Tesla's full self-driving AI features have proven controversial as of late. With multiple high-profile videos showing the electric cars moving towards humans, many are losing confidence in the technology. However, one YouTuber has set out to ease minds with a new Tesla experiment.
Tesla Experiment: Will it hit a cat?
YouTuber Mat Watson of CarWow released a 12-minute Tesla experiment last week looking to answer one simple question. Will a Tesla's full self-driving AI let a car run over a helpless little kitty cat? Or will it brake to avoid one of the planet’s most adored animals?
For the test, Watson had two cars: a Tesla Model 3 and a Volvo V90. The YouTuber tested both against a variety of obstacles, from foam cars to taxidermy cats. As it turns out, both vehicles performed better than expected.
For example, both self-driving cars braked correctly when faced with a fake car. Afterwards, Watson put a cardboard cutout of Tesla CEO Elon Musk on the road, asking: “Will the Tesla kill Elon?” The answer: no, it will not turn Elon into roadkill.
Both cars performed well, breaking before smashing Elon Musk. However, the results are interesting. The Model 3 braked sooner but still had less distance between obstacle and car than the Volvo. Additionally, the Volvo also tightened Watson’s seatbelt, protecting him as well as the pedestrian.
Watson then tested to see if the cars would hit large animals, represented by a large kangaroo plush. They didn't. However, smaller animals did face the full power of the vehicles. Taxidermy dogs and cats were used, and they did not survive.
Read More: Tesla engineers claim Elon Musk is misleading the public on Autopilot dangers
So, are they safe?
For the most part, this Tesla experiment does prove that the self-driving cars will not hit a pedestrian in a test scenario. However, it's important to note that the FSD AI is still in a beta form, and it is far from perfect.
Videos of Teslas swerving towards oncoming traffic and pedestrians or being confused by the moon aren't uncommon. However, it is unfinished software, and drivers are not expected to use it without being ready to take over the controls.
So, yes, it’s fairly safe to use if you're paying attention, but it's still far from the sci-fi dream of full self-driving. For example, if you're intoxicated after a night out, you definitely shouldn't trust the AI to take you home safely.
Maybe, one day, self-driving cars will be flawless, but currently they're not. So, if you do own a car with automatic driving capabilities, stay aware while using them.