the data be stored for long enough to be of any use?
Based on the increasingly milquetoast protections currently afforded to personal privacy, the answer to these questions is probably "yes!"
Interestingly, there is no question about the possibility of the technology being available in the future, as much of it is already available today. It will be for future politicians to decide whether or not the invasion of privacy is worth the potential for solving murders and other serious crimes.