Powered by AI: the flying taxis which will cause a productivity revolution

Martin Warner

Flying taxis have long been the preserve of sci-fi movies, but they will soon be in the skies above us, changing the complexion of urban travel.

eVTOL (electrical Vertical Take-Off and Landing) aircraft which use battery powered technology, will occupy the sub-2000ft airspace above and between cities. In a city such as London they will make journeys of five to 35 minutes, hovering, taking off and landing vertically – with around 10 rooftop stops serving multiple routes.

Morgan Stanley predicts the passenger eVTOL market will be worth around $1.5 trillion by 2040, but many commentators believe that is an under-estimate. The first passenger vehicles are likely to launch within three years and within 15 years the infrastructure is likely to be in place to support passenger routes.

As an eVTOL pioneer developing a six-seater prototype through my company, Autonomous Flight, I know how exciting this opportunity is for the future of urban transport, but I am also aware of the scale of the technological challenge – and the need to persuade regulators and the public that flying taxis will be a safe mode of travel.

The Washington DC collision between an American Airlines jet and an Army Black Hawk helicopter, leading to the deaths of all 67 people, highlighted safety concerns about the role of air traffic controllers in policing low altitude airspace.

The role of AI in eVTOL technology

But the rapid advance of machine learning has been a game changer for the eVTOL industry, speeding up the development and application of the technology and providing enhanced safety mechanisms.

Powered by lithium-ion batteries, the first-generation aircraft are being designed to reach cruise speeds of 125 mph and will have a range of 100 miles on a single charge. Later models will be fully autonomous. Understandably, securing certification is a complex four-stage process, with seven sub-stages, and bringing a single electric aircraft to production costs tens of millions of pounds.

We have deployed machine learning extensively to develop our prototype. Simulation tests that might previously have taken five years can now be completed in a few weeks. A large language model learns from historical data to predict system performance under untested conditions; it recommends the next round of tests, which will provide the most relevant information, enabling engineers to focus on critical areas, reducing development timelines but without compromising safety or performance.

AI is transforming engineering in the same that CAD revolutionised design engineering a generation ago. It will also revolutionise eVTOL maintenance; AI algorithms can analyse vast amounts of data from onboard sensors and flight logs, identifying potential issues before they escalate. This will enable predictive maintenance, reducing downtimes and preventing costly repairs.

Checkpoint indicators provide a wealth of information that has never previously been available, telling us everything that is happening on board the aircraft, from temperature casing to battery analysis. Through dynamic interfacing if something goes wrong with the Flight Management System it is automatically rebooted, which a traditional checkpoint system cannot do. The AI overlay builds in a layer or redundancy that means the aircraft can still fly if the battery fails; if two propellers fail it will tell us where the integration points are and conduct a radical overhaul.

The end goal is a zero point of failure – the holy grail in aircraft safety.

But AI will also improve the experience for passengers. Before they even leave the aircraft, hot point indicators will provide invaluable insight about the traffic conditions around them. What is the fastest route to their next destination? What is the best form of travel? Are there last-minute tickets available for the hot new theatre show that has just opened?

The productivity revolution

This is the application of AI that interests me most because the next big tech wave will be the revolution in human productivity. AI gives us a unique opportunity to organise our lives with minimum effort for maximum value, defining and measuring our hourly output and creating more time capacity.

Using the same architecture from our proprietary large language model, which has powered the development of an eVTOL prototype, we have developed warpspeed, an AI-enabled app that will revolutionise personal and workplace productivity.

By constantly learning about the user, AI will take on autonomous functionality, undertaking multi-step tasks to deliver unprecedented productivity gains. An AI agent will do 80% of the heavy lifting, which will be especially valuable for entrepreneurs running smaller businesses, but will help anyone organise their life more efficiently. A multi-dimensional calendar will autonomously plan schedules; a tagging and organisation tool will turn rough notes into action-driven lists that can undertake tasks; and a task implementation tool will autonomously make daily routines more productive.

What does it mean for us?

If Humanity 1.0 was the age of information abundance, driven by automation in the 1990s and the organisation of information through the development of the worldwide web, search and ecommerce in the 2000s, then Humanity 2.0 will be the age of information meaning. Machines will understand human context and outpace human thought – and life will speed up consequently.

For now, there is plenty of commentary around the limitations of AI – the withdrawal of an Apple feature that inaccurately summarised news stories from the BBC and other media outlets, for example – but this will be a short-lived phase in its development.

Once an AI agent can do tasks without our knowledge and with 100% reliability, we will be in an era where speed is no longer a meaningful measurement. The productivity revolution will be complete. The challenge will be how we partner with AI; how we police its applications; and how we live with its consequences.

Martin Warner is the founder of Autonomous Flight