Transport

Excited reports of new applications of Artificial Intelligence (AI) regularly fill the media, yet its technical functionality and prospective impacts are difficult to comprehend. ‘Artificial Intelligence’ is perhaps rather a misnomer, in any case. ‘Machine Learning’ may be better description of its core capability – employing massive computer power, specialist chips and neural net algorithms to absorb, analyse and utilise huge amounts of data and identify meaning and relationships, so as to be able to answer rapidly any questions posed. The most common approach, pioneered by ChatGPT, scrapes vast amounts of written text from the internet to ‘train’ the algorithm, though ‘steal content’ might be a more appropriate label, some believe. 

Other approaches of particular interest included DeepMind’s Alphafold, which absorbs protein structure data, both amino acid sequences and three-dimensional structures determined by physical methods, so as to be able to predict 3-D structures as yet undetermined from known amino acid sequences – a huge achievement for which two Nobel Prizes were awarded. DeepMind has recently used historic weather patterns to match current weather conditions and generate forecasts much faster and more accurately than conventional methods that depend on computer modelling of the atmosphere.

For transport, a potentially major AI application would be commercially viable road vehicle automation through its ‘understanding’ of all possible types of in-motion situations and incidents, but here I am struck by the slow rate of progress compared with other digital implementations. Perhaps the problem is the magnitude and range of behavioural and environmental variant conditions that must be coped with, and the requirement to achieve publicly acceptable rates of error that lead to crashes. Perhaps a new breakthrough in the underlying technology is still needed – the British business Wayve is one to watch, applying AI to automated driving. Or possibly we need to be clearer in how we think about technological innovation and its comparison with equivalent human behaviour.

There is a fairly fundamental distinction in the approach to innovation in digital industries compared with transport technologies and other traditional industries. In digital industries, product lifecycles are short, which means that rates of innovation must be high, so that missing out on even one generation of improvement can be fatal, as Andrew McAfee, an astute academic commentator, has pointed out. Moreover, sustained exponential improvement means that sudden shifts in capability occur, such as computers getting small enough to fit into our pockets, or broadband getting fast enough to permit online meetings without specialised equipment.

All this means that competition between suppliers is intense because potential markets are so large, and since many of these markets are winner-takes-all (or nearly so), coming second is a not a viable strategy. At the same time, there is a rich open-source software movement from which all businesses can benefit. Venture capital is available for tech start ups that promise early returns. The general approach to innovation is to be agile, to aim for the rapid development of a ‘Minimum Viable Product’, the first functional version of a product, designed to be launched quickly and cheaply, to gather user feedback and support. 

In contrast, product lifecycles in traditional industries based on mechanical and electric power engineering, are much longer. The Boeing 747 jumbo jet aircraft, for instance, was manufactured between 1968 and 2023, a 55-year production run, and even now President Trump is acquiring a second-hand model. The life cycle of a car tends to be around 6-8 years between full model changes. Manufacturing techniques are substantially proprietary, with only limited knowledge transfer between businesses. Patient capital is needed but can be hard to find, given alternative uses.

Another distinction between digital and traditional innovations is the scope for cost reduction. Moore’s law reflects the observation that the number of transistors in an integrated circuit doubled about every two years or so, a compound growth rate of some 40%, with commensurate cost reduction. The cost reduction curve for mechanical engineering arises from increased experience of manufacture and refinement of design, which are not transformative of the business. 

Andrew McAfee argues that organisational structures are consequentially different:

So there is a clash between both the approach to innovation and the organisational structure required for successful digital businesses, and that traditionally seen as appropriate for road vehicle manufacture.

But the operational environment for transport can certainly benefit from AI applications in stand-alone digital technologies such a digital navigation (AKA satnav in the road context), as offered by Google Maps, Waze, TomTom and others. These have penetrated the market quite fast, since the vehicle technology itself does not need to be changed. Similarly for offerings based on digital platforms, which make a market virtually rather than physically, exemplified by online retail pioneered by Amazon, a winner nearly taking all. Uber is an example of a commercially viable application of a digital platform in the transport sector that matches suppliers with customers, although it has only recently become profitable, after experiencing operating losses every year since its founding in 2009.

Yet putting digital and traditional mechanical/electrical engineering technologies together in a single product, the state-of-the-art modern car, has proved difficult, particularly for the legacy auto manufacturers, even at the present early stage of vehicle automation involving driver assistance and ‘infotainment’ – in-vehicle access to media and navigation, often via a personal smartphone. The transition to ‘driverless’ operations would be even more demanding since the solution suppliers are not in a position to fully control the operational environment on roads (adoption of automation in fixed track environments like rail is a much more persuasive concept).

A major complication for the auto industry is how, at the same time as pursuing automation, to accommodate the switch to electric propulsion, involving another new type of technology – battery chemistry – with many potentially conflicting requirements including energy storage, rate of power delivery, rate of charging, decline in performance over time, weight, safety and flammability. In the past, businesses with expertise in battery development and manufacture did not make vehicles, so they either partnered with auto businesses or sold their products into the market. For the vehicle manufacturers, the strategic decision had been whether to partner with a battery maker, who may or may not turn out to have a leading product, or to buy batteries as a commodity in the market, with the possibility of not having access to the best technology.

Battery development itself seems to fall somewhere between the digital and legacy models outlined above. As a field, electrochemistry is relative opaque, with apparently no theoretical framework that would delimit ultimate performance. The cost and speed of development has meant that action has shifted from academic labs, which generally publish their advances, to secretive industrial labs. The UK effort to promote academic research has taken the form of The Faraday Institution that has built a large, collaborative, multi-disciplinary research community, which, however, has as yet to deliver its full potential. 

The challenges faced by traditional car manufacturers generally are exemplified by the emergence of BYD (‘build your dreams’), now maker of the best-selling EVs in China. The business was founded as a battery manufacturer in 1995, buying a small car manufacturer in 2003 and subsequently building an EV product line based on in-house battery technology development, most recently including ultra-fast charging capable of adding up to 470 kilometres of range from a charging session lasting just five minutes. BYD has also announced advanced driver assistance capabilities at no additional charge (unlike Tesla, which makes a substantial charge for such capabilities). Warren Buffett, the legendary US investor, took a significant stake in BYD in 2008 when it was an obscure battery maker, an investment that reflected a prescient judgement and demonstrates the possibility of ‘picking winners’, even in this rapidly developing emergent market.

As well as addressing the strategic imperatives arising from vehicle electrification and automation, car manufacturers must now grapple with the uncertain global economy and in particular the impact of the Trump tariffs on market opportunities and supply chains that had been developed on the assumption of continued globalisation. Moreover, the Trump administration’s aversion to green technologies is a headwind to EV sales in the US.  The task of top management in this business environment is not enviable, and, indeed, the current turnover of CEOs is exceptional.

The incremental improvements that will emerge from strong competition will doubtless benefit consumers, but without being transformative of road travel. The historic transport technologies – railways, the bicycle, the internal combustion engine for a variety of road vehicles, the jet engine – each allowed a step change increase in speed of travel, which in turn permitted a corresponding step change in access to people and places, the true benefit of mobility. Vehicle electrification and automation do not allow a step change in either the speed of travel or in access, so the motivations for adoption are to improve the quality of the environment and of the journey – the latter a choice for consumers to make based on value for money.

A further issue that affects technological innovation is the wish to maximise safety and minimise risk to users and the public by means of regulation. For traditional transport technologies, regulation of vehicles, infrastructure and drivers is well established and improves incrementally. For digital technologies, particularly the application of AI, the speed of development and the commercial imperatives make regulation a contentious matter.  In Britain, we have comprehensive legislation in place to regulate automated vehicles (AVs), the Automated Vehicles Act 2024, based on thorough analysis by the Law Commission. This allocates responsibilities when a vehicle is functioning in self-driving mode, with regulations for implementation intended in 2026, as well as pilots of self-driving taxi- and bus-like services promised for next year.

While admirably clear and comprehensive in the style of traditional transport regulatory legislation, there must be a question as to whether this elaborate apparatus may deter deployment of AVs that exploit fast evolving digital technologies and algorithms, or propose creative novel operational environments. A contrasting approach is illustrated by a paper authored by researchers from Swiss Re, an insurer, and Waymo, Google’s AV business, which compared third party insurance claims, both property damage and bodily injury, for 25 million fully autonomous vehicles miles with a matched human driven mileage, finding a reduction of about 90% for the former. This impressive outcome suggests that standard insurance may be sufficient to ensure adequate safety of AVs, since any shortcomings in performance would result in higher insurance premia.

As noted above, digital navigation (AKA satnav) has quickly and extensively penetrated the road transport sector, the consequences of its benefits for users being available at low or no cost, no interaction with the vehicle’s core technologies, and an absence of regulation. The leading offerings predict journey times and provide traffic-aware route recommendations based on real-time road conditions, utilising data that includes live traffic flow information, incidents, and historical data to predict delays and suggest optimal routes. The providers, such as Google Maps, Waze and TomTom, are reticent about how they do this. So I asked Gemini, Google’s AI tool, how this was achieved, which provided a seemingly full explanation, summarised as follow:

“In summary, Google Maps predicts journey times by:

• Measuring the distance between the start and end points.
• Determining the fastest route using algorithms like Dijkstra’s or A*.
• Applying speed limits for different road segments.
• Overlaying real-time traffic data from various sources to understand current congestion.
• Utilizing historical traffic patterns to predict future traffic.
• Employing machine learning models to analyse and forecast traffic conditions.
• Considering user-reported incidents and other relevant factors.

The combination of these data points and sophisticated algorithms allows Google Maps to provide highly accurate and dynamic journey time predictions that adapt to changing conditions. The accuracy can still be affected by unforeseen events or the availability of real-time data in certain areas.”

I wondered whether more might be learned from DeepSeek, the Chinese AI app that requires much less training effort and cost than its rivals, and offers as an option more insight into the ‘thought’ processes involved. Deepseek’s answer begins:

“Okay, so I need to figure out how Google Maps predicts journey times. Let me start by recalling what I know. I use Google Maps a lot, and it usually gives me an estimated time of arrival that’s pretty accurate. But how does it do that? I guess it’s not just using the current traffic, right? There must be more to it.

First, maybe they use the speed limits of the roads. Like, if a highway has a speed limit of 65 mph, they might calculate the time based on that. But that seems too simplistic because traffic varies. Sometimes you can go faster, sometimes slower. So maybe they also look at real-time traffic data. I remember that Google collects data from people’s phones using the Google Maps app. If lots of people are on a particular road and their phones are moving slowly, Google can tell there’s a traffic jam and adjust the estimated time accordingly. That makes sense…..”

Deepseek concludes:

“To sum up, Google Maps likely uses a combination of real-time traffic data from GPS signals, historical traffic patterns, road characteristics (speed limits, type of road), user-reported incidents, and possibly external factors like weather and events. Machine learning models process all this data to predict how long a journey will take, considering the best route based on current and predicted conditions.”

Regrettably, this discursive conversational ‘exposure’ of thought process adds almost nothing, although both Gemini and Deepseek mention the use of machine learning. What does not emerge is the nature of the core requirement for predicting optimal routes and journey times in different and dynamic circumstances, which must be a real-time model of the road network and the vehicles travelling thereon. This might be termed a ‘digital twin’, a concept of growing interest, although at present this is largely limited to machine and asset performance monitoring.

A Digital Roads of the Future project, based at Cambridge University, is exploring how digital twins, smart materials, data science and robotic monitoring can work together to deliver resilient roads that can measure and monitor their own performance over time, to make the highway system safer and greener. However, this does not extend to consideration of real-time traffic conditions and the responsive behaviour of traffic.

A recent report on the potential economic benefits of digital twins, commissioned by the Department for Transport, suggested a figure of £110 million (present value over a ten-year appraisal period) for network capacity management through reduced congestion under business-as-usual conditions. Yet the possible complementary contribution of existing digital navigation providers was not considered

It seems that the providers of digital navigation are well advanced in modelling road network performance through digital twin construction of some kind, and application of the relevant rapidly-gathered real time data. What I find remarkable is the disregard of these ‘commercial’ developments by the research community. I am aware of only four relevant papers in the research literature, three of which I authored. Two are case studies (this and this) of the impact of ‘smart motorway’ schemes that failed to deliver expected benefits.

The third, my summary paper, identifies three behavioural changes as a result of the wide use of digital navigation:

• diversion of commuters and other local traffic to take advantage of faster travel made possible when the capacity of major roads is increased, thus pre-empting space intended for longer distance business users;
• diversion of traffic from congested major routes to minor roads, previously used only by those with local knowledge;
• the prediction of journey times under expected traffic conditions helpfully reduces uncertainty of time of arrival, which road users regard as the main detriment arising from road traffic congestion.

The fourth published paper is by authors from DeepMind, the London-based AI business, acquired by Google in 2014, who employed machine learning to improve the performance of Google Maps by comparing predicted journey times with outturn as observed at the end of a trip, and modifying the calibration of the model to bring these into line. This is a rare example of comparison of prediction with outturn for a transport model, with the aim of improving and better validating the model for future use.

This lack of consideration of the impacts of digital navigation applications in the peer-reviewed academic literature is remarkable, given the vast number of papers that model the expected impact of AVs (despite the present paucity of observation data) and the sizable number that discuss the impact of ride hailing (exemplified by Uber). It is equally surprising that road authorities disregard the impact that digital navigation is having on their networks, and the opportunities that likely exist for beneficial collaboration with the digital navigation providers. There is low-hanging fruit here, ripe for picking, unencumbered by regulatory barriers, new infrastructure requirements or significant public expenditure, given that the technology is already in place and in use by millions of road users. A potential boost to economic growth at low cost – what’s not to like?

One encouraging development is that consultancy SYSTRA is to use TomTom’s digital maps, location technology and traffic analytics to build transport models that replicate real-world traffic conditions, these to be used to perform analyses on the potential impacts of infrastructure investments, schemes and policies, and to support clients in evaluating outcomes. Orthodox transport modellers will surely need to consider whether their traditional approaches can compete with dynamic models like these, based on huge data inputs analysed by machine learning. If one element of the new digital technology era is a harbinger of significant change to current transport planning practice, this is surely it.

This blog post is the basis for an article in Local Transport Today of 17 July 2025.

In general, transport in the UK has been regarded as a well-regulated sector. It is not surprising, given the obvious potential for public harm, and the high profile of any errant behaviour or system failures. Vehicles, drivers, infrastructure, and service provision by operators are all subject to safety and customer protection regimes. But technology and business practice move on, generating new issues to be addressed. Generally, this happens on a rather piecemeal basis when new modes, technologies, or commercial models emerge, but not always then.

Accordingly, under the Boris Johnson government, the Department for Transport launched its Future of Transport Regulatory Review in 2020 in the form of a call for evidence, initially focused on three areas: micromobilty; flexible bus services; and Mobility as a Service. Other topics that have been considered under this broad initiative included the future of flight, and modernising road vehicle standards. There has also been a look at ‘regulatory sandboxes’ as a means to introduce temporary or specific place-based regulatory flexibilities to support innovation. Yet addressing such topics ad hoc seems to have so far yielded little in the way of useful regulatory advance in any of these fields, let alone others of concern such as secure digital payment systems, customer data protection, and standards of information provision, for example.

One area that has been subject to full consideration is the regulation of automated road vehicles (AVs), an area where the last government seemed to believe there to be a major new industrial development and employment opportunity. The Law Commission was asked to recommend a regime to govern responsibilities when vehicles are operating in driverless mode, whether as ‘robotaxis’ or individually owned vehicles. The outcome was the Automated Vehicles Act 2024, which had broad political support. There is now in place a comprehensive legal framework that developers, operators and insurers of AVs might see as helpful in making clear their legal responsibilities, which, however, might be seen as quite onerous compared with arrangements in countries where the main development of the technology is taking place, the US and China in particular. UK developers of AVs are limited in number and their commercial prospects are unclear.

To be sure, the regulatory regime being put in place for AVs is forward looking and intended to encourage innovation. But the implications for the wider transport system and the management of the road network as a whole, particularly in urban areas, are consequential matters that also need careful consideration, a topic recently reviewed by the International Transport Forum and the subject of an exploratory project report by the National Infrastructure Commission.

Another fast-moving technological development area is aviation. Here a comprehensive sectoral regime operated by the Civil Aviation Authority is well established to oversee safety, security, the efficient use of airspace for maximum public benefit, and to protect the consumer interest. The CAA has set up an ‘innovation hub’, the purpose of which is to create an environment where innovation in aviation can develop in line with the CAA principles. These include ‘pilotless‘ planes and drones. A recent review endorsed the view that CAA is a world class regulator that is fit for purpose, delivering high quality services to the aviation and aerospace industry and the consumer. Moreover, the CAA and DfT are consulting on the creation of a new UK Airspace Design Service that would act as a single guiding mind for modernising the design of UK airspace, exemplifying the CAA’s proactive approach to innovation.

A further sector-wide regulator is the Office of Rail and Road, which originally was solely charged with regulatory oversight of railways, including holding Network Rail to account for stewardship of its network, its funding, as well as protecting the consumer interest in respect of services provided by the train operating companies. The ORR carried out a consultation in 2015-16 on its approach to innovation, following a prompt from the government. The ORR has taken the view that innovation is led by the rail industry and where appropriate it would be supported by the overall regulatory framework, including the duty to promote improvements in railway service performance and promote efficiency and economy on the part of providers of railway services. So, compared with the CAA, the ORR’s approach is relatively passive.

Since 2014, the ORR also has some oversight of roads, but this is limited to the engineering performance and efficiency of National Highways in England in respect of its network, not to other roads. The ORR’s remit does not allow it to address the benefits to road users of new road investments – whether the outturn from investment in new road capacity meets expectations – nor whether the strategic road network is being efficiently utilised (unlike the CAA’s concern with the efficient use of airspace). Again, there is a lack of intention actively to foster technological innovation for the strategic road network, or to analyse and prioritise the optimal use of capacity, and identify where enhancement is most justified.

For road vehicles more generally, there are important examples of regulatory approaches that both drive and support technological innovation. Tackling the harms from tailpipe emissions of petrol and diesel enginned vehicles has been effected by regulations within the portfolio of the Department for Transport that have driven down emissions of air pollutants and by the ZEV Mandate, a legal requirement for a minimum percentage of each manufacturer’s new car and van sales to be zero-emission each year.  Regulations made by local authorities to create urban clean air zones (in London known as the Ultra Low Emission Zone) also help incentivise the switch to electric propulsion, a once-in-a-century transport technology innovation. While the propulsion systems and physical impacts of individual classes of vehicles are a DfT matter, concern for overall emissions and air pollution lie elsewhere – with the Department of Environment, Food and Rural Affairs (Defra) and Department for Energy Security and Net Zero (DESNZ).

When it comes to new modes of travel, the Department for Transport again seems to be following rather than leading, with actions in fits and starts, for example in making regulations allowing trials of rental e-scooters to be fast tracked and expanded in response to proposals from local authorities. The intention is to understand usage, safety, and environmental impacts, and to explore changing travel patterns since the coronavirus pandemic and as e-scooters become more embedded in public life. Only e-scooters participating in official rental trials may be used legally on roads. It is expected that a decision will be made in due course about legalisation of e-scooters as a vehicle class, and about the appropriate regulatory regime, which may include a requirement to park rental e-scooters in prescribed bays. E-bikes are already legal and treated as ordinary bicycles if of limited power, but the problem of inappropriately parked dockless rental e-bikes prompts complaints and pressure to regulate.

In contrast to the supportive approach to e-scooters, and a permissive approach to e-bikes, as innovative mobility technologies, there has been inaction in several other significant areas. This includes updating the historic taxi and private hire legal regime, with minimal recognition by the authorities of the merits of ride hailing as an innovative means of summoning a taxi, exemplified by Uber’s struggles to get acceptance in London. This involved a number of court cases, which found in favour of Uber, while the company lost some other cases concerned with the employment status of drivers, which nevertheless has not seemed to impact on the provision of the service. Uber’s digital request and booking/payment system has proved very popular with users as a better means of getting a taxi compared with licenced black cabs or conventional private hire vehicles, as well as being sufficiently attractive to drivers so as to offer customers an acceptable level of service. But this outcome was no thanks to a taxi regulatory regime having a remit to foster innovation. On the contrary, taxi regulation is historic, not modernised for the digital age, rightly concerned with the protection of customers, less defensibly having the effect of protecting the trade of established black cab drivers. There is a marked contrast between this reluctance to welcome innovative approaches to taxi services and the progressive stance towards automated vehicles, including robotaxis.

In terms of road usage, beyond the urban areas where e-scooters and ride hailing taxis operate, regulation of the inter-urban system roads is confined to vehicle and driving licence holding, vehicle type approval and performance – handled by a number of specific agencies like DVLA and DSA – with speed limits and other rules of the road embodied in the Highway Code – an historic regime that has not generally recognised advances in innovative technologies.

One technological innovation that has been widely adopted is digital navigation, applicable to all travel modes to replace conventional maps. In the road context, this is generally known as satnav and is offered by a number of providers including Google Maps and Waze via free-to-use smartphone apps, and others who provide input to equipment installed by vehicle manufacturers. A key element of many of these offerings is to indicate the fastest route in the light of prevailing traffic conditions. Digital navigation is changing the way drivers use the road network in three main ways: local users are attracted to new capacity on major roads, pre-empting the additional capacity intended for longer distance business users and so weakening the economic case for the investment; longer distance users can divert to local roads that offer faster journeys than on congested major routes, roads that are well suited to active travel; and predictive journey time information allows more efficient use of the network.

At present, provision of digital navigation is a free-for-all, funded by sales of direction-finding to businesses that wish to attract clients to their premises, as for Google Maps; or by the sale of services to vehicle manufacturers that fit navigation as original equipment, as for TomTom. Road authorities generally seem to be paying little if any attention to these developments. Yet there are opportunities being overlooked to make better use of the road network, both urban and inter-urban, including buses services.

Curiously, legislation has long been in place to allow providers of digital navigation to be regulated. This dates back to 1989 and requires what were then termed ‘dynamic route guidance systems’ that take account of traffic conditions to be licensed by the government. The intention was to facilitate the introduction of a pilot route guidance system, known as ‘Autoguide’, that had been developed by the government’s Transport and Road Research Laboratory and required roadside equipment to be installed – a good example of forward-looking regulation to support innovation. The licence could include conditions concerning the roads that should not be included in the route guidance, the provision to road authorities of information on traffic conditions, as well as the right to install roadside equipment. In the event, the Autoguide pilot did not proceed, due, I suspect, to the limited computing power available at that time. Nevertheless, the legislation has remained on the statute book although no licences have been applied for or granted.

The possibility of achieving more efficient use of the road network through provision of better information about traffic conditions to both road users and road authorities was recognised early on by John Wardrop, a traffic analyst at the then Road Research Laboratory, in a seminal paper published in 1952. He postulated two equilibria. The first states that, under equilibrium conditions, traffic arranges itself in congested networks in such a way that no vehicle can reduce its costs (time and money costs) by switching routes. For this to happen in practice, drivers would need to have perfect knowledge of all feasible routes and travel times. Digital navigation may be seen as improving such knowledge, thus enhancing network efficiency, yet with a number of independent providers that may offer conflicting advice, it is hard to assess to what extent increased efficiency is being achieved.

Wardrop’s first equilibrium assumes that road users make decisions without regard to the impact their choices may have on others – a ‘selfish’ equilibrium. According to his postulated second equilibrium, the average journey time would be at a (lesser) minimum if all users behave cooperatively in choosing their routes to ensure the most efficient use of the whole system. This would be the case if an omnipotent central authority could command them all which routes to take. Traffic flows satisfying Wardrop’s second equilibrium are generally deemed system-optimal, and the loss of efficiency from this to the selfish equilibrium is an example of what is known as ‘the price of anarchy’.

Economists argue that a more socially optimal outcome could be achieved if the costs imposed by the marginal road user on others, by adding to congestion, could be internalised by a congestion charge, thus modifying behaviour by reducing demand through higher vehicle operation costs. However, implementing road pricing is difficult in practice, and there are issues of equity, so a question worthy of investigation is to what extent a more socially optimal outcome could be achieved through flexing the routing advice offered by providers of digital navigation – ‘nudging’, not compulsion

The likelihood that improved operational efficiency could be achieved through digital navigation is suggested by the wide use of such routing and navigation systems by the competitive road freight sector. Everyday experience of online shopping indicates the use made by logistics businesses of digital technologies to manage, track and predict flows of goods, often offering delivery time slots of two hours or less, all done algorithmically. This points to techniques to achieve operational efficiency on congested road networks that might be extended to the generality of traffic in a way that could be far more cost effective and less carbon generating than civil engineering technologies.

The competing providers of digital navigation services are generally uncommunicative about their operations, while highway authorities appear to show no interest in the impact of this technology on the functioning of the road networks for which they are responsible. One kind of opportunity for mutual benefit would be when the network is under stress, for instance on the occasion of major incidents, peak holiday flows, bad weather and the like. It is probable, and certainly worth further investigation, that coordination between highway authorities and digital navigation providers could make better use of available capacity. There is also the possibility of improving operations at normal times, including avoiding routing through traffic via unsuitable minor roads. There is therefore likely to be scope for coordination that would improve outcomes for road users. This might be through the parties acting voluntarily for mutual benefit, or through the updating and implementation of the 1989 regulatory regime.

Although, by and large, the transport sector is functionally regulated to avoid calamity, there are obvious gaps that could usefully be filled to the benefit of users and to give innovators a sound basis upon which to progress their new ideas – or at least try them out with appropriate dispensation. This is particularly the situation in respect of the operation of the road network. As well as concern for safety and customer protection, all regulators should have a remit to encourage innovation that offers user benefits, since making effective use of existing infrastructure is preferable to creating new capacity, given the need to achieve the Net Zero climate change objective.

The new Labour government has set up a Regulatory Innovation Office, with the aims of supporting regulators to update regulation, speeding up approvals, ensuring different regulatory bodies work together smoothly, informing the government of regulatory barriers to innovation, and setting priorities for regulators which align with the government’s broader ambitions. This is an ambitious and potentially extensive agenda. Yet it will be disappointing if the emphasis solely on novel and exciting new technologies, a possibility suggested by one initial topic, the role of drones as a form of connected and autonomous vehicle technology – surely a rather niche market. In contrast, there is much scope for encouraging innovative approaches in the mainstream emerging transport technologies, where the new Office could hopefully play an influential role.

This blog post is the basis for an article that was published in Local Transport Today 0f 31 October 2024.