《英國交通研究實驗室：2025彌合差距：交通領域人工智能應用障礙與突破路徑研究報告（英文版）（19頁）.pdf》由會員分享，可在線閱讀，更多相關《英國交通研究實驗室：2025彌合差距：交通領域人工智能應用障礙與突破路徑研究報告（英文版）（19頁）.pdf（19頁珍藏版）》請在三個皮匠報告上搜索。

1、PPR2060AuthorsPhil MartinOmar Jamal Sanaz Bozorg Tejal Khullar Robin WorkmanMay 2025Overcoming the Barriers to AI Adoption in TransportBRIDGING THE GAP TRL Limited 2025This report has been produced by TRL Limited(TRL)under a contract with the UK Department for Transport.Any views expressed in this r

2、eport are not necessarily those of the UK Department for Transport.The information contained herein is the property of TRL Limited and does not necessarily reflect the views or policies of the customer for whom this report was prepared.Whilst every effort has been made to ensure that the matter pres

3、ented in this report is relevant,accurate and up-to-date,TRL Limited cannot accept any liability for any error or omission,or reliance on part or all of the content in another context.This report shall be cited as:Martin,P.,Jamal,O.,Bozorg,S.,Khullar,T.,Workman,R.(2025).Bridging the Gap:Overcoming t

4、he barriers to AI adoption in transport.Published Project Report PPR2060.TRL Limited,Crowthorne.DOI:10.58446/ansu5904PPR 20601.Establishing clear governance and regulation2.Balancing innovation and data protection through collaboration3.Enhancing technical skills and expertise4.Building public confi

5、dence in AI adoption5.Ensuring investment in infrastructure to support innovation and impactOur colour key(below)connects the key insights from the 8 key barriers(pages 7-21)to the 5 key opportunities to transform transport by overcoming these barriers(pages 23-28).FOREWORDEXECUTIVE SUMMARYNAVIGATIN

6、G AI ADOPTIONEnabling the Future of UK TransportRewriting the RoadmapFROM THEORY TO PRACTICETHE 8 KEY BARRIERS1.Data Discoverability and Quality2.Infrastructure for AI Integration3.Regulation and Governance4.Skills and Workforce Readiness5.Decision-Making and Accountability6.Public Trust and Social

7、Acceptance7.Cost to Implement8.Environmental ImpactsTHE 5 KEY OPPORTUNITIES1.Establishing clear governance and regulation2.Balancing innovation and data protection through collaboration 3.Enhancing technical skills and expertise 4.Building public confidence in AI adoption 5.Ensuring investment in in

8、frastructure to support innovation and impact ACKNOWLEDGEMENTS REFERENCES1233457791113151719212324252627282931CONTENTS12PPR2060 TRL Ltd 2025|However,achieving fundamental change and overcoming existing and perceived barriers to the adoption of AI will be an ongoing mission.I am delighted to endorse

9、this reports call for further coordinated effort across the transport industry to create an AI ready ecosystem within the transport sector,because by working together we can realise the benefits.Prof.Richard CuerdenDirector,Safety Advisory and TRL AcademyHonorary Professor of Transport Safety,Univer

10、sity of NottinghamI am reminded of Niccol Machiavellis warning that people and organisations dont like change.Of course,this is true some of the time,in some circumstances,and for some people.It is also true that there are a reasonable number of us,often the majority,who when helped with the adoptio

11、n process,embrace learning,and thrive through discovering new and innovative ways to improve our world and live our lives.The use of Artificial Intelligence in transport will become more prevalent.In my opinion,the choices we now face must focus on taking on the perilous conduct of leadership in thi

12、s space.Specifically,how will we enable the best AI solutions to be adopted at scale?The alternative is to be passive and watch on,as AI inefficiently and undemocratically integrates into the journeys we make,impacting accessibility,efficiency,choice,safety,cost,and data privacy and security.This pa

13、ssive option is unattractive,not least because it carries much greater risks to our economy,communities,and environment.The motivation to act is therefore clear.AI represents a significant opportunity to enhance mobility for everyone and to mitigate the unacceptable levels of harm caused by todays t

14、ransport systems every day in the UK about 400 people are killed or injured on our roads and transport accounts for a third of all carbon dioxide emissions.This work builds on a growing body of accredited literature,which describes the challenges and obstacles that many in the transport sector exper

15、ience when it comes to introducing AI-driven solutions.These barriers are technical and non-technical(organisational,societal,economic and environmental)in their origin,requiring diverse system-wide perspectives to be able to solve the wicked problem of improving the adoption of AI in transport.Brin

16、ging together a diverse group of over 60 stakeholders from across the AI and transport ecosystems,this work co-created a cross-sectoral approach for overcoming the barriers to the adoption of AI in transport.Having this strong foundational understanding and broad consensus on the nature of these bar

17、riers enabled stakeholders to collaborate,and through industry led workshops explore and develop practicable solutions.This brings me back to Machiavellis leadership point.All those who contributed to this activity should be rightly proud of their collective achievement.Because of their input,five k

18、ey opportunities have been identified to drive the change that would be necessary to improve the adoption of AI in transport.The power behind these recommendations is that they set a direction and give the public,private and academic sectors a common framework.“There is nothing more difficult to tak

19、e in hand,more perilous to conduct,or more uncertain in its success,than to take the lead in the introduction of a new order of things.”Niccol Machiavelli(1513)The Prince,Ch.6.The integration of Artificial Intelligence(AI)in the transport sector presents a significant opportunity to enhance efficien

20、cy,safety,and sustainability.Yet,AI adoption faces several barriers that hinder its full potential,including technical and non-technical challenges.These challenges contribute to a fragmented data landscape,slowing innovation and limiting AIs transformative potential.To address these issues,a struct

21、ured approach to overcoming these barriers is needed.This study aims to identify key barriers to AI adoption in the transport sector and to propose actionable solutions that will support AI integration in the UK transport network.To achieve this,the study employed a mixed-methods approach,incorporat

22、ing an extensive literature re-view and two industry-led workshops where experts from across the UK transport sector,academia,and AI providers participated.The literature review identified 12 key barriers,categorised into technological,organ-isational,societal,economic and environmental and cross-se

23、ctoral collaboration themes.The first work-shop,held at TRLs Smart Mobility Living Lab in London,brought together industry stakeholders in the trans-port sector to refine and prioritise these barriers based on real-world experiences.Some of the initial barriers identified were found to overlap and w

24、ere merged,resulting in a condensed final list of eight barriers.These eight barriers were then explored in greater depth during a second workshop,which was held at the Transport AI Conference in Manchester.Participants engaged in collaborative discussions,further refining the list of key barriers a

25、nd developing a set of strategic recommendations to address these challenges.The findings of this study provide a practical strategy to accelerate AI adoption in transport.Stakeholders identified five key opportunities to drive change:1.Establishing clear governance and regulation2.Balancing innovat

26、ion and security through collaboration3.Enhancing technical skills and expertise4.Building public confidence in AI adoption5.Ensuring investment in infrastructure to support innovation and impactThese recommendations highlight the need for a coor-dinated effort across the transport industry to creat

27、e an AI-ready ecosystem,ensuring that AI can effec-tively support a future efficient,safe and sustainable transport system.EXECUTIVE SUMMARYFOREWORD34PPR2060 TRL Ltd 2025|Enabling the Future of UK TransportThe UK transport sector is at a critical crossroad,as other industries rapidly advance in leve

28、raging AI to meet customer demands.Todays modern transport network faces an opportunity to revo-lutionise mobility in the largely untapped world of AI.Integrating AI into our daily travel is not just an option it is a necessity to efficiently tackle todays significant levels of emissions,persistent

29、conges-tion and safety concerns.Imagine a transport system where AI is embedded at the heart of every journey,where predictive algorithms anticipate and mitigate congestion before it happens,re-routing traffic for optimal flow.Envision a world where trains self-diagnose and self-repair,minimising de

30、lays and maximising efficiency.Picture a road network serviced by autonomous vehicles for personal and commercial use,communicating in real-time to streamline traffic management,update road users on roadworks and ultimately enhancing safety.Certainly,this sounds ambitious.algorithms are fragmented a

31、nd human oversight prohibits accelerating our industry to the next level.AI can help us to move beyond outdated legacy systems and give us the confidence to fully incorporate machine learning in our decision making.However,are we prepared for the ethical dilemmas that may arise?Who is accountable wh

32、en an AI-driven vehicle makes a critical decision?How can we ensure fairness and prevent algorithmic bias from sustaining existing inequalities?The successful delivery of meaningful change is made possible when we collaborate and adapt to overcome the challenges before us.These questions must be add

33、ressed head-on,not ignored,as we navigate the advancement of AI in the transport sector.The time for incremental adjustments has passed.The industry stands at a pivotal juncture,demanding a fundamental transformation to move from a reactive model,to one that is proactive and predictive.By embracing

34、opportunities for change and implementing AI,the transport sector has the potential to revolutionise mobility,redefining how people and goods move.This shift can directly address the evolving needs of passengers and freight customers alike,enabling a smarter,more adaptive and future-ready transport

35、system.Rewriting the RoadmapEach day,AI is integrated into various aspects of our lives.For instance,AI-driven systems provide real-time updates on bus and train occupancy,instantly notifying us of delays and allowing us to make informed travel decisions.Even parking has become more efficient,with t

36、he ability to reserve spaces in advance being introduced,eliminating the frustration of searching for a vacant spot.Traffic signal systems are able to dynamically adapt to congestion levels,facilitating an efficient journey.If a bus is full,AI can promptly detect overcrowding in real time and direct

37、 passengers to an alternative vehicle,ensuring a more comfortable and efficient journey.As intelligent transport system solutions continue to evolve,their potential to optimise mobili-ty is becoming increasingly evident.constraints,and a lack of leadership further impede progress.Transport data rema

38、ins fragmented without a culture that fully understands and values such data,despite developing technical expertise and strong incentives for investment.Despite these advancements,some scepticism persists within the industry.Many remain attached to traditional habits or question the extent to which

39、AI-driven systems can truly enhance our travel experiences.Issues related to data discoverability,privacy,security,and ethical considerations contribute to hesitation among transport users.Organisational resistance,legal and monetary What will the future of transport look like?Envision a world where

40、 AI goes beyond mere assistance,but anticipates,adapts,and refines itself to meet our needs before we even realise them.A seamlessly intelligent transport system that allows us to embrace a reality where hesitation fades,trust grows,and technology propels us towards a safer,more efficient,and sustai

41、nable future.The journey to AI-driven transport is just beginning.To reach this vision,we must first confront the barriers standing in our way.This paper explores these challenges in greater depth.In an increasingly interconnected world how can we safeguard data and give passengers the confidence to

42、 embrace AI in their daily travel?While AI is increasingly becoming a part of our daily journeys,its widespread adoption continues to encounter significant challenges.Existing barriers to the adoption of AI in the transport sector hinder innovation and limit the potential of AI to transform mobility

43、.NAVIGATING AI ADOPTIONThe integration of AI in transport faces numerous barriers that must be understood and addressed to fully unlock its potential.Data is often confined to silos,56PPR2060 TRL Ltd 2025|Methods OverviewThis report follows a structured approach to identifying and analysing barriers

44、 to AI adoption in transport.It incorporates insights from academic literature with industry expertise gained through two interactive workshops.The first workshop confirmed and prioritised the barriers,while the second explored the barriers in more depth to identify potential solutions.A growing bod

45、y of literature focusing on the barriers to AI adoption across many sectors underscores the critical role that overcoming these barriers has in unlocking the benefits of AI.To establish a solid technical foundation for this work,scientific and grey literature were both reviewed,including research re

46、ported in non-academic publications,policy papers and industry white papers.Barriers to AI adoption were identified from this research and categorised into four key themes:1.Technological barriers2.Organisational barriers3.Societal,economic and environmental barriers4.Cross-sector collaboration.Foll

47、owing the literature review,an interactive cross-industry workshop was convened at TRLs Smart Mobility Living Lab in London,bringing together key transport stakeholders.These included representatives from the Rail Safety and Standards Board(RSSB),Railway Industry Association(RIA),Rail Delivery Group

48、(RDG),Partners,Confederation of Passenger Transport(CPT),Transport for London(TfL),Local Council Roads Innovation Group(LCRIG),Research Institute for Disabled Consumers(RiDC),Alan Turing Institute(ATI),UK Research and Innovation(UKRI),London Legacy Development Corporation(LLDC)and several independen

49、t industry experts.This workshop began with an interactive presentation summarising key findings from the literature review,which identified 12 preliminary barriers to AI adoption.Participants were invited to engage with the findings through digital interaction tools,enabling them to prioritise barr

50、iers relevant to their specific domains.The group was then divided into two discussion panels,where they explored:How each barrier manifests in different transport sectors.Examples of specific projects where these barriers had been encountered.Prioritisation of barriers based on impact and feasibili

51、ty of resolution.A key outcome was the finding that some of the initial barriers identified were found to overlap,and so these were merged to result in a condensed final list of eight barriers.The discussions provided real-time insights into industry perspectives,refined the AI adoption barriers and

52、 informed the focus of a second workshop.The second workshop took place at the Transport AI Conference in Manchester,where 65 experts attended from across the transport industry,including rail,road,maritime,aviation,and urban mobility.Stakeholders came from a diverse range of organisations,including

53、 public transport authorities,consultancies,research institutions,and AI solution providers.This ensured a broad scope of perspectives and expertise,enriching discussions on AI adoption challenges.Eight tables were formed to facilitate focused conversations on AI adoption challenges,with each table

54、focused on one of the eight identified barriers.This session adopted a World Caf approach,where some participants remained at a table while others rotated to contribute fresh perspectives.The structure stimulated a dynamic exchange of ideas,fostering collaboration across industry sectors.The worksho

55、p objectives were to validate and refine the shortlist of 8 key barriers to AI adoption and explore the root causes of each barrier.The output was to identify potential solutions and opportunities for overcoming these barriers.To facilitate this,three key questions were deliberated by workshop atten

56、dees in regard to the barrier on each table:What are your experiences with the root causes of your barrier?What does a future transport system look like to you with your barrier removed?What actions should the transport sector take to overcome your barrier and achieve this future transport system?Fo

57、r each identified barrier,stakeholders provided valuable insights,outlining practical concerns and offering recommendations tailored to the address the specific challenges faced by organisations implementing AI.Following a thorough review of all discussions and suggested actions,the recommendations

58、were grouped into broader thematic categories to ensure a structured and consistent approach.This process resulted in the identification of five overarching opportunities for action to overcome these barriers to AI adoption.By highlighting these opportunities,a clear and actionable roadmap was creat

59、ed that reflects the real-world priorities of stakeholders,ensuring that AI can be successfully integrated into transport systems through practical,industry-driven solutions.FROM THEORY TO PRACTICE78PPR2060 TRL Ltd 2025|Definition:AI in transport faces challenges with using fragmented and often inac

60、cessible data due to proprietary restrictions which creates a lack of standardisation.This limits access to reliable,fair and ethical data AI-driven data and applications.Case Study Traffic Regulation Orders(TROs)Despite its purpose to serve cities and counties with regulations to observe road safet

61、y,congestion and local air quality,how this is enforced with AI remains unclear.Currently deployed by over 150 local authorities across the UK these regulations are often non-digitised and vary in formats.This makes it challenging to anticipate and understand how to improve our road network.For inst

62、ance,without access to high quality and timely data(particularly quality rich data),providing real-time updates for road closures remains delayed,hindering route optimisation for logistics companies,causing delivery delays and increased fuel consumption.4 In other words,urban transport systems that

63、can benefit from AI adoption currently suffer from fragmented data-sharing frameworks,particularly in smaller cities.5 Hence there is a need for centralised protocols and funding to ensure equitable access to AI capabilities.Without such a national framework in place,real-time traffic information op

64、erators cannot dynamically optimise routes,leading to inefficiencies.Root Causes Lack of centralised data management making it difficult to locate,access,and integrate data for AI applications.Incomplete,outdated,or inaccurate data leads to poor data quality and undermines the reliability of AI mode

65、ls.An unbalanced number of uniform data formats and metadata are challenges that limit the scalability of AI solutions.Literature Review Given the nature and the sheer volume and heterogeneity of data generated within the transport ecosystem,data encompasses diverse formats.Across all modes,includin

66、g road,rail,air,and maritime,data is multifaceted,from sensor readings and traffic flow data to passenger information and maintenance logs.As highlighted by the UK Government Digital Service report1,this fragmented data landscape makes it difficult to identify and access the specific datasets needed

67、 for particular AI applications.Transport data often resides in multiple public and private databases(e.g.,bus operators,train operating companies,local authorities).Gaining a consolidated dataset is difficult,particularly when commercial interests or privacy concerns limit data sharing.Thus,emphasi

68、s is placed on the need for standardised data formats and common data platforms to facilitate data sharing and interoperability across different stakeholders.2 Research has shown that the lack of standardised data governance frameworks and metadata schemas hinders the discoverability of relevant dat

69、asets.3 This lack of discoverability translates to increased time and resources spent on data acquisition and preprocessing,significantly slowing down the AI development lifecycle.How do we discover what data is available and how this can be accessed?We must create a clear pathway on how we manage d

70、ata ownership,data aggregation issues and data usage effectively.Legacy systems and outdated contractual agreements restrict access to critical data.This challenges users when requesting permissions in particular aggregated data from various sources.There is a need to recognise and specify key data

71、requirements early on in the planning stages of AI projects.By identifying the purpose of the data,organisations can condense large data volumes of information into actionable insights.National Data Library:Stakeholders agreed that a more open collaboration between academia,industry,and the public s

72、ector is needed.Offering standardised access to open source data allows users to locate required information in one place,accelerating innovation.Improve communication between systems.By standardising the data format systems across the transport sector,such as rail and road,better passenger experien

73、ces can be developed.Maximise data use by taking a multi-faceted approach to explore the technical,organisational,and financial considerations.Can we incentivise commercial enterprises to share data points?DATA DISCOVERABILITY AND QUALITY910PPR2060 TRL Ltd 2025|Definition:Transport systems currently

74、 lack the physical and digital infrastructure needed to support real-time AI deployment.Integrating AI with current frameworks and legacy systems is not only costly,but also complex.This leads to an ineffective deployment of AI at scale.Case Study Maritime Autonomous Surface Ships(MASS)Infrastructur

75、e gaps are not confined to road transport but expand to other sectors like maritime as a barrier to adopting Maritime Autonomous Surface Ships(MASS).These independently operated ships require remote control centers and operational control for each level of autonomy,necessitating upgrades to port inf

76、rastructure and along maritime routes.However,many ports and maritime hubs still depend on traditional logistics frameworks,limiting seamless integration.Key infrastructure challenges include insufficient 5G/6G coverage for real-time AI-driven decision-making,and outdated docking and refuelling stat

77、ions suggesting a lack of port readiness.However trials such as those concluded with Yara Birkeland,were often disrupted due to poor digital infrastructure,such as unreliable communication networks,which delay real-time decision-making for route optimisation and cargo transfers.10 For MASS to thrive

78、,governments and industry stakeholders must invest in smart ports,AI-compatible traffic management systems,and real-time data-sharing networks.Root Causes Conflicting communication networks between software and hardware due to rapid advancements in technology causes development to lag.Since AI syste

79、ms require vast amounts of data,and organisations often struggle with inadequate data storage solutions,risks such as data breaches and unauthorised access represent key root causes.Outdated legacy IT infrastructure and systems are often incompatible with modern AI technologies,making integration di

80、fficult and slowing down adoption.Literature Review Development of AI in transport systems requires robust infrastructure to support data sharing and real-time applications.This includes the systems and processes needed for collecting,storing,processing,and managing the vast amounts of data required

81、.AI-based transport systems often rely on real-time communication and connectivity to function effectively.This includes networks and technologies that facilitate the exchange of data between vehicles,infrastructure,and control centres.6 Reliable and high-speed connectivity is therefore crucial to d

82、eploying applications such as AVs,smart traffic management,and optimising EV charging based on user demand and connected infrastructures.7The transport industrys reliance on real-time data processing necessitates high-speed seamless connectivity via 5G networks and Internet of Things(IoT)infrastruct

83、ure.However,the fundamental network infrastructure to support these technologies must be present in both urban and rural areas,as without this the industry faces challenges due to limited broadband coverage.Beyond these challenges,unauthorised persons and persistence of that access is a significant

84、challenge for AI integration in transport infrastructure.Cyber security concerns and the reliance on IoT devices and interconnected systems increases vulnerabilities to cyber-attacks,data breaches,and unauthorised access.8 These threats can disrupt critical services and compromise the safety and eff

85、iciency of AI systems.9The need for interplay between software development and on-field operatives emerged as a critical area for improvement to enhance data utility.Cyber security is a significant challenge,as reliance on(IoT)devices,V2X and interconnected systems continue to expose a higher risk o

86、f encountering cyber threats.Should organisations allow open access to certain data points which could enable us to make better use of existing infrastructure?Stakeholders suggest that the rate at which we develop and expand our infrastructure can be achieved by capital revenue splits and exploring

87、the most beneficial case for businesses to adopt.Physical barriers,such as access to advanced connectivity(i.e.5G and 6G),must be accessible and balanced for all geographic regions to allow road and rail sectors to offer AI functionality without compromising on passenger data use concerns.INFRASTRUC

88、TURE FOR AI INTEGRATION1112PPR2060 TRL Ltd 2025|Definition:Outdated or fragmented regulatory frameworks often fail to address the unique challenges posed by AI,such as accountability,liability,and risk management.This results in inconsistencies between governing bodies(national,regional and local),s

89、uppliers and operators leading to a disjointed governance landscape.Case Study Regulation and Governance Challenges in Autonomous Vehicle DeploymentAutonomous Vehicles(AVs)are advancing rapidly,with OEMs integrating technologies that offer self-driving solutions.However,a key issue that impedes its

90、implementation is how we address the accountability and liability frameworks for these technologies.This creates regulatory inconsistencies that complicate deployment for suppliers and operators.This lack of consistency for suppliers and operators results in in varying safety,testing,and operational

91、 requirements across jurisdictions.This regulatory patchwork makes it difficult for manufacturers to produce AVs that comply with all requirements,increasing costs and delays.A study by Sheffield Hallam University in 2024 highlights the urgent need for harmonised governance structures to support AV

92、integration15.They identified that,for cities at the forefront of connected and automated vehicles,growth is underpinned by proactive policy making to support these vehicles.Hence,establishing clear national and international standards would streamline compliance and provide legal clarity for liabil

93、ity in AV-related incidents.Root Causes Rapid development of AI technologies often out paces the creation of clear and consistent regulations The lack of standardised frameworks such as universally accepted governance standards creates ambiguity,making it difficult to implement AI responsibly String

94、ent data protection laws(such as GDPR and DPA 2018)create data privacy and security concerns.Literature ReviewLegal frameworks within the transport sector have been described as outdated,particularly in logistics and public transit.11 The lack of harmonised governance models creates a legal vacuum t

95、hat leaves stakeholders uncertain about compliance requirements.12 For instance,the European Unions General Data Protection Regulation(GDPR)imposes strict data privacy requirements,which can conflict with the data-intensive nature of AI systems used in transport.13 This regulatory ambiguity discoura

96、ges investment and slows down the deployment of AI-driven solutions.AI governance remains a key issue due to organisational and technological barriers that hinder data control and liability attribution for AI-driven decisions.In the rail sector for instance,outdated legal frameworks and fragmented g

97、overnance slow AI adoption.RSSB is working to align the Rail Technical Strategy with the Whole Industry Strategic Plan(WISP),yet sector-wide gaps in liability,data sharing,and compliance remain.Tools like the Managing Disruption Toolkit and climate risk modelling require clear data ownership and har

98、monised contracts.Modernising these frameworks is critical to enabling AI-driven solutions that improve safety,sustainability,and operational resilience.14 This lack of clear policies on data ownership,privacy,and usage rights limits AI adoption.Stakeholders emphasised how using“all”data available t

99、o enable AI-driven innovations is challenged with concerns over data privacy,ownership,and competition.It became apparent these sensitivities create resistance to sharing data across stakeholders,hindering the creation of unified and interoperable AI systems that could enhance transport services.AI

100、is a broad subject area that requires clear and rigid governance.At present,the lack of collaboration and coherence hinders decision makers from leveraging their understanding to provide solutions faster.Existing tension surrounding data usage,particularly regarding commercial sensitivity,is driven

101、by a lack of harmonised regulations across different governance levels.Solving this barrier opens up faster progress to integrate AI in transport in a manner which is bounded by structural governance and strategy on how AI can be used and what levels of passenger data can be utilised.Consistency bet

102、ween government sources will allow for Transport strategy groups and local authorities to collaborate on future transport systems.Priority roadmaps and consistent forms of communication will provide stakeholders with the mechanisms and tools needed to share data between private and public organisati

103、ons,offering more flexible and adaptable regulatory frameworks.REGULATION AND GOVERNANCE1314PPR2060 TRL Ltd 2025|Definition:At present there is a shortage of skilled professionals with the technical expertise in developing,deploying and maintaining AI-driven systems.Resistance from within traditiona

104、l transport sector industries,where concerns about job displacement exist,increase resistance to a future AI-enabled transport system.Case Study Bridging the skill gap in the transport industryThe UK is believed to be facing a significant skills gap,with a shortage of engineers and AI specialists ne

105、eded to maintain and regulate vehicles equipped with Advanced Driver Assistance Systems(ADAS).With projections indicating that by 2030,44%of UK cars will incorporate such systems,this necessitates a workforce that is skilled and trained to ensure we continue to operate safe roads and highways.18At p

106、resent however,there is a lack of structured,widely recognised training programmes tailored to AI in transport and this in turn has resulted in limited awareness among operational staff about the practical use cases and benefits of AI,leading to resistance or underutilisation.Transport for West Midl

107、ands(TfWM)have identified a need for approximately 60,000 new workers over the next 15 years to support the regions transport infrastructure,including roles in digital and AI sectors.19However,without sustained investment in education and training,the adoption of AI in the UKs road transport sector

108、may continue to be impeded by this critical skills shortage.Root Causes The industry currently faces a lack of AI literacy and technical expertise making it difficult to manage AI effectively.There is a gap in accessible,high-quality training programs to upskill the current workforce,leaving employe

109、es unprepared to work with or alongside AI systems.There is a mismatch between education and industry needs,causing fears of role redundancies leading to a lack of motivation to engage with AI.Literature ReviewThe transport sector faces significant challenges in adopting and scaling AI-driven soluti

110、ons due to a persistent skills gap.The transport industry traditionally relies on manual and operational skills,leaving it ill-prepared for the data-driven demands of AI technologies.This skills gap is further compounded by the rapid pace of technological advancement,which outstrips the ability of e

111、ducational institutions and training programs to keep up.16 As a result,organisations struggle to recruit and retain talent with essential expertise in network systems,data engineering,machine learning,and AI ethics,which hinders progress on projects on solutions that could enhance efficiency and sa

112、fety.This shortage of technical capacity often forces organisations to rely on costly external expertise,making AI integration unsustainable in the long term.Reports from the CIHT and other studies emphasise the need for cross-modal training programmes and partnerships between academia and industry

113、to address these gaps and build a workforce capable of adapting to AI-driven workflows.17 Beyond technical skills,workforce readiness is also hindered by cultural resistance to change.Many employees in the industry are sceptical of AI technologies,fearing job displacement or increased complexity in

114、their roles.Without effective change management strategies and clear communication about the benefits of AI,organisations face significant challenges in fostering a culture of AI dependency.Many organisations lack the necessary technical expertise to utilise the full potential of AI,which hinders th

115、eir ability to integrate solutions effectively.As such there is a strong need for cross-sector leadership to direct and guide organisations through this transition.There is a need to accommodate and account for a learning curve phase to the adoption of AI,for which decision makers should factor in t

116、he additional cost and time needed to exercise this.Sourcing skilled professionals is challenging as most train in the public sector but pursue work opportunities in the private sector due to remuneration.Retaining talent is vital in creating the necessary workforce to implement AI effectively.Devel

117、op training that targets all levels of AI literacy and all applications for AI-based systems.These training initiatives should be designed to upskill and equip employees with the tools and understanding needed to operate and manage AI systems,reducing dependencies on outsourcing activities.Risks,suc

118、h as cyber security risks,may arise when inexperienced employees utilise AI software,such as ChatGPT and Gemini within the workforce.Focusing on capability development across the workforce can mitigate the risk of data breaches and security issues.Senior leaders have a responsibility to discuss the

119、use of AI within the workforce and the positive impact this can have when carrying out daily tasks.Employees should be supported and reassured of the benefits of AI to tackle stigmas surrounding redundancies.Fostering an open-minded workforce culture where learning and developing is encouraged will

120、help retain skilled professionals within the sector,thus accelerating AI implementation.SKILLS AND WORKFORCE READINESS1516PPR2060 TRL Ltd 2025|Definition:AI-related decision-making requires clear accountability frameworks.Uncertainty on who is responsible for the actions of an AI-based system,partic

121、ularly following a mistake,limits the ability of organisations to make informed,ethical,and transparent decisions,resulting in a reluctance to deploy AI at scale.Case Study GATEway ProjectTRL conducted a review to understand public perception and acceptance of driverless micro-transit vehicles in th

122、e GATEway project in Greenwich.21 Passenger experiences were analysed,focusing on safety,satisfaction,and willingness to pay for such services.While participants were generally open to the technology,a critical concern emerged:What happens if something goes wrong?This highlights the need for clear a

123、ccountability and robust decision-making frameworks.A key lesson learned from the trial was that participants felt more comfortable with a safety operator present,indicating hesitancy in fully relying on AI decisions.The lack of a clear liability model created reluctance among regulators and insurer

124、s,as uncertainty over responsibility in case of system failure posed a major barrier to widespread adoption.Furthermore,the project highlighted the necessity of clear policies on human intervention and liability assignment,as regulatory and ethical considerations remain central to building trust in

125、AI-driven transport.These findings reinforce the successful deployment of AI-driven solutions.While the technology holds significant promise,its successful deployment depends on establishing well-defined accountability structures and a transparent regulatory framework to ensure both public confidenc

126、e and operational safety25.Root Causes Lack of clear accountability framework can create a challenge in assigning responsibility when errors occur.The absence of legal and ethical understanding makes it difficult to validate AI decisions,build public trust,and ensure safe data management.Literature

127、Review Effective decision-making and accountability are critical for the successful adoption of AI in transport systems.Senior leaders,decision-makers,and project managers must possess foundational AI knowledge to ensure informed,transparent,and ethical deployment of AI technologies.Without such lit

128、eracy,organisations risk failing to align AI applications with public expectations,ethical standards,and regulatory frameworks.Research has shown the need for decision-makers to address human and social factors,such as trust and resistance to automation,while leveraging AIs potential for economic an

129、d operational improvements.20 Leaders equipped with AI knowledge are better positioned to communicate benefits,mitigate public concerns,and establish accountability frameworks for AI implementation.Stakeholders stressed the importance of strengthening organisational leadership in AI,ensuring that le

130、aders and decision-makers have the necessary knowledge and expertise to confidently guide AI adoption and integration.Stakeholders highlighted why we need to foster a cultural shift,transitioning from risk-averse and blame-focused mindsets towards a more open,experimental,and proactive approach that

131、 encourages innovation and continuous improvement.How can AI truly transform transport safety if systems remain siloed?Collaboration across different transport sectors is key.By enabling connected solutions,AI tools can be seamlessly integrated,ensuring that data flows efficiently between vehicles,i

132、nfrastructure,and traffic management systems to create a more adaptive and intelligent transport network.Stakeholders found that bureaucracy and uncertainty over AI implementation decisions results in multiple unnecessary approvals and a reduction in the efficiency of the AI innovation process.Stake

133、holders proposed that creating dedicated offices or teams can guide cross-sectoral AI implementation in an informed and structured manner.Stakeholders suggested the government should recycle AI solutionsrather than reinventing them,the sector should prioritise reusing existing AI tools where possibl

134、e to accelerate adoption.DECISION-MAKING AND ACCOUNTABILITY1718PPR2060 TRL Ltd 2025|Definition:Widespread scepticism about AI reliability,safety,and fairness,combined with the lack of extensive real-world testing,can impact public trust and acceptance of AI-enabled transport systems,preventing solut

135、ions from deploying at scale and benefiting society.Case Study Bus Open Data Service(BODS)The Bus Open Data Service(BODS),launched in 2020 under the Bus Services Act 2017,aimed to improve bus travel in England by providing open access to timetable,real-time vehicle location,and fare data.24 It uses

136、standards such as TransXChange(timetables),SIRI-VM(real-time tracking),and NeTEx(fares).Key milestones included deadlines for operators to publish this data between 2020 and 2023.Despite its benefits,BODS has been criticised for excluding accessibility information,affecting passengers with disabilit

137、ies.This resulted in distrust from vulnerable passengers from using public buses due to the lack of accessibility data in real-time systems,which caused missed connections and reduced travel independence.The findings indicate that good communications,along with AI literacy and awareness,is becoming

138、increasingly important to improve public trust and social acceptance.Without transparency and clear regulations,scepticism may continue grows hindering AI adoption timelines.Root Causes Public trust in AI is low because people are unsure whether it is safe,fair,or reliable.Lack of real-world testing

139、 and unclear accountability makes them worry about biased decisions.Misconceptions and fear of the unknown result in public concerns that AI could replace jobs or change their roles in ways they do not understand.Limited AI training and knowledge make this fear worse.Some senior staff prefer traditi

140、onal ways of working and see AI as disruptive.Without a clear AI strategy and proper training,adoption remains slow.Literature Review Public fearfulness around AI privacy in smart cities,particularly in mobility and governance systems,have been debated as key public concerns towards AI use.21 Extens

141、ive collection of personal data,and its potential mismanagement,raises concerns about surveillance and trust,leading to resistance to AI adoption.To address this,continuous public engagement and transparent data practices,including explainable AI,strong privacy regulations,and ethical frameworks sho

142、uld be considered to build trust and acceptance.Furthermore,social resistance to adopting AI technologies is often driven by fear of job loss,lack of understanding,and apprehension toward workplace changes.This resistance is not only seen among frontline workers but also among senior staff members,w

143、ho may prefer established methods and resist investing time in learning new AI-based systems.This is due to lack of clear communication and well-defined AI strategies,where employees often assume the worst-case scenario,further deepening their reluctance to adopt.Recently,Great Britain implemented A

144、I in MOT testing,using a data-driven approach to enhance efficiency,target resources more effectively,and maintain high safety standards.Currently each year there are 66,000 testers conducting 40 million MOT tests across 23,000 garages in Great Britain,with a team of 300 examiners auditing garages a

145、cross the country.23 AI-driven insights now analyse data,identify patterns,detect anomalies,and focus inspections where they are needed most,improving compliance monitoring and regulatory oversight,whilst also reducing unnecessary resource use.However,concerns arose about workforce displacement duri

146、ng this process,signalling the need for transparency when adopting AI-systems at scale.Stakeholders highlighted that the transport sector has struggled to convey AIs potential in a way that resonates with the public,with operators and strategic leaders driving misunderstandings and mistrust.Discussi

147、ons with stakeholders revealed that AI in transport requires changes in how trust is communicated.The lack of effective communication employed by the media on AIs benefits and implications has contributed to scepticism.Stakeholder feedback indicated passengers welcome the use of digital methods to i

148、mprove their journey experience.How can we replicate the same confidence through implementing AI without compromising on trust and the service delivered?It has been suggested by stakeholders to stop using the term“AI”without context:Many people react negatively to the term“AI”due to misconceptions a

149、bout its impact.Instead,clearly communicate AIs benefits to the user by explaining how it improves efficiency,safety,and convenience.Stakeholders emphasised that AI adoption should be presented in a way that highlights how it benefits transport,rather than replacing jobs.Transport networks should fo

150、cus on creating AI systems that work together seamlessly,rather than implementing isolated solutions.Ensuring AI is seamlessly integrated and provides real value will help reduce resistance from both the public and industry professionals.PUBLIC TRUST AND SOCIAL ACCEPTANCE1920PPR2060 TRL Ltd 2025|Def

151、inition:Adopting AI in transport requires substantial investment,with the returns on this in-vestment largely unknown,making the implementation of AI-driven solutions risky for public transport agencies and operators.This could limit AI-enabled benefits to those who can afford to invest,reducing acc

152、ess and widening the data-divide gap.Case Study EV Charging DesertsElectric vehicles(EVs)are widely recognised as a key solution for reducing carbon emissions and fostering sustainable transportation.However,the high capital cost of establishing EV charging stations has emerged as a significant barr

153、ier to mass adoption,especially in areas with limited access to funding or investment.Rural areas are therefore at risk of becoming“charging deserts”as commercial viability for private investment is questionable due to higher connection costs and lower demand.30A key challenge in these areas is the

154、high cost of grid connections and installation.Unlike urban centres,where population density ensures frequent use of charging stations,rural locations experience lower demand,making it difficult for investors to recover costs.As such the AI benefits that can be introduced through EV charging such as

155、 predicting demand and grid management may remain restricted to those who have access to financial resources.Root Causes Investment costs and funding gaps encompass the procurement of advanced hardware and software necessary for AI implementation.It also includes the substantial costs associated wit

156、h hiring or training a specialised workforce to manage these technologies.AI systems heavily rely on vast amounts of high-quality data.The processes of collecting and managing this data can be extremely costly,especially when dealing with varied data sources and formats.Not being clear on what data

157、is required and instead sourcing AI driven data before finalising the end goal.Literature Review AI has the potential to enhance efficiency and reduce costs in the transport sector,but economic barriers hinder widespread adoption.AI-driven predictive maintenance has been demonstrated as a cost-savin

158、g tool in multi-modal transport,yet high upfront costs and technical barriers necessitate tailored financial models.25 Public sector funding constraints further slow digital infrastructure upgrades and AI adoption,while limited digitalisation in port logistics reduces supply chain efficiency.26 Econ

159、omic challenges also stem from fragmented systems.Long-term cost savings can be achieved from digital transformation,but this requires harmonised policies and shared funding.Investment incentives and public-private partnerships are critical to overcoming these obstacles.At present,high costs are ass

160、ociated with digital and physical infrastructure,as well as the increasing cost of power for supply and cooling.27 AI can also support sustainability goals while cutting costs,as demonstrated by the AI-powered EV Charging Initiative,which uses predictive algorithms to optimise energy use and costs.2

161、8 Similarly,AI-driven traffic management can ease congestion and reduce costs,but high infrastructure costs also hinder the adoption of advanced technologies like autonomous vehicles and Intelligent Transport Systems.29 Addressing these economic challenges requires innovative business models,sustain

162、able funding,and strategic investments to balance costs and cost savings to unlock AIs full potential in transport.Securing targeted innovation funding is essential to lower the financial barriers to AI adoption in transport.This includes government-backed grants,industry funding pools,and venture c

163、apital incentives that specifically support AI-driven transport solutions,which are essential in lowering the financial barriers.Through the development of shared data centres and cloud-based infrastructure,multiple organisations will have access to a shared AI resource.Therefore,collaboration,espec

164、ially between public and private sectors,would reduce the duplication of efforts needed to address common challenges.It is important to encourage long-term financial models that reduce upfront capital costs,making AI adoption more accessible to a wider range of stakeholders.AI procurement processes

165、were found to vary widely,leading to inefficiencies and inflated costs.Establishing consistent procurement guidelines,frameworks,and pre-approved vendor lists would streamline purchasing and reduce risks.Cost of lost productivity should not be overestimated.The cost of implementing an AI-solution is

166、 not necessarily portrayed by the price of a single system.Understanding the economics of investing in infrastructure and data points due to market trends and rapid changes in development is a necessity to pricing the cost of implementation.The understanding of the costs of AI implementation can be

167、exacerbated by a skills gap,increasing reliance on costly external expertise.Investing in in-house training,knowledge-sharing,and collaboration can reduce long-term costs.COST TO IMPLEMENT2122PPR2060 TRL Ltd 2025|Definition:While AI-driven transport solutions can reduce the impact of transport opera

168、tions on the environment,the computational systems necessary to do so consume significant amounts of energy and water.Sustainable AI models are therefore needed in the drive to mitigate the impact of the transport sector on the environment.Case Study AI-Powered Route Optimisation in Maritime Shippin

169、gMaritime shipping is a major contributor to global CO2 emissions,projected to reach 5%by 2050 without intervention.One effective solution is AI-powered route optimisation,which analyses real-time data to identify the most efficient travel paths.Studies show this can reduce fuel consumption and CO2

170、emissions by up to 10%,lowering costs and environmental impact.34Traditional navigation relies on static planning,leading to inefficiencies and excessive fuel use.AI-driven systems dynamically adjust routes based on weather,ocean currents,and congestion,optimising efficiency.To maximise these benefi

171、ts,industry leaders and policymakers must invest in digital infrastructure and regulatory support.Without strategic adoption,shipping will continue to face inefficiencies and rising emissions.AI-powered route optimisation presents a scalable solution for a greener,more efficient maritime sector.Root

172、 Causes AI systems,particularly those used for complex tasks like autonomous driving and real-time traffic management,require vast amounts of computational power.This translates to substantial energy and water resource consumption by data centres,leading to increased environmental impact.Resource in

173、tensity of hardware production involves the extensive extraction process of rare earth minerals,which contributes to environmental degradation,Rapid advancements in AI technology are leading to frequent hardware upgrades,resulting in a growing volume of electronic waste.Literature Review The environ

174、mental impact of AI on commuting is complex and has resulted in a notable rise in worldwide energy usage with high levels of greenhouse gas(GHG)emissions.Research consistently pointing to the massive energy consumption of training and running AI models,particularly large language models.31 This incr

175、ease poses a twofold challenge:addressing the escalating computational requirements while reducing environmental effect.Further challenges associated with mitigating the direct and indirect environmental impacts has meant computing AI poses significant environmental repercussions.Direct impacts such

176、 as resource consumption are primarily negative in nature and are associated with using vast amounts of;energy demand,raw resources and water to power and cool data centres that process and compute data.32 Indirect impacts are those which stem from applications of AI such as smart grid technology or

177、 digital twin simulation.Nonetheless efforts are needed to measure how we advance AI processing and hardware without leading to increased levels of electronic waste,increased levels of GHG and over-extraction of raw materials.AI applications,including route optimisation,fuel efficiency improvements,

178、and support for green mobility initiatives have the potential to significantly reduce carbon emissions,enhance air quality,and promote eco-friendly transport options.AI-powered route optimisation has been shown to significantly reduce CO2 emissions,especially in congested urban areas.33 However,real

179、ising these benefits depends on robust infrastructure,regulatory frameworks,and efficient data-sharing mechanisms.There is a need to focus on climate-conscious and sustainable AI deployment strategies to minimise environmental and ecological harm.Resource-sharing models are required to balance deman

180、d for data centre resource to avoid using unnecessary energy.Stakeholders raised the question of“Are we over relying on nuclear energy as our source?”We should be inspired to solve how we can meet the growing processing demands of AI by utilising the cleanest energy sources and reducing the need to

181、extract rare earth minerals.There is a need to assess AIs resource consumption and its impact on achieving net zero goals.Improper disposal of e-waste can release hazardous substances into the environment,posing risks to ecosystems.It should be ensured that AI-driven systems calculate and consider t

182、he impact of energy requirements,emissions,ecological harm and waste before adoption.ENVIRONMENTAL IMPACTS23PPR2060 TRL Ltd 2025|Five key opportunities totransform transport as we know itUnclear rules,lack of responsibility,and business concerns make it challenging to implement AI in transport and h

183、ighlight the need for clear and consistent AI governance frameworks.These issues were discussed under Regulation and Governance,Data Discoverability and Quality,and a structure for AI Integration.Establishing a clear and consistent AI governance framework that fosters collaboration and removes regul

184、atory uncertainty is a necessity for the transport sector.A lack of harmonised regulations,unclear responsibilities,and commercial sensitivity are major barriers to AI adoption in transport.Creating a structured,coordinated and continuously evolving approach to AI governance,both nationally and loca

185、lly,will enable decision-makers at all levels to more effectively implement AI solutions.Data privacy,ownership,and competition concerns currently make it difficult for private and public organisations to share information transparently.This slows progress and prevents the development of unified and

186、 interoperable AI systems.Consistency and cooperation across the sector must be promoted and enabled,with effective governance practices collaboratively and openly shared.Instead of an overly rigid framework,action should be focused on a flexible approach to governance that balances accountability w

187、ith innovation.Developing a priority roadmap and establishing consistent communication channels between actors within the transport AI ecosystem will support understanding of current and future regulatory expectations,whilst allowing AI systems to evolve for a certain future.By encouraging a culture

188、 of collaboration,rather than control,AI can be integrated into transport systems ethically,securely,and effectively.Making a concerted effort to establish transparent decision-making processes and promoting shared responsibility will ensure AI adoption benefits both public and private sectors,while

189、 maintaining public trust.Establishing clear governance and regulationOPPORTUNITY 12526PPR2060 TRL Ltd 2025|Many transport organisations operate in isolation,restricting their access to essential data,knowledge and AI-ready resources.Commercial partnerships are required to improve data discoverabili

190、ty,resource sharing,and AI infrastructure development.While solutions exist that address this challenge,these channels should be better harnessed by all to bridge the gap in public and private sector cooperation.Shared data centres and AI resources must be developed as national assets to allow both

191、the public and private sectors to access computational power and high-quality datasets without duplication of effort and cost.Public-private investment should be incentivised to develop shared AI infrastructure,such as a National Data Library,to reduce the costs and increase efficiency of AI solutio

192、ns across the sector.To overcome concerns around commercial sensitivity,the creation of standardised frameworks for data access,security,and intellectual property rights is a necessity.A trusted data-sharing ecosystem,supported by clear regulatory protections,will give organisations confidence to co

193、llaborate without fear of losing competitive advantage.Balancing innovation and data protection through collaborationWorkforce culture,leadership,and technical skills lim-it the adoption of AI in transport.These challenges were raised in discussions on kills and Workforce Readiness,Decision-Making a

194、nd Accountability,and Public Trust and Social Acceptance.A proactive learning culture,leadership engagement,and workforce training must be encouraged to embed AI skills across the public and private sectors.A national AI skills strategy should be introduced,and local delivery mechanisms put in place

195、,to ensure the UK workforce from senior leaders to frontline workers have the knowledge and confidence to work with AI-driven systems.To bridge technical skill gaps,collaboration with industry,academia,and training providers to share technical data analysis skills,AI literacy,and ethical AI consider

196、ations will enable employees to work alongside AI systems effectively rather than viewing them as a threat.Strong leadership is crucial for AI adoption.Promoting AI literacy at the executive level must be considered to ensure senior leaders understand the potential,risks,and implementation strategie

197、s for AI.A lack of leadership engagement often results in slow decision-making,risk-averse behaviour,and missed opportunities for AI-driven innovation.Fostering a supportive work culture and promoting AI adoption through transparency,employee engagement,and reskilling opportunities will strengthen a

198、cceptance.Employees need to see AI as a tool that enhances their work rather than replaces it.The early implementation of good change management practices,such as open communication,clear role transitions and training,will ease resistance and create a workforce that embraces AI as an enabler of prog

199、ress.Enhancing technical skills and expertiseThe challenges of data sharing,resource collaboration,and AI infrastructure are key barriers to AI adoption in transport.These issues emerged in discussions on Data Discoverability and Quality,Infrastructure for AI Integration,and Regulation and Governanc

200、e.OPPORTUNITY 2OPPORTUNITY 32728PPR2060 TRL Ltd 2025|Gaining public trust and tackling widespread uncertainty about the benefits of AI,its negative impacts and perceived risk towards workforce displacement is a key opportunity for enabling AI adoption in transport.These discussions were put forward

201、in Public Trust,Social Acceptance,Skills and Workforce Readiness.Transport providers and technology developers must transparently and clearly communicate the benefits arising from AI solutions,rather than simply regurgitating the term AI or other jargon which can alienate users.By clearly explaining

202、 the positive impacts the AI solution will deliver,customers become more understanding of how it will be experienced in daily travel.Coordinated public awareness campaigns and user-friendly guides are essential stepping-stones to demystifying AI use cases to the end user.The industry must work toget

203、her to demonstrate how AI brings value to customer travel aspirations,or to support those in the industry working with it.Cross-sector consistency in messaging can be the trigger for much-needed cultural and attitudinal shifts.While there is a need to integrate systems to reduce friction,policy make

204、rs should implement robust data protection measures that ensure compliance with privacy regulations to alleviate concerns.The public will need assurance that their data is being responsibly and securely handled.Sensitively communicating how the risks of data security and protection have been ethical

205、ly managed is critical to building confidence.Mechanisms for enabling public involvement to collaboratively design and implement AI systems will be vital to building public trust in AI.Transport operators and AI developers should consult with a broad and representative range of stakeholders througho

206、ut the life-cycle of an AI product.They should provide clear,accessible information,such as how their AI systems are designed to function,their decision-making processes,and the steps they have taken to ensure physical and digital risks are appropriately managed.Building public confidence in AI adop

207、tionThe need to balance investment in physical and digital infrastructure with the intended operational benefits and environmental impacts can be both a driver and barrier to innovation for AI in transport.These challenges were raised in discussions on cost of Implementation,Infrastructure for AI In

208、tegration and Environmental Impacts.A key challenge to the adoption of AI in transport is overcoming the actual,and perceived,costs and impacts of AI solutions throughout their value chain.To enable this,an evidence-based approach for evaluating the value chain impact of AI solutions must be develop

209、ed.Bridging this evidence gap will empower organisations to undertake impact analyses and have greater certainty when developing business cases for implementing any future AI solution.A focus on securing targeted AI innovation funding through public-private partnerships is crucial for sharing the in

210、vestment burden,bridging resource and expertise gaps and accelerating AI solutions for social benefits.Collaboration between policymakers,local authorities,the private transport sector,and AI developers alleviates financial burdens and delivers pilot projects.Tax incentives,grants,or subsidies can e

211、ncourage transport organisations to invest in AI solutions and make adoption more financially viable.The opportunities for utilising exisiting physical and digital infrastructure cost-effectively should be explored to better leverage these assets for the benefit of future AI applications.To enable t

212、his,a national AI infrastructure registry should be developed to improve the discoverability and utilisation of these assets and collaboratively accelerate the early adoption of AI in transport.Longer-term national and local capital infrastructure programmes that focus on modernising physical and di

213、gital infrastructure are required to ensure the UK transport sector remains at the forefront of the global race in developing AI-driven transport applications.It was recognised that,although we have not yet fully exploited existing physical and digital infrastructure,enhancements in capacity will ce

214、rtainly be needed in the future with increased adoption of AI solutions in transport.It is vital to develop AI solutions that aim to mitigate their impact on the planet.These must align with net-zero strategies by using clean energy to power data centres and by minimising ecological damage.Strategie

215、s such as resource-sharing models to minimise our carbon footprint should be considered as a collaborative measure that results in reducing the impact of AI solutions.Ensuring investment in infrastructure to support innovation and impact.OPPORTUNITY 4OPPORTUNITY 52930PPR2060 TRL Ltd 2025|Acknowledge

216、mentsWe would like to extend our sincere gratitude to all participants of the stakeholder workshops for their invaluable contributions.Their expertise,insights,and collaborative spirit were instrumental in identifying and addressing the key barriers to AI adoption in transport and the opportunities

217、for overcoming these barriers.We acknowledge the participation of representatives from across all transport and mobility sectors,including public transport authorities,consultancies,research institutions,and AI solution providers.Invaluable contributors to these workshops were:Adam Parkinson,Digital

218、 Catapult,Senior Industry Lead TransportAdam Sobey,The Alan Turing Institute,Programme Director,Data-Centric EngineeringAmy Burns,Transport for the North,Assistant AnalystAndreas Zachariah,TravelAi,CEOAndrew Browning,SchemeFlow,CEOAndrew Hamilton,Yunex Traffic,Senior Product ManagerAndy Graham,Trans

219、port Technology Forum,Lead Connected VehiclesAnna Jordan,BridgeAI&Alchera Technologies,Transport Expert Working Group&CEOAshleigh Filtness,Loughborough University,Professor of Transport Human FactorsAtul Kumar,DIMTS LTD,General ManagerBarclay Gauld,Anxend,Chief Technology OfficerBob Hickish,City Sci

220、ence,Principal Transport&Sustainability ConsultantCatriona Swanson,Manchester City Council,Strategic Lead-Sustainable TransportCecilia Oram,Sustrans,Head of Programme,Behaviour ChangeChris Hillcoat,KPMG LLP,Associate DirectorClaire Stocks,Greater Manchester Combined Authority,Senior Strategy Officer

221、Dan Cullen,LCRIG,Head of Sector InsightsDaniel Chick,Journey Alerts Ltd,Chief Technology OfficerDan Piner,South Western Railway,Senior Design and Innovation ManagerDaniel Saunders,Basemap,Head of ProductsDavid Learman,South Yorkshire Mayoral Combined Authority,Software ConsultantDavid Milner,Create

222、Streets,DirectorDevon Barrett,Podaris,Chief Technology OfficerEmily McCaffery,Transport for Greater Manchester,Apprentice Project ManagerFazilat Dar,Transport for London,Data Science Manager Lead/Head of Profession Data ScienceFilip Kolodziejski,techUK,Autonomous Vehicles Programme ManagerFred Ewing

223、,Meridian Business Support,Head of Transport and InfrastructureGerard Butler,Transport for London,Service Performance Manager,Journey PlannerGordon McCullough,Research Institute for Disabled Consumers,CEOHannah Tune,Transport for Greater Manchester,ITS Development ManagerHelen Bowkett,Lower Thames C

224、rossing,Head of Traffic and EconomicsJames Bullen,Transport for West Midlands,Product Manager LeadJez Smith,Rail Delivery Group,Rail Data Marketplace Programme LeadJim Lupton,Independent Railway Engineering ConsultantJohn Bradburn,WSP UK,Technical DirectorJohn Taylor,Confederation of Passenger Trans

225、port,Operational Technical ExecutiveJonathan Raper,TransportAPI,CEOJoshua Jiao,Transport for the South East,Analysis ManagerKris Beuret,Independent ExpertLlewelyn Morgan,SYSTRA,Director of InnovationLuca Mitchell,Warwickshire County Council,Transport Data TechnicianMartin Tidd,Rail Delivery Group,Pr

226、ogram ManagerMartine Harvey,UK Research&Innovation,Digital Transport Innovation LeadMatt Toozs-Hobson,Digital Catapult,SpecialistMatthew Peck,AtkinsRalis,Innovation DirectorMax Sugarman,ITS UK,Chief ExecutiveNicholas Turner,London Legacy Development Corporation,Data and Digital LeadNick Reed,Reed Mo

227、bility,Independent ExpertNick Ruxton-Boyle,Citisense,Technical DirectorNicolas Collignon,Kale AI,CEOPeter Overbury,Zircon Software,Head of AI EngineeringPhil Evans,Travel&Transport,PartnerRichard Dolphin,Transport for Greater Manchester,ITS ManagerRick Rowe,Transport for London,Head of Policy Licens

228、ing&RegulationRobin Hay,Railway Industry Association,Technical&Innovation ManagerRobin Lovelace,University of Leeds,Professor of Transport Data ScienceRowan Davies,Loughborough University,ResearcherSam Knight,Transport for Greater Manchester,Policy OfficerSam Li,Transport for Greater Manchester,Seni

229、or Innovation OfficerSandhya Nagaraju,Connected Places Catapult,Senior Transport ModellerStelios Rodoulis,CIHT Bus Centre of Excellence,Head of Bus Centre of ExcellenceSteven Alexander,WSP UK,Director of Transformation&Digital InsightsStuart McLay,National Express,Head of RetailTao Cheng,UCL SpaceTi

230、meLab,ProfessorThomas Ableman,Freewheeling,FounderTom Weeks,Informed Solutions,Transformation Director&Chief Strategy OfficerVahib Puri,Rail Safety and Standards Board,Director of Sector Strategy and TransformationZhenzhen Wang,Transhumanity AI,Project ManagerStatement of AuthorshipThis report was p

231、repared by TRL researchers Dr Phil Martin,Dr Omar Jamal,Dr Sanaz Bozorg,Tejal Khullar,and Dr Robin Workman.All authors were responsible for the conceptualisation,research,and writing of this report.This included designing the report structure,designing and facilitating the workshops,analysing worksh

232、op discussions,synthesising key findings,and drafting the final document.Declaration of FundingSupport in organising the final workshop was provided by Landor Links,whose in-kind contributions helped facilitate an engaging and productive workshop.The Department for Transport(DfT)provided grant fundi

233、ng and guidance for the workshops,but had no further influence over the analysis or conclusions presented in this report.3132PPR2060 TRL Ltd 2025|REFERENCES1.Government Digital Service,(2025).Artificial Intelligence Playbook for the UK Government.Department for Science Innovation Technology Report.V

234、iew here.2.Hangl,J.,Behrens,V.,&Krause,S.B.,(2022).Drivers,and Social Considerations for AI Adoption in Supply Chain Management:A Tertiary Study.Logistics,6(63).View here.3.Rjab,A.,Mellouli,S&Corbett.,J.(2023).Barriers to artificial intelligence adoption in smart cities:A systematic literature revie

235、w and research agenda.Government Information Quarterly 40(3).View here.4.Benson,M.,Tyers,R.&Cunningham,S.(2024)Traffic Regulation Orders(TROs).House of Commons Library Research Briefing.Doucment Number CBP 6013.View here.5.Rahman,S.,Islam,M.,Hossain,I.,&Ahmed,A.(2024).Utilizing Ai and data analysis

236、for optimazing resource allocation in smart cities:a US based study.International journal of artificial intelligence,4(7),70-95.View here.6.Department for Transport,(2023).Transport Data Strategy Innovation through data.Transport Data Strategy Report.View here.7.Shankar Iyer,L.(2021).AI enabled appl

237、ications towards intelligent transportation.Transportation Engineering,5,1-11.View here.8.ODwyer,E.,Pan,I.,Acha,S.,&N.,S.(2019).Smart energy systems for sustainable smart cities:Current developments,trends and future directions.Applied Energy,237,581-597.View here.9.Alamoush,A.(2024).Trends in port

238、decarbonisation research:are we reinventing the wheel?Current Opinion in Environmental Sustainability,71(1).View here.10.Merz,M.,Grtli,E.I.,Mrkrid,O.E.,Tangstad,E.J.,Fossy,S.,and Nordahl,H.(2023).A gap analysis for automated cargo handling operations with geared vessels frequenting small sized ports

239、.Maritime Transport Research,5.View here.11.Chakwizira,J.(2022).Regulatory Frameworks,Policies,Norms and Standards.In:Odiyo,J.O.,Bikam,P.B.,Chakwizira,J.(eds)Green Economy in the Transport Sector.Springer,Cham.View here.12.Floridi,L.,et al.(2018).AI4PeopleAn ethical framework for a good AI society.M

240、inds and Machines,28(4),689-707 View here.13.Montero-Pascual,J.F.(2023).Creating a common European mobility data space.Robert Schuman Centre for Advanced Studies.View here.14.RSSB.(2022)What We Plan to Deliver in 2022-23.View here.15.Marson,J.,Dickinson,J.and Parkes,s.(2023)Is the UK ready for auton

241、omous vehicles?A comprehensive investigation into the technical,societal,and policy readiness.Sheffield Hallam University Research Archive.View here.16.Polydoropoulou,A.,Thanopoulou,H.,Karakikes,I.,Tsirimpa,A.,Pagoni,I.,&Tsouros,I.(2023).Creating a methodology matrix tool to research the effects of

242、automation on the transport labour force:A European focus.Transportation Research Procedia,72,1090-1097.View here.17.CIHT.(2023).The role of data and artificial intelligence in achieving transport decarbonisation.Chartered Institution of Highways&Transportation.View here.18.Institute of the Motor In

243、dustry(IMI)(2023)Meeting the demand for skilled vehicle technicians in the age of ADAS.View here.19.Wheeler,P.(2022)TfWM launches Transport Skills Academy to close 60,000 worker skills gap,Interchange UK.View here.20.Koh,L.Y.and Yuen,K.F.(2023)Public acceptance of autonomous vehicles:Examining the j

244、oint influence of perceived vehicle performance and intelligent in-vehicle interaction quality.Transportation Research Part A:Policy and Practice,178.View here.21.Fernndez-Medina,K.,Delmonte,E.,Jenkins,R.,Holcombe,A.,&Kinnear,N.(2018).GATEway Trial 1:Deployment of a micro-transit vehicle in a real-w

245、orld environment.TRL Limited.Prepared for Innovate UK.View here.22.Bezai,N.,Medjdoub,B.,Al-Habaibeh,A.,Larbi Chalal,M.,&Fadli,F.(2021).Future cities and autonomous vehicles:analysis of the barriers to full adoption.Energy and Built Environment,2(1),65-81.View here.23.Ubaldi,B.,Le Fevre,E.M.,Petrucci

246、,E.,Marchionni,P.,Biancalana,C.,Hiltunen,N.,Intravaia,D.M.,and Yang,C.(2019).State of the Art in the Use of Emerging Technologies in the Public Sector.OECD Working Papers on Public Governance,No.31.OECD Publishing,Paris.View here.24.Department for Transport(2025)Bus Open Data Service(BODS).View here

247、.25.Wiese,T.(2024).Predictive Maintenance Using Artificial Intelligence in Critical Infrastructure:A Decision-Making Framework.International Journal of Engineering,Business and Management(IJEBM),8(4),1-4.View here.26.Yang,Y.,&Hsieh,Y.(2024).The critical success factors of smart port digitalization d

248、evelopment in the post-COVID-19 era.Case Studies on Transport Policy,17 View here.27.Chen,L.,Chen,Z.,Zhang,Y.(2023).Artificial intelligence-based solutions for climate change:a review.Environmental Chemistry Letters,21.View here.28.Wolniak,R.,&Stecua,K.(2024).Artificial Intelligence in Smart CitiesA

249、pplications,Barriers,and Future Directions:A Review.Smart Cities,7,13461389.View here.29.Shahedi,A.,Dadashpour,I.,&Rezaei,M.(2023).Barriers to the sustainable adoption of autonomous vehicles in developing countries:A multi-criteria decision-making approach.Heliyon,9(5).View here.30.Cruden,A.and Jone

250、s,C.(2024).Watering the EV charging deserts and leaving no-one behind?View here.31.Organisation for Economic Co-operation and Development(OECD).(2022).Measuring the Environmental Impacts of Artificial Intelligence Compute and Applications:The AI Footprint.OECD Publishing.View here.32.Vinuesa,R.,Aziz

251、pour,H.,Leite,I.et al.,2020.The role of artificial intelligence in achieving the Sustainable Development Goals.Nature Communications,11(1),p.233.View here.33.Abduljabbar,R.D.(2019).Applications of Artificial Intelligence in Transport:an overview.Sustainability,11(189),2-24.View here.34.Omdena.(2024)

252、.AI-powered ship routing reduces fuel consumption and emissions.View here.33AI in transport offers significant potential for efficiency,safety,and sustainability,but its adoption faces challenges that must be overcome to unlock its potential.Following an extensive literature review and discussion of

253、 AI use in transport across two industry-led workshops,eight technical and non-technical barriers were identified as key areas needing attention.These findings were developed into five strategic opportunities for the sector:establishing clear governance and regulation,balancing innovation and data p

254、rotection through collaboration,enhancing technical skills,building public confidence,and ensuring infrastructure investment to drive impact.Collaborative action on these opportunities can help create an effective AI-ready ecosystem to modernise our transport network.DOI:10.58446/ansu5904PPR2060 TRL Limited 2025