Thoughts on the Future

Gazing at the sky, imaging possibilities

Everything is predictable… in hindsight

I spend my life studying and applying AI, so you would’ve thought writing an article on it would be second nature. Yet I had a thousand ideas of what to write. The future is… big.

Ultimately, I decided to build from a core premise—we’ll take a wildly optimistic view of the short to medium term, assuming we maintain our current trajectory and all goes well. This is just one vague set of possibilities to explore.

Let’s discuss various aspects of our lives and how we might be impacted.

Transport

Leonardo Da Vinci originally sketched a concept for a self-propelling cart (sometimes referred to as the world’s first robot) capable of autonomous driving in 1478. It feels as though we’ve been waiting for self-driving cars ever since then. The promise is transformational - over one million people die globally each year in road accidents, and that’s not counting those who are involved, injured, or have their lives affected. Humans are terrible drivers. Maybe you aren’t, but drive one stretch of motorway and it’s hard not to come to that conclusion. Even if we aren’t “bad” skill-wise, some drive drunk, high, tired, angry, or on the phone, eating, telling the kids to be quiet, or just plain daydreaming. What if, instead of relying on humans to keep 100% concentration and a high skill level to keep roads safe, we could delegate that to robots? We’d dramatically reduce road accidents and road rage incidents. We’d also see traffic evaporate since these autonomous vehicles could be connected and communicate, they could effortlessly avoid each other, go at the fastest speed possible safely, and reroute automatically before coming to a blocked path.

Well, even though we’ve been waiting a while for self-driving cars, for many people, they are already here.

“The future is already here—it’s just not evenly distributed.” — William Gibson.

In many cities worldwide, particularly in the USA, you can see autonomous vehicles regularly driving around the streets. They are still being tested, but it’s at scale and in the real world. The likely outcome is that companies like Tesla, Waymo, or the UK’s Wayve continue to push forward and progress in the adoption of this technology, and we continue to see more journeys taken in these robots on wheels. In the West Midlands, the automotive industry has great historical, cultural, and economic significance - we should be paying a lot of attention to this direction of travel, and try not to get distracted no matter how many crap Jaguar adverts get made.

The fact that the cars will be autonomous will have secondary effects - not only will travel be dramatically safer, but it will be more accessible since you don’t need a driving license to be able to benefit. There would be no need for a steering wheel, since… you won’t drive. So, the car would become a system optimised for entertainment, work, or sleep rather than driver comfort. In fact, there would be no need for many things we associate with a car today - even wing mirrors, although there are non-obvious effects through redesigning the car and removing these features, such as how blind people use wing mirrors to know where the front of the car door is and navigate around to a door handle.

Cars would be more efficient and able to minimise fuel consumption. There’d be little need for huge multi-storey car parks - you could be dropped off exactly where you wanted to go, and the car could drive off anywhere until you beckon it back. Many think you won’t own a car but will use a service (similar to how you travel by air - you think of the airline, not the plane). Cars are typically parked for around 95% of their lifetimes - it doesn’t make sense to own one if you can grab a ride anytime, companies like Uber are prime candidates to shape this future. If you did own a car, you could easily have it join a ride-share fleet whilst you weren’t using it - the car becomes an actual asset, earning you money whilst you aren’t using it. Tesla believes this will be a reality for many.

Imagine this future for a second - what could we do with all of the space currently taken up by eye sores known as car parks? How much happier would we be without having to stew in traffic or deal with idiots on the road? What else may change? On-street parking? There’s no need for it. Signs? Mainly for humans anyway. Garages? Everyone will suddenly be installing home gyms they never use. You’d need fewer and narrower lanes since autonomous vehicles require less space and wiggle room.

That is, if humans are removed from the transportation loop, of course. If not, it’s likely to be a slow crawl to change rather than all at once. Even then, small changes year-on-year compound into dramatic changes over a decade. But it’s hard to imagine that if accidents trend to zero with this technology and we see the main cause as human input, we won’t adapt as a society. We might look at human driving in the future like we look at drunk driving now - “How reckless? I can’t believe they’d do that.”

The contradiction to keep in mind here is that there are tremendous technological challenges to solve before this becomes reality, but it’s also not that far-fetched or potentially that far away. There are also some consequences we should take very seriously.

What happens to all the jobs that rely on driving? There may be more demand for other services, such as cleaning ride-share cars or new types of signalling and network design. But truck drivers, lorry drivers, taxi drivers, delivery drivers, bus drivers… they are dramatically reduced or disappear entirely. What about all the jobs in the driving supply chain? Mechanics look very different as a service. Wider public transport would take a huge hit. The car manufacturing industry and supply chain would completely change. All in all, if self-driving cars hit, then millions of jobs will be lost immediately or impacted immeasurably.

Health

Healthcare looks very different in the near future. Doctors and medical professionals are often busy, tired, hungry, distracted, in a rush to judgement or just lacking curiosity… and that’s once you’ve waited a few weeks for an appointment and can still be bothered to investigate things yourself. AI will mean there’s a doctor in your pocket, 24/7, completely and utterly devoted to you and your health. It can know everything about you - not just what the current worry or ailment is, but your diet, your lifestyle, how you’ve slept for the last few weeks, what stress you have at work, the extent to which your kids are impacting your sanity, and your medical history. Not just your medical history - that of your immediate family and anyone you’re related to. It can know all relevant medical research, all treatment pathways that are potential candidates, and how effective those treatments were for people with your genetic make-up. It has all the time in the world for your questions, and what’s more, if you ever do want to talk to a human doctor, then your conversations can be digested into a transcript and summary, meaning you don’t need to repeat yourself and they can be up to date with minimal effort.

I say “if” you want a human doctor involved. For the short term, there’s little question that most people will. That’s the reality that Microsoft and others are working on. They are looking to use AI to make doctors’ and nurses’ lives easier by having AI take care of the clinical pathway paperwork, communication, admin, and more. That will mean your medical professionals should have more time and attention dedicated to you, which is transformational for you as a patient and for the medical profession.

However, there is no guarantee you will always want the “human in the loop”. Let’s imagine AI keeps improving from where it is today. One study showed diagnosis accuracy rates for complex diseases of 74% for human doctors, but when they had access to ChatGPT for help, they increased to 76% accuracy. That’s great - AI as a co-pilot for doctors improves efficiency (slightly). What’s really interesting, though, is that if you were to trust ChatGPT alone, removing the human from the loop, then the diagnosis accuracy went to 94%. Human professionals held the AI back.

The other side of this is bandwidth - how many diagnoses can you get right, and how quickly can you do them? With AI, you have almost immediate feedback. With humans involved, you introduce multiple bottlenecks. Someone needs to review case notes. They need to make recommendations. These often need sign-off from others. All of this takes time and resources. Having the human as part of the process is a very comfortable thought for us, but how long will it survive when it costs time and money?

This tension between AI efficiency and human oversight isn’t unique to healthcare—it represents one of the fundamental questions we’ll face across every industry transformed by AI. When does human involvement add critical value, and when does it merely slow things down? This balance is what many refer to as the “human in the loop” problem.

Human In The Loop

While keeping humans in the loop sounds reassuring and practical, there’s a fundamental challenge: computers think 20,000 times faster than humans. By involving humans, you incur additional time and costs. Does that mean you shouldn’t involve humans? Not at all—it should always be necessary in many domains. But will it be the case? The problem is that if you keep a human in the loop and a competitor doesn’t, they decrease costs and increase convenience—two massive advantages in any market.

But can these systems be trusted to that extent? Despite God-like abilities, they are also weirdly and deeply flawed. You should picture their capability as Swiss Cheese - impressive but with random holes, and many of them. Without significant workarounds, they are also only trying to predict what is plausible, which is very different from what is true. You absolutely cannot trust them for now - we’ll discuss why this is in-depth another time. In future, these flaws will likely trend to zero with the skill and scaffolding of software engineering around them.

The truth is, though, that the human bottleneck is likely why change occurs slowly and then all at once. After all, despite our having God-like technology, a lot of the NHS still runs off Windows XP. Inertia and legacy are real. So, I can buy the version of events that small changes year on year lead to a huge transformation over a decade.

Human-Computer Interactions

As AI transforms transportation and healthcare, it will also fundamentally change how we interact with technology itself. It’s hard to know the effects of a generation of people growing up thinking that when a machine makes a mistake, it means they just need to talk to it more. Back in our day, if a machine made a mistake, you’d give it a whack, turn it off and on, or throw it out completely. Now? If it gives you the wrong answer, you keep chatting and negotiating until it gets it right.

Relationships with AI and the way humans interact with computers will undoubtedly change, and some of this change will be incredible.

The Digital Revolution left people behind. If you couldn’t access forms easily, navigate websites, or were uncomfortable with email… well, you struggled to access even the basic benefits everyone else took for granted.

The AI Revolution promises a different world, where the interface is simply language. That opens the doors to everyone who can talk to access all services and benefits—incredible. But what about those who can’t use language? Well, AI is now capable of reading brain waves from scans to understand thoughts. Yes, you heard that right. AI can read your mind. The average accuracy of brain-wave-to-word translation is around 50%. That is incredible, and it will only get better.

What’s more impressive? This is all from the same technology - Neural Networks. We’ve learned that, despite our yearning for beautiful theory and discipline expertise, if we throw enough compute power, data, and a neural network at the problem, it will tend to figure it out. The current hypothesis dominating strategic investment is that more compute power gives more intelligence.

All countries in the world are currently in the race to AI, including the UK which is dedicating billions in its AI Opportunities Plan. I can give my critiques of the plan separately, but my biggest is - let’s imagine they succeed. What then? There is no discussion of a vision for the world where those plans actually work. What happens to jobs? To the economy? To the human experience? This is the job of government, and many top AI labs are warning of imminent Artificial General Intelligence - we should at least plan as if that were a real possibility, but nobody has seemed willing to so far.

It’s part of the work I’ve done in the West Midlands, which I’m stepping down from and I hope whoever leads next will continue. In our Economic Case for AI, the feedback was that we were the only strategy people had heard taking seriously addressing job losses. That is worrying.

AI Unlocks The Doors

So, what else can we expect to see? Well, what’s your favourite sci-fi film? Picture the tech in that film… that’s what we’ll get, more or less. We’re already capable of most of it. The only difference is that the film will be the Instagram version of the tech. Highlight reels only. You don’t see the messiness. The no make-up wake-up. The crap. Our technology experience over the next few years will feel like a sci-fi movie, just more crap. We’ll have machines capable of so much, yet that are also so stupid at the same time.

There is a convergence of technology that makes progress even more exciting and unpredictable. Along with those leading the AI chase (OpenAI, Anthropic, Google, X.AI, Meta) are surges in progress across Quantum Computing (Google, Microsoft, D-Wave), Bio-Technology (Neuralink, Microsoft, Meta), Space Travel (SpaceX, Blue Origin, Firefly), Robotics (Tesla, Figure, NVIDIA) and more. Advances from these other fields will wildly change our human experience, though each presents significant challenges on its own. I can see a path for AI to be the key to figuring out the rest, meaning a crescendo of activity around unlocking powerful AI. That better come with wild progress in Cyber Security, or your email being hacked is the least of our issues.

The future is unknown, but I bet three things. 1) I’ll be wrong. 2) It will happen faster than anyone is ready for. 3) It will be wild - are you ready?