Saturday, 5 September 2026
Artificial intelligence is pushing the automotive industry towards a future in which vehicles do considerably more than drive themselves. Cars are increasingly being developed as intelligent platforms capable of understanding their surroundings, communicating with passengers, anticipating journeys, interacting with infrastructure and potentially operating independently when their owners are elsewhere. That wider transformation was explored during the AI4 2026 panel “The AI-Powered Vehicle: Smarter, Safer, and Fully Connected,” where representatives from May Mobility, Tensor and the Toyota Research Institute discussed how artificial intelligence is changing vehicle development and what autonomous transport could eventually mean for cities, dealerships, insurance, logistics and everyday travel.
The three organisations represent different approaches to the same transformation. May Mobility is developing autonomous transport technology for deployment through cities, fleets and ride-hailing partners. Tensor is pursuing a privately owned Level 4 autonomous vehicle designed from the beginning around AI, while Toyota is exploring autonomous driving alongside a broader philosophy in which AI assists and improves human driving rather than necessarily removing the driver completely.
May Mobility has been developing an autonomy architecture that combines learned AI models with predictive modelling of the surrounding environment and real-time reasoning. Jacob Crossman, Senior Vice President of Autonomy at May Mobility, described how the company had been able to learn not only from the behaviour of its own vehicles but also from other road users observed by its sensors. Combining those observations with its existing decision-making technology allowed the system to handle increasingly complicated situations. The broader objective is to create autonomous systems capable of responding to unfamiliar circumstances rather than depending entirely on enormous amounts of location-specific driving data.
Tensor is approaching the problem differently. Rather than adding autonomous technology to an existing conventional vehicle, the company has designed its Robocar around autonomous operation from the outset. The vehicle incorporates more than 100 sensors, including cameras, lidars, radars, microphones and other monitoring equipment intended to provide extensive awareness of both its external environment and what is happening inside the vehicle.
The company’s ambition extends beyond replacing the driver. Tensor envisages the vehicle becoming a personal AI assistant that can understand natural-language instructions and eventually connect journeys with other aspects of its owner’s daily life. Instead of simply entering a destination, a passenger could ask the vehicle to stop at a particular entrance or alter a journey conversationally. Connected with a person’s calendar and other applications, a vehicle could potentially recognise an upcoming meeting, calculate additional journey time caused by congestion and recommend leaving earlier.
Toyota’s research highlights another possible direction for automotive AI. Rather than assuming autonomous technology must ultimately eliminate human driving, the company is investigating whether increasingly capable AI could make people better drivers. One example discussed at AI4 was an AI driving instructor capable of helping inexperienced motorists improve their skills. Such technology could potentially identify dangerous habits, anticipate risks and assist drivers before mistakes become accidents.
The distinction is important because fully autonomous vehicles are unlikely to arrive uniformly around the world. Road quality, regulation, infrastructure, driving behaviour and consumer acceptance vary considerably between countries. Technology capable of operating without a driver in a carefully mapped American or Gulf city may face very different conditions in markets where road markings, infrastructure or driving behaviour are less predictable.
The panellists were consequently relatively cautious about how quickly complete autonomy will dominate global vehicle sales. When asked to estimate the share of new vehicles that could be fully autonomous 15 years from now, their estimates generally remained below or around one quarter of worldwide sales. The discussion suggested that technological capability may eventually advance faster than the regulatory systems determining where and how autonomous vehicles can operate.
For the property industry, one of the most significant consequences could be the changing relationship between cars and urban land. Private vehicles spend much of their lives parked, meaning offices, shopping centres, residential developments, airports and city centres have historically been designed around substantial amounts of parking infrastructure. If autonomous vehicles can deliver passengers and then leave independently, parking no longer necessarily needs to be located immediately beside a building. Shared autonomous fleets could reduce that requirement further by keeping vehicles moving between users.
Over time, this could release valuable urban land currently occupied by surface parking areas and multi-storey garages. For developers and city authorities, sites previously required for parking could potentially accommodate housing, commercial development, public spaces or other uses. Building design could also change, with greater emphasis on passenger drop-off areas, autonomous vehicle waiting zones and charging infrastructure while conventional parking requirements decline.
Hotels and the wider travel industry could experience another change. Once passengers no longer have to concentrate on driving, travelling time becomes usable time. People could work, communicate, rest or consume entertainment while moving between destinations. An autonomous vehicle capable of travelling for several hours while passengers work or sleep could alter the perceived distance between cities and potentially affect competition between road, rail and short-distance air travel. Hotels could eventually interact directly with vehicles, coordinating arrival times, parking, luggage handling and other services before guests reach their destination.
Autonomous vehicles could also influence how cities manage traffic. Vehicle-to-vehicle and vehicle-to-infrastructure communication has been under development for years, but AI potentially allows cars, traffic signals, mobile devices and transport infrastructure to exchange information and respond more dynamically. Vehicles could receive warnings about dangerous road conditions detected by cars ahead, while traffic systems could potentially adjust signals and vehicle speeds to improve flows through congested areas. The challenge is that many of these benefits become substantially greater only when a meaningful proportion of the overall vehicle fleet and surrounding infrastructure is connected.
Commercial transport may move faster towards autonomy in some areas because the financial case can be easier to demonstrate. Trucks operate for long periods, while drivers represent a significant component of transport costs. Long-haul freight routes, distribution centres, ports and logistics corridors could therefore become important environments for autonomous transport. That development would also affect logistics real estate as warehouses and distribution centres adapt to autonomous vehicle movements, automated loading, charging infrastructure and more integrated fleet-management systems.
Car dealerships could face an equally significant transformation. The traditional model assumes that customers travel to a physical location to inspect, purchase and service vehicles. Autonomous cars could reverse part of that relationship. A demonstration vehicle could theoretically travel to a prospective customer’s home, provide a test journey and return to the dealership independently. That could reduce the importance of some traditional showroom functions while increasing the role of technology, servicing and customer-experience facilities.
Vehicle maintenance could change in a similar way. A car capable of identifying a technical problem could potentially schedule its own service appointment, travel to the workshop when its owner does not require it and return after the work has been completed. This could allow dealership workshops to operate more intensively outside conventional customer hours and potentially alter the design and location requirements of automotive service properties.
AI could simultaneously create new dealership revenue opportunities. Vehicles capable of extensive personalisation may support software services, digital upgrades, customisation and aftermarket products throughout their operating lives. The economic relationship between manufacturer, dealer and customer could therefore become less dependent on the original vehicle transaction.
Insurance presents another potentially significant disruption. When a person is driving, responsibility for an accident traditionally rests largely with the driver and their insurer. When a Level 4 autonomous system controls the vehicle and the passenger cannot intervene, responsibility increasingly moves towards the manufacturer or technology provider. Tensor told the AI4 audience that it intends to assume liability when its vehicle is operating autonomously at Level 4, while responsibility would return to the human driver during manual operation.
That distinction could eventually produce more dynamic insurance models. Vehicles could record how much distance was travelled autonomously and how much was driven manually, potentially allowing insurance exposure and pricing to reflect who or what was controlling the vehicle. The relationship between automotive manufacturers and insurers could consequently become considerably closer as responsibility gradually shifts from individual driving behaviour towards the performance of autonomous systems.
Privacy represents another major challenge. Intelligent vehicles generate enormous amounts of information through cameras, microphones and other sensors. That information could potentially reveal where people travel, who accompanies them and what happens inside their vehicles. Autonomous vehicles will therefore have to establish trust not only through driving performance but also through their handling of personal information, cybersecurity and accountability when systems fail.
One of the most difficult technical problems remains uncertainty. Driving environments are inherently unpredictable: pedestrians change direction, vehicles behave unexpectedly, weather affects visibility and unusual circumstances arise that cannot all be individually programmed. May Mobility’s approach involves allowing its system to evaluate multiple potential outcomes before selecting an appropriate response, particularly when uncertainty increases.
The industry’s progress will therefore depend on considerably more than developing cars capable of remaining within lanes without human intervention. The larger transformation involves creating machines capable of understanding people, roads and increasingly the wider infrastructure around them. That brings autonomous vehicles directly into questions concerning urban planning, commercial property, logistics, insurance, energy, dealerships and tourism.
The most important consequence of the AI-powered vehicle may therefore not be that people eventually stop driving. It may be that cities, buildings and businesses gradually stop being designed around the assumption that every vehicle requires a human driver.
Source: CIJ.World Research & Analysis Team
