Autonomous Tech Stalled Through Lack of Innovation

Autonomous cars were expected to be the next big thing in mobility but, after endless hype for their potential and broken promises of imminent deployment, we don’t seem to be any closer.

The last few years have been muddled by both contraction and consolidation in AV development. Aurora Innovation, which acquired Uber’s self-driving car division in 2020, has been burning through its cash and may run out of funds by mid-2024. While the technology could be pivotal to the future of ride-hail services, Lyft also sold off its AV division. Ford and Volkswagen had collectively injected more than $3Bn into Argo AI but they pulled the plug on that venture in October 2022.

Autonomous tech development, as well as research into ADAS and other safety tech, continues on at automakers, start-ups and many other enterprises. Given the recent setbacks and the lack of notable advancements in recent years, it’s getting harder to envision a near-term future for self-driving vehicles.

Early success raised the bar too quickly

Mike Ramsey, vice-president and analyst in automotive and smart mobility at Gartner, attributes the pie-in-the-sky expectations to the quick progress that occurred during the early days of AV research and development. He said now that the dust has settled, automakers and autonomous tech companies remain focused on tackling big and distinct obstacles but no current solutions are ready for commercialization.

For example, algorithms have become better but they are not capable of making decisions in all environments. Sensors, another important component, have improved but have yet to provide the drive-anywhere solution that full autonomy requires. “Then you have regulation, which is still completely spotty,” said Ramsey, noting that the regulatory landscape hasn’t evolved much over the last five years. “There just simply isn’t any regulation. In Europe they’ve moved a little bit further along on this but not in the US. In many other countries, it’s totally unclear from a licensing, registering and simply operating and insuring standpoint. Are these vehicles legal, can they operate and what’s the process for getting them approved?” Consequently, Ramsey said that no matter how much progress is made technologically, the industry can’t move forward because the regulations aren’t there to dictate what’s needed to operate safely in most locations.

Jennifer Dukarski, an attorney and shareholder at Butzel Long, concurred with Ramsey’s assessment. She feels that AV makers have yet to find a good way to show the National Highway Traffic Safety Administration that they are safe to be on the road in a large volume. “We’re still a ways out,” said Dukarski. “Right now, the sum of the parts is better than the whole. When you have it in its elemental form, when you’re just using it in lane keep assist or in detection, you’ve got some systems that are starting to prove themselves. When you start bundling them, running them all into one vehicle, it’s really challenging because the examples we’ve seen have all had their major issues.”

The business case for autonomy

Ramsey said there is also one big obstacle that no one likes to discuss: is there actually a business case for autonomy? “Does it make sense to do this in any way?” he questioned. “It does from a partial autonomy standpoint. The systems that give me hands off driving in certain situations, we now know that you can charge money for that and there’s a business model for that.” He is less certain of the potential for fully autonomous robo-taxis, which will require a massive infrastructure investment to ensure they operate safely. “It’s questionable whether there’s an actual driving need for this in any way,” he affirmed. “I think that is going to be the long-term problem that inhibits the growth of AV applications.”

Dan O’Dowd, CEO of Green Hills Software and founder of The Dawn Project, a safety group aiming to make AVs and computers safer for humans, is a bit more bullish on the future of AVs. He pointed to the success of Google’s Waymo, whose cars can go “thousands of miles” in select cities without requiring human intervention. “If you stand on a San Francisco street corner for a while, you’ll see a self-driving car drive by,” said O’Dowd. “There are hundreds of robo-taxis, they’ve killed no one. Same thing in Phoenix.”

O’Dowd said that Cruise, General Motors’ self-driving division, has made headway as well and has expanded to Austin, Texas. He believes that both Cruise and Waymo are having success because they’re doing it slowly and learning new territory, mapping out every road as they go. However, Cruise and Waymo have billions of dollars to continually invest in autonomous driving, which begs the question: are time and money the solution? Is the lack of both why others have failed? “There are not going to be 10 trillion-dollar self-driving car companies,” O’Dowd answered, alluding to the race to be the first to commercialization. “In fact, there’s unlikely to even be one but that’s what they’re going after. Spending billions makes sense if you have billions.”