What Will It Take to Make Driverless Cars Safe?

Why Trust Techopedia

Where did it all go wrong for driverless cars? After successful trials in 2019 and go-lives in 2020, autonomous vehicles (AVs) were being hailed as the future of transport. Then, late last year, things took a turn: accidents, lawsuits, layoffs, software glitches, cars pulled off the road and a steady stream of bad PR.

Today, the three companies vying for AV dominance – Waymo, Cruise, and Tesla – are under investigation over safety concerns. Now, a study has found that while AVs are generally less prone to collisions than humans, in certain conditions they fare worse.

With the negatives piling up, are driverless cars still a good idea? We look at the technical and ethical barriers to adopting fully autonomous cars and consider what can be done to address both.

Key Takeaways

  • After a promising start, Waymo, Cruise, and Tesla are being closely watched by US transportation safety agencies.
  • A series of incidents, minor and major, have triggered an AV backlash from a wide range of opponents.
  • Local police, fire departments, and community groups in US cities all say self-driving cars have been more menace than benefit.
  • Yet some very big names have committed big money to the AV rollout.
  • The most recent evidence suggests driverless cars are actually pretty safe. Will a nervous public be convinced?

Is a Driverless Backlash Underway?

AV companies have had a bad run of it recently. In April, GM-owned Cruise announced that it had returned to manual driving, pausing autonomous operations “to focus on rebuilding trust with regulators and the communities we serve.”

This followed a cascade of problems that kicked off in October 2023 when one of its robotaxis struck and nearly killed a pedestrian in San Francisco.

Even before the accident, a local anti-AV activist group had cataloged hundreds of mishaps and near misses, culminating in a stealth campaign to disable the company’s cars by placing traffic cones on their hoods.

Advertisements

San Francisco’s police and fire departments weighed in, complaining of multiple incidents where self-driving cars had blocked or complicated rescue operations.

In other US cities, Google’s Waymo robotaxis suffered a similar fate: colliding with a tow truck, clipping other cars, blocking traffic, and getting confused around construction sites.

The National Highway Traffic Safety Administration has launched a probe to try and understand exactly what’s going on and who’s to blame.

Tesla is under the microscope after a massive recall of its AVs last year, as is Amazon-backed Zoox. Is it time to call the whole AV experiment off?

Tesla is under the microscope after a massive recall of its AVs last year, as is Amazon-backed Zoox. Is it time to call the whole AV experiment off?
A total recall on robotaxis? Source: Vanity Fair/Ronald Grant Archive/Alamy

The Truth About Autonomous Vehicle Safety

Not so fast. Despite a spate of bad press, a recent study has found that, while AV car safety merits some concern, they’re generally safer than the alternative.

Researchers at the University of Central Florida looked at 37,000 accidents involving AVs and human-driven vehicles. Assessing various autonomous vehicle safety metrics, they found that driverless cars were generally less likely to be involved in accidents than those driven by people – though they underperformed human drivers in specific situations.

Self-Driving Car Accident Rate vs. Human

According to the estimated self-driving car safety statistics, Level 4 vehicles – those with the greatest degree of autonomy – performed well.

  • They were notably less likely to be involved in accidents causing minor injury (< 36%) and 90% less likely to be involved in a fatal crash.
  • The risk of a rear-end collision was roughly half what human-driven vehicles register.
  • The risk of a t-bone collision was about one-fifth.
  • They were 1/50th as likely to run off the road.

AVs fared less well in conditions with variable lighting (neither light nor dark). The study found they were more than five times as likely to be involved in a mishap around sunrise and sunset.

Navigating turns can also be problematic, with the double odds of an accident while turning compared to human drivers.

Ethical Questions of Autonomous Driving

While twilight and turning pose some challenges, overall, the safety figures favor AVs. Where it gets tricky is analyzing the ethics. Some of the biggest questions AV companies and lawmakers are wrestling with are philosophical.

A 2015 study by the University of Toulouse considered what an AV might do if, for example, it had to choose between hitting a jaywalker or colliding with an oncoming car. How should the AV’s software decide? Is it possible to write an algorithm capable of understanding the context and balancing the ambiguities that arise in a ‘lesser of two evils’ dilemma?

Those questions extend beyond the hypothetical. If you knew your self-driving car would sacrifice you in favor of a group of pedestrians, would you still want one? And if the algorithm ultimately chooses who or what to collide with, who bears legal responsibility, you or the manufacturer?

In earlier days, it was easier to assign blame. Cars didn’t come with a hundred million lines of code, GPS trackers, or remote kill switches. If you sat behind the wheel and hit a car or pedestrian, you bore at least some culpability for what happened. Courts would look at the specifics and decide how much.

Increasing levels of in-car automation feed into a natural nervousness people have about ‘letting go of the wheel’ and trusting their lives to a large, heavy machine that moves at high speed.

But giving in to fear ignores the sunnier reality of the Florida study’s key findings: across major categories, driverless cars have proven to be safer – much safer. While bad news gets attention, there are still plenty of reasons to be pro-AV.

Compelling Use Cases

Teodora Kaneva, Head of Smart Infrastructure and Systems at techUK, told Techopedia that the use cases for autonomous vehicles extend well beyond urban robotaxis.

Self-driving cars could be used to augment public transport, fill in existing transport gaps at lower cost, or connect to existing hubs in cities with a ‘mobility as a service’ (PAYG car share) model. Kaneva said:

“Connecting workplaces to train stations, for example, could contribute to reducing congestion and freeing up space for greater active travel. These use cases can help support an integrated and widely accessible transport network across the country.”

And consumer applications are just the start. Driverless vehicles “can also be deployed off-road; supporting worksites and airports to improve their efficiency and protect workers from harm.”

One option sometimes mooted for the trucking industry is ‘platooning,’ where a lead truck uses a human driver, followed by a series of AV-driven trucks in the back that auto-mimic the lead truck’s actions: maintain speed, braking, accelerating, slowing down, and stopping – all in tandem.

AVs could also help smooth out the transportation experience while governments deal with decaying road and rail networks straining under population growth and a steady stream of new car purchases.

Anyone who commutes by car knows the near-daily pain of accidents, traffic jams, lack of parking, and potholed roads and highways.

Guided by AI, autonomous vehicles could buy time for upgrading national transport infrastructure and de-stress the experience of traveling by car.

Regulation the Enabler

Before any of that becomes even a remote possibility, driverless car companies will need to sort out their technical failures. They’ll also need a consistent legal and regulatory framework that clears up questions about risk and responsibility.

In the UK, the Automated Vehicles Act became law in May 2024, paving the way for a mass deployment of AVs in the country by as early as 2026.

Kaneva says there is still work to be done, however:

“The industry needs enabling regulations to provide a clear path to removing the driver and commercializing the technology. It also needs a clearer sense of the evidence they will need to obtain authorization under the Act once regulations are in place.”

In an analysis published earlier this year, techUK laid out other twists & turns that have to be negotiated before general acceptance and adoption become realistic.

Self-Driving Car Safety Barriers to Overcome

Social AcceptanceBetter InfrastructurePublic InvestmentVehicle ChargingData & ConnectivityManufacturing Capacity

Building public trust in driverless vehicles is crucial, with effective communication and exposure to the technology seen as essential for gaining public acceptance.

AV services will need to be wedded to existing transport infrastructure, so ensuring roads are in good repair is vital.  Driverless vehicles should only need minor modifications to operate safely.

Especially for deploying AVs in rural or semi-urban areas where profitability could take longer. A Westminster-backed funding model could remove some of the financial risks for AV companies.

Since most AVs will be EVs, a widespread charging infrastructure is a non-negotiable requirement. That also means looking closely at local grid capacity and connection queues.

Reliable digital networks will be needed to support AV business models and services. That shouldn’t mean, however, that automated vehicles need continuous connectivity to operate.

As demand for AVs increases, the UK will need to develop domestic manufacturing capabilities to produce ‘native’ driverless vehicles and keep the economic benefits within the country.

In the longer term, techUK’s Kaneva says this last part is crucial.

“Right now, industry needs to adapt conventional vehicles to drive themselves. That introduces cost and time to getting vehicles on the road. We believe the government has a role in incentivizing the manufacture of vehicles that are designed from the ground up to be safely and securely driven autonomously. This could make the UK the global home of advanced manufacturing.”

The Bottom Line

Despite the limitations, markets have grasped the promise of AVs and aren’t letting go. In 2019, there were around 30 million cars with partial or full (level 4) automation operating globally. This year the number is expected to shoot past 54 million.

The worldwide market for AVs is on a similar growth trajectory. In 2021 it was valued at over $24 billion. By 2026, it’s projected to be worth over $60 billion.

And China is eying the market, racing forward, and handing out licenses while other countries apply the brakes over safety concerns.

The next move lies with lawmakers and regulators, who need to establish a safe and level playing field so the industry can flourish.

FAQs

How safe are autonomous cars?

What are the safety challenges for autonomous vehicles?

What is the accident rate for autonomous vehicles?

Will autonomous cars reduce accidents?

Is Waymo safe?

Advertisements

Related Reading

Related Terms

Advertisements
Mark De Wolf
Technology & iGaming Journalist
Mark De Wolf
Technology & iGaming Journalist

Mark is a seasoned tech journalist covering esports, igaming, GambleFi, Web3, and topics at the intersection of blockchain and gambling. His work has appeared in Redshift, Investing.com, Energy Central, Marketing, and The Startup. He’s an honors graduate of the Ryerson University School of Journalism, where he studied under senior reporters from The New York Times, BBC, and Toronto Star.