What Even Is a Self-Driving Car? Busting the Hollywood Myths First

If you’ve ever watched a sci-fi blockbuster where a sleek, self-driving car zooms through bumper-to-bumper traffic, dodges flying debris, and even takes a detour through the air to avoid a collision, you’re not alone. For decades, Hollywood has sold us a very specific, very flashy version of autonomous driving: indestructible, all-knowing, and ready to replace your human driver the second you step into it. From the driverless taxis that outrun villains in *Spider-Man: Homecoming* to the rogue, hacker-controlled autonomous fleet that causes chaos across the globe in *Fast & Furious 8*, the message is always the same: self-driving cars are already here, they’re cool, and they can do pretty much anything.

It’s a fun premise for a movie, but it’s also wildly misleading. If you walked out of a theater today and went to a car dealership asking for a “self-driving car” that matches what you saw on screen, the salesperson would probably laugh you out of the room. The reality of autonomous driving is far less flashy, far more incremental, and far more… well, boring, in the best way possible. It’s not a magic switch that flips from “human drives” to “car drives itself” overnight. It’s a staircase, one step at a time, and we’re only a few steps up right now.

Let’s start by tearing down the most common myths first, then break down what the different “levels” of autonomous driving actually mean, in plain English, no engineering degree required.

First, let’s name the biggest Hollywood lies you’ve been sold

Before we get into the technical details, let’s get the obvious out of the way. The self-driving cars you see in movies almost never match real-world tech for three big reasons: 1. They’re built for drama, not accuracy. A self-driving car that safely obeys speed limits and stops at stop signs doesn’t make for a thrilling car chase. Movie makers add flying capabilities, indestructible armor, and hacker-proof systems because they serve the plot, not because they’re technically feasible. For the record: no self-driving car in existence can fly, no production model has bulletproof plating, and every system ever built can be hacked if someone tries hard enough. 2. They skip the boring, unglamorous work. Movies never show the years of mapping every inch of road, testing millions of miles of edge cases, or fixing the thousands of tiny glitches that make the difference between a car that stops at a stop sign and a car that thinks a plastic bag blowing across the road is a pedestrian and slams on its brakes. That unsexy, behind-the-scenes work is 99% of what autonomous driving development actually is. 3. They erase the messy middle. Almost every movie set in the future assumes we’ve already jumped straight to fully driverless cars (what the industry calls Level 5 autonomy) with no messy in-between phase. In reality, we’re living in that messy middle right now, and we’re going to be here for at least another decade or two.

Even outside of Hollywood, automaker marketing has done us no favors. You’ve probably seen ads where a driver is napping, reading a book, or even sitting in the backseat of a car while the “self-driving system” handles the highway. Those ads are, at best, misleading, and at worst, dangerous. In the U.S., the National Highway Traffic Safety Administration (NHTSA) has repeatedly fined automakers for using terms like “autopilot” or “self-driving” in marketing materials that imply drivers can take their hands off the wheel, when the systems are only designed to assist a fully attentive human driver.

---

The autonomy levels: not a switch, a staircase

In 2014, the Society of Automotive Engineers (SAE), the global body that sets technical standards for the auto industry, released a standardized framework for autonomous driving levels, ranging from 0 (no automation) to 5 (full automation, no human needed ever). These levels are the universal language the industry uses to talk about how much a car can do on its own, and they’re the only framework you need to understand to cut through all the marketing hype.

Let’s walk up the staircase one step at a time, with real-world examples you can actually buy or ride in today.

Level 0: No automation

This is every car built before the 1990s, and plenty of cheap, base-model cars sold today. The human driver does 100% of the work: steering, accelerating, braking, watching for pedestrians, reacting to stop signs. The car might have safety features like anti-lock brakes (ABS) or electronic stability control (ESP) that step in to prevent crashes in emergencies, but those features don’t actually control the car’s direction or speed for you—they just kick in if you mess up. You are 100% responsible for everything that happens behind the wheel.

Level 1: Single-function driver assistance

This is the first step up the staircase, and you’ve almost certainly used features that fall into this category. Level 1 systems handle one single driving task at a time, while you do everything else. The most common examples are:

• **Adaptive cruise control (ACC)**: The car maintains a set distance from the car in front of you, speeding up or slowing down automatically, but you still have to steer.
• **Lane-keep assist (LKA)**: The car gently nudges the steering wheel if you drift out of your lane, but you still have to do all the steering.
• **Automatic emergency braking (AEB)**: The car slams on the brakes if it detects you’re about to hit something, but you still have to do all the accelerating and steering.

You can’t combine these features on most Level 1 cars—if you turn on adaptive cruise, you still have to steer, and if you turn on lane-keep, you still have to control your speed. You’re still responsible for 99% of the driving. The only Level 1 feature that’s standard on almost all new cars sold in the U.S., EU, and China today is automatic emergency braking, which has been shown to reduce rear-end crashes by 50% according to IIHS data.

Level 2: Partial driving assistance (this is almost certainly what you have if you bought a new car in the last 4 years)

Level 2 is the first level where the car can handle two driving tasks at the same time: steering *and* accelerating/braking. The most common Level 2 feature package combines adaptive cruise control with lane-centering (a more advanced version of lane-keep that keeps you centered in your lane, not just nudges you back when you drift) to let the car handle both speed and steering on highways or in stop-and-go traffic.

This is the level that has caused the most confusion for consumers, because automaker marketing has spent years calling these systems “Autopilot” (Tesla), “ProPILOT” (Nissan), “Honda Sensing” (Honda), “Toyota Safety Sense” (Toyota), and dozens of other brand names that imply the car can drive itself. It can’t.

The key rule of Level 2, no matter what the automaker calls it: you are still the driver, you are still fully responsible for everything that happens, and you must keep your hands on the wheel and your eyes on the road at all times. If you take your hands off the wheel for more than a few seconds, almost all Level 2 systems will beep at you, flash a warning, and if you still don’t respond, slowly slow down and pull over to the side of the road to stop.

To put this in perspective: if you’re driving a 2024 Toyota Camry with Toyota Safety Sense, set adaptive cruise to 70 mph, and take your hands off the wheel to text your friend, you’re breaking the law in all 50 U.S. states, and if you crash, you’re 100% liable for all damages and injuries. The car is just a tool to help you, not a replacement for you.

A 2022 study by the AAA Foundation for Traffic Safety found that 60% of drivers who own cars with Level 2 features believe it’s safe to take their hands off the wheel while using them, and 25% believe it’s safe to take their eyes off the road. Those beliefs are wrong, and they’re the cause of dozens of crashes every ye