A Glimpse Into the Future of AI-Driven Electric Cars
Last week, I took part in a radio show on AI electric cars of the future. It was based on an article from the BBC’s TopGear, which used to be my favorite car program, but they changed hosts, so I rarely watch it now.
It got me thinking about how fast things are moving and how much misinformation is out there regarding how and when autonomous driving will be viable (it isn’t yet, regardless of what Elon Musk says), the state of electric vehicles today, the very low probability for hydrogen to replace batteries, and where else AI will be introduced into our cars.
What is particularly interesting and concerning is how fast China is moving with these technologies and how much better, faster, and cheaper its electric cars are, even though it currently has a serious quality control problem.
Let’s talk about the rapid advancements in autonomous driving, the current state of EVs, and the role of AI in the automotive industry. Then, we’ll close with my Product of the Week: an automotive cheat tool my wife bought me that I am impressed with.
Electric Car Status
The pressure on electric car sales has been increasing rapidly of late, mainly because demand for these cars has declined, slowing market growth. This slowing is due to three things: continued quality problems, particularly with Tesla’s new Cybertruck, some of Musk’s political views, and a lack of marketing execution.
As a result, Tesla has taken a hard hit in terms of sales momentum. Given it’s the biggest player, for now, in the electric car market, its problems are having a huge impact on the industry.
But Tesla isn’t alone. Virtually all of the traditional carmakers have had issues pivoting to electrics. Even though electric cars should be automotive appliances that rarely break, they have been breaking a lot — so much so that their repair rate adjusted for volume is far higher than that of internal combustion engine (ICE) cars, resulting in recent surveys indicating that more than half of the current electric car owners want to go back to ICE vehicles.
Additionally, several automakers who indicated they would be 100% electric by 2030 are rethinking this move and pivoting to plug-in hybrid electric vehicles (PHEVs).
However, PHEVs have their own problems as they effectively have two power trains, which means massively increased complexity and highly varied usage models. What I’m getting at is that most of the newer PHEVs with around a 40-mile electric range won’t use their gas engine much, but gas engines need to be run, particularly if they are using ethanol, which can gum up the injectors and cause the engine to fail if the gas gets too old.
Here is a piece of advice: If you have a PHEV, find a gas station that sells gas without ethanol. Use that gas, and you won’t have as big a problem. By the way, the same goes for motorized equipment like snow blowers that are left unused for much of the year.
At the same time, the U.S. and much of Europe agreed to switch from the current plug standards for electric cars to Tesla’s plug standard. This change will open Tesla superchargers to everyone and reduce much of the confusion surrounding which plug goes to which car.
The Tesla standard is now called the North American Charging Standard, or NACS. By 2026, most electric cars in the U.S. should have converted to that standard. Older electric cars can use an adapter, but don’t bother buying one until your car gets the software update that will allow it to talk to a Tesla Supercharger.
By the way, if you haven’t tried both types of chargers, the old U.S. standard required a fee and an ATM-like experience: you first had to enable the charger with an app, and then it might allow your car to charge — although often the chargers were broken. Tesla drivers, on the other hand, could just plug in.
However, Tesla chargers have issues with gas-powered cars parking in the charging spaces, Tesla drivers leaving their vehicles longer than necessary, people breaking the chargers, and theft (which isn’t only a Tesla charger problem).
As I see it, the problem is that we began this shift to EVs before we had the needed technology. Lithium-ion batteries were developed after decades of slow progress in battery technology for electric devices like cell phones, iPods, and laptops. Plus, you might recall many of those batteries catching fire, a problem that continues to plague us.
Lithium and cobalt, both used in current lithium-ion batteries, are challenging to mine. Lithium burns hot enough to melt aluminum and some types of steel, as I discovered when a battery nearly burned my house down after melting the steel screws on its containment chamber. Additionally, lithium batteries break down rapidly when overheated, such as during fast charging.
Big Changes Coming
Over the last 30 years, we have ramped up battery development massively. By 2027, we should have viable alternatives to lithium-ion, including solid-state batteries, sodium-ion batteries, and some new lithium versions (cobalt mining is also pretty nasty).
These coming technologies promise ranges of 750 miles or more, 15-year, 1M-mile service life, and other advantages that would make the need for charging away from your home infinitely less pronounced because how often do you drive over 750 miles?
These new batteries, coupled with ever more powerful chargers, should result in sub-10-minute charging times without the current danger of prematurely aging your batteries. China’s XPeng already has a car with a near 500-mile range at a really attractive price, suggesting it is becoming as significant a threat to existing carmakers as Japan once was. Until now, Tesla has had a considerable range and price advantage in this market.
So electric propulsion will become more convenient and competitive with gas cars, but what about AI?
Autonomous cars will have five levels. Level 5 means they never need a driver. Currently, most cars are hovering at Level 2+, which pretty much entails the ability to avoid things like the car in front of you.
Here is a video that does a nice job of explaining the different levels.
Mercedes has started rolling out cars equipped with Level 3 systems, but only in Nevada and California. Level 4 systems will arrive around 2027, and level 5 systems are due around 2030.
But driving the car isn’t just where AI will play. In other areas, restrictions against using the technology are not tied to whether the technology is mature enough yet to avoid accidents and deaths. In-car AI digital assistants are coming. Nvidia has been developing this technology for some time. In 2021, they showcased what an in-car digital assistant might look like and how it would work.
One of the big announcements with electric and gas cars today is that even within the same brand, the controls may be vastly different. There are advanced features like automatic parking that people don’t use because they either don’t know how to or don’t trust them. Imagine being able to ask the car in natural language to do what you need to be done, like “Please park behind the car on my right,” or “Surprise me and take me to a new place to eat that I’d like.”
Other advantages are that when there is a problem, it could tell you what the problem is and whether it needs immediate attention, unlike the current “fix engine” light that can mean nothing serious or pull the heck over right now and get out of the car.
I watched a video earlier today showing people using jumper cables to try to get a car air conditioner to work. But an in-car AI digital assistant could have said that their air conditioner needed servicing and what that would mean, and it didn’t mean hooking up jumper cables and then driving down the road with them on.
Wouldn’t it be nice to know about problems before you get stuck someplace, or what it means when a light comes on you’ve never seen before? This happened to me yesterday when a light showing a foot and a pedal popped up. It’s meant to suggest you slow down, but that same light can mean different things on other cars. Some people get confused by these lights. Simply being able to ask what the notification means would be incredibly helpful.
Add to that, being able to help you select entertainment content, alert you to approaching dangers (for me right now, forest fire alerts would be really handy), and get information from cars in front of me about approaching hazards like black ice road conditions. If you’ve never experienced black ice, I have. It’s no fun at all, and it isn’t clear that sensors in even autonomous cars can see it yet, so getting an early warning would be incredibly helpful.
Finally, over time, the car would get to know you and better adjust to your unique needs, which might change between times of day and days of the week. For instance, you might like the financial reports on the way to work but comedy or soft jazz coming home. Current car technology will adjust between drivers, but next-generation technology will go much further.
Wrapping Up
I expect most of the current drama about electric vehicles will be fully resolved by 2030, and on higher-end cars well before that.
In addition, the coming wave of AI technology will allow us to create a deeper relationship with our vehicles. Since generative AI is conversational, imagine having discussions with a car on a long drive while the car keeps you far safer than you can keep yourself. AI doesn’t get distracted, have substance abuse problems or sudden medical conditions, and when vehicles can talk to each other, they can avoid human-driven cars that are having issues and potentially reduce accidents by over 90%.
The question becomes: once the car is smart and can drive itself, will you need to own one, or would you instead subscribe to a car service, like Uber, without the driver and save the money you’d spend on your car for other things? The younger generations already avoid driving successfully, so while those of us of a certain age may be tied to ownership, they are not.
I love driving, but I’m also willing to use this new technology so I can enjoy events without worrying I may have had too much to drink (I don’t drink when I’m driving) and enjoy the company as much as those riding with me seem to.
Regardless, the future of electric cars with internal digital assistants is coming, ready or not.
Carly On-board Diagnostic Scanner
I have several on-board diagnostics (OBD) readers, but the problem is that you can’t update them. You need to buy a new reader whenever you get a new car, and even with the reader, the information it supplies doesn’t always tell you how to fix the problem.
Carly, which consists of a wireless OBDII reader, an app, and a subscription, can automatically fix some things, provide access to features that typically available only to dealers, and, with its advanced service, offer information on how to fix the problems it identifies.
You can enable Carly for a single car or all brands for around $76 with the premium package. Right now, that package is $10 cheaper than the basic package for a promotion. To cover all brands, it is around $34 more — allowing you to configure your car and read codes. But if you want help understanding the codes, the cost jumps another $36.
On top of this, you’ll need to pay for a subscription that can vary between $9 and $20 a month, depending on how much service you get. Still, the information provided to the home mechanic is invaluable and, so far, well worth the expense because mechanic costs are pretty steep these days. So Carly can save you a ton of money.
Carly enabled me to activate features on my Audi TTS that I didn’t know existed. On my Audi E-Tron GT (electric), it not only enabled some new features but also helped me diagnose and correct an issue with a sensor that was unplugged during wrapping, something that would have otherwise taken hours to diagnose and fix.
Carly is the smartest OBDII reader I’ve found, and likely a precursor to one of the AI features I expect to come with the approaching wave of AI-driven electric cars, making it my Product of the Week.
Editor’s Note: The images featured in the Product of the Week section are credited to Carly Solutions.