AI took the wheel: Inside my first autonomous car experience

On a recent trip to San Francisco, I found myself sitting in the backseat of a taxi with no one in the driver’s seat. As the vehicle pulled out into traffic and navigated the bustling streets of downtown Chinatown, I couldn’t help but feel like I had stepped into the future. This was a Waymo self-driving car – fully autonomous, driven entirely by artificial intelligence. For someone from the UK, where such vehicles are still largely a concept, this experience was both exciting and slightly disorienting.

It struck me that, while driverless vehicles might be familiar to some in the US and China, in the UK and Europe, we’re not quite there yet. The knowledge that these Level 4 autonomous vehicles exist, and are actively in use in densely populated areas, may not be as widespread. This gap in awareness made me consider the uneven pace of AI adoption. While we’re witnessing incredible advancements, significant challenges and uncertainties remain before such technology becomes a seamless part of everyday life, especially for autonomous vehicles.

The AI behind the wheel

As the vehicle seamlessly navigated through the dense roadways of San Francisco’s city centre, it became clear that this wasn’t just about following a pre-set route. The AI was constantly reading the road, adapting to changes in real time, and making decisions with a level of precision that was both impressive and slightly surreal. I had already been struck by how confidently the car moved through junctions and responded to signals, but there was one moment that truly stood out.

We were weaving through normal city traffic and inching forward towards a busy junction when a large bus swung into view, struggling to navigate the tight corner.  It was the kind of situation that would make any human driver a little anxious: it was clear there wasn’t quite enough space for the bus to make the turn, and cars were lined up behind us. With safety a priority of its operation, the Waymo taxi gently came to a halt.

A car ahead of us remained still for just a moment before carefully shifting into reverse, intending to create room for the bus to complete its wide arc. With human-like intuition, our own car reacted in tandem, easing back without hesitation. We watched as the bus slowly made the turn. Then, as soon as it was safe, our car continued on its way as if the whole encounter had been perfectly orchestrated.

It wasn’t just the flawless execution that amazed me – it was the calm, calculated manner in which the car responded to a dynamic situation that would have tested even an experienced human driver. The AI didn’t just follow predetermined instructions; it actively navigated a complex scenario, demonstrating an incredible level of situational awareness.

More than a buzzword, a glimpse of the future

The global buzz around AI, especially since the release of generative models like ChatGPT, has been hard to ignore.  ChatGPT represented a breakthrough in solving the challenge of human-like language interpretation and production, demonstrating that AI had effectively tackled the problem of natural language processing. Yet, as with many groundbreaking technologies, the progress has been more measured than anticipated.  While the breakthrough was undeniable, subsequent advancements have been incremental, leading some to question if the initial hype has outpaced practical adoption.

Analysts and industry experts have begun to caution about an ‘AI Bubble’, suggesting that while there have been significant investments, the anticipated rapid transformation hasn’t materialised as quickly as some expected. This highlights the delicate balance between technological advancement, regulation and society’s readiness – particularly in areas as critical as autonomous driving – along with identifying appropriate use cases, or “problems to solve”.

Despite these challenges, my experience with Waymo gave me a tangible glimpse into what the future might look like.  Waymo’s level of AI maturity, with its Level 4 autonomous vehicles, was beyond what I expected. It offered a striking example of AI technology fully integrated into core operations. However, much like the breakthrough of GPT, we now seem to be entering a phase of more incremental advancements, a time where the technology is awe-inspiring, but where the true transformation requires sustained effort and refinement.

What became clear is that very few organisations have deployed AI in such an advanced way. Achieving this level of integration demands a significant investment—not just in technology, but in time, skills, processes, and money. It’s not a simple plug-and-play solution. For those interested in a deeper exploration of the stages of AI adoption within business, I’ve written more about them in a separate paper.

The challenges of autonomy: Why progress takes time

Despite the futuristic allure of fully autonomous vehicles, the road to widespread adoption remains challenging.

The answer lies in the intricate and expensive nature of autonomous vehicle development. Building a fully autonomous vehicle requires massive investment in research, testing, and infrastructure – similar to how space exploration or high-speed rail systems take years of planning and billions in funding before they become viable. It’s one of the reasons why many car manufacturers have opted to partner with tech companies specialising in AI. These development cycles span years, even decades, as the technology continues to evolve. For traditional automakers, it’s like running a marathon in an ever-changing race, where the finish line keeps moving. This long timeline makes it difficult for car manufacturers to see a clear path to profitability in the near term. Instead, many have shifted their focus to incremental innovations like advanced driver-assistance systems (ADAS)—features that offer immediate benefits such as enhanced safety, convenience, and efficiency without requiring full autonomy. It’s a more manageable leap, offering short-term gains while laying the groundwork for the autonomous future.

Moreover, regulatory uncertainty also remains a significant barrier. Around the world, governments are still working to establish clear frameworks for managing autonomous vehicles, particularly regarding safety standards, liability, and insurance.

Richard Damm, chairman of the United Nations Economic Commission for Europe (UNECE) Working Party on Automated and Connected Vehicles, has stated that globally harmonised regulations for Automated Driving Systems (ADS) are expected by mid-2026. Meanwhile, Tesla recently announced plans to launch its Full Self-Driving (FSD) system in Europe and China by early 2025, pending regulatory approval. For manufacturers, the risk of navigating these uncharted regulatory waters is daunting and why pour billions into developing fully autonomous systems when the legal infrastructure isn’t ready yet?

Public trust and acceptance remain significant hurdles. Companies like Tesla have been at the forefront of autonomous technology, but incidents involving early-stage systems have received notable media attention, raising concerns about the readiness of driverless cars – and the technologies they rely on. Tesla, in particular, has faced scrutiny for its ambitious timeline, with some questioning whether advancements have sometimes outpaced safety assurances. This has led to a more cautious approach from both consumers and other manufacturers, as the industry navigates how best to introduce these innovations responsibly.

Finally, many car companies are currently focusing on electrification, driven by upcoming stricter emissions targets and growing demand for greener vehicles. The number of electric cars sold in the first quarter of 2024, surpassed the annual total from just four years ago. With such regulations and raised demand for greener vehicles, autonomy could continue as a longer-term vision with the current focus remaining with electrification to meet emissions targets and offer a more immediate return on investment.

The future is now, but the journey has just begun

As the car came to a stop and I stepped out of the Waymo taxi, I was left with a powerful realisation: what once seemed like science fiction is now a reality. The technology is here, but the road ahead is far from straightforward.

While the AI driving this car was impressive, it also underscored the myriad of challenges that lie ahead – technical, regulatory, and societal. Ensuring that the AI systems have access to accurate and comprehensive data, that they can adapt to an ever-changing environment, and that they gain the public trust are all critical to the future of autonomous vehicles. However, perhaps one of the greatest hurdles will be navigating the complex regulatory landscape. Governments and policymakers must establish clear safety standards, liability frameworks, and licensing protocols before fully autonomous cars can become a reality on all public roads.

Despite these challenges, it’s clear that we’ve made great progress and will continue to do so. But it does prompt critical questions as to the broader implications for society and the environment. The road ahead is filled with promise (and soon, autonomous vehicles maybe!), but it demands a commitment to understanding and addressing its many impacts, so we can shape our future thoughtfully and beneficially.