The problem of self-driving cars: AI tech, human error, and data harvesting

Article by John Fitch

Cars have come a long way since the very popular KITT in the ’80s TV series Knight Rider. While we may not be on the verge of seeing self-aware cars like the artificially intelligent KITT, we do appear to be at the dawn of self-driving cars. The technology for AI-driven cars has been around for a couple of years now, and Google and Uber have both experimented with these new vehicles.

But is it really as simple as creating this tech and then rolling it out? Evidently not as many companies have run into a series of problems getting their self-driving technology off the ground. And one of those problems is that they need to rely on us in an emergency. Humans are easily distracted, and if we’re not expecting to take the wheel, our reaction times can be severely slowed down. We can’t exactly be relied upon to take over the wheel without too much warning. Because of this, companies developing self-driving cars are now trying to develop emergency backups that don’t rely on riders.

Google was the first company to catch onto this idea and created a car with full autonomy in 2012 — no pedals or steering wheel — and it looks like they are set to go through with this model once its self-driving company, Waymo, launches in full. And this could be for the best as Delphi, an international automotive industry supplier, reports that cars with full autonomy have improved safety features and allow people who can’t drive to become more mobile.

It’s not just humans that could stand in the way of self-driving cars, as weather conditions could also pose a big problem. Autonomous cars rely on cameras and LiDAR sensors to see surrounding cars and road markings. Test drives have shown that inclement weather could be the downfall of these sensory systems, as excessive rain and snowfall can consume the cameras and radars. So far, Ford is the company to come closest to a solution after creating 3D maps that include information and data about what’s on the roads and what cars should expect from the surrounding topography and road signs. So even though a car may not be able to use road markings, it can use all the data stored in the map to locate itself.

These issues shouldn’t be too difficult for the tech sector to tackle as long as mechanics and engineers continue to develop solutions. However, there are two, potentially bigger, issues that need to be dealt with before a hard-launch: shareability and data privacy. Thankfully, the industry already has an ace up its sleeve for responding to these two pressing issues — the blockchain.

Blockchain technology can guarantee all users of self-driving cars peace of mind that their data is secure, and that it won’t be harvested by the vehicle companies. This tech will also make it easier for riders to share vehicles, as bills can easily be split and contracts shared. And since a third party won’t be required to organize this smart billing, rides won’t end up costing a small fortune.

Blockchain-based payments and contracting (even blockchain based ride-sharing companies) are a logical addition to the advance of self-driving cars. Without blockchain, consumers may not be able to unlock all the potential benefits of self-driving cars. And without the decentralization and transaction-ledger features of blockchain technology, everyone from car manufacturers to ride-sharing providers to cities could harvest data and, compromise user privacy at will. As the blockchain is completely decentralized and can be replicated on a number of nodes, such data can be shared transparently or kept private as needed.

So while self-driving cars need to make sure their AI tech can stand up to some of the more physical challenges of being out on the road, it seems that blockchain can solve some of the digital problems that will emerge as the technology advances.

Interested to learn more about how AI is shaping the world in which we live? Contacts us to learn more about this futuristic tech. We also host executive briefings to help organizations incorporate this new tech.