Automation is often discussed as a story about vehicles. But the deployment of Connected, Cooperative and Automated Mobility (CCAM) is, just as much, a story about people and jobs, which ones fade, which ones transform, and which entirely new professions emerge. RESKILLING's  published Deliverable D3.1 puts hard structure on that question, and the picture it paints is far more nuanced than "robots replace drivers."

Led by ZLC, D3.1 consolidates evidence from previous EU projects, 22 scientific and policy documents, and 60 real-world job advertisements into a single, standardised map of the CCAM workforce. Here's what stakeholders across the value chain need to know.

The headline: 110 jobs, 33 job families, one common language

At the heart of the deliverable is a mapping exercise that identifies 110 distinct job profiles impacted by CCAM and consolidates them into 33 functional job families. Crucially, every profile is classified using the international ISCO-08 occupational codes and skill levels, and aligned with ISCED education levels.

Why does that matter? Because it turns a sprawling, hard-to-compare list of emerging roles into a shared vocabulary that employers, educators, public authorities and workers' organisations can all use, and it connects each role directly to the training pathways that will be designed later in the project (Task T4.3).

The mapping is deliberately cross-modal: while road transport is the primary focus, most roles apply across road, rail, urban and maritime contexts, and to both passenger and freight mobility. Predictive Maintenance Analysts, IoT Developers, Cybersecurity Experts and MaaS Platform Managers don't belong to a single mode, and neither should the training built for them.

A value chain, not just a technology stack

One of the distinctive features of the RESKILLING approach is that it frames CCAM as a complete ecosystem, not merely a set of technologies. The deliverable organises roles across nine functional stages of the value chain:

1. Research & Innovation (a driver of CCAM rather than an outcome of it)

2. Design & Development: engineers, software and systems developers, cybersecurity, data science, HMI/UX

3. Manufacturing & Assembly

4. Recycling & Circular Economy: including EV battery second-life and decommissioning

5. Infrastructure Deployment & Operations

6. Mobility Services & Logistics

7. Policy & Regulation

8. Market & Commercial Development

9. Training & Education

This layered view makes something important visible: the technical core of CCAM is essential, but so are the legal, commercial, educational and societal roles that determine whether automated mobility is actually adopted, trusted and made inclusive.

CCAM Value Chain. Source: ZLC, 2025.

Who disappears, who transforms, who emerges

The analysis is candid about disruption, but it resists a simple "jobs lost vs. jobs gained" framing. Three dynamics run in parallel:

Roles that transform. Existing professions such as mechanics, clerks, planners and drivers don't simply vanish; they shift toward digital oversight, incident management, remote supervision and service personalisation. A driver increasingly becomes a supervisor of automated functions; a mechanic increasingly diagnoses sensors, electric drivetrains and V2X components.

Roles that emerge. Entirely new professions appear, including Autonomous Systems Experts, AI Specialists, AMoD (Autonomous Mobility on Demand) services modellers and programmers, Software Integration Engineers (V2X), Electric Car Battery Support Agents, and Warehouse Assistants for Smart Hubs. Some, like the Insurance & Risk Assessment specialists for autonomous systems, are established professions opening an entirely new branch of knowledge that must be built "almost from scratch."

Roles that contract. Manual, routine tasks, much of the daily work in freight, warehousing and construction, are progressively automated. Many of these lower-skill profiles will diminish, though certain conventional roles will persist by redirecting toward the specific tasks that still genuinely need human hands and judgement.

The deliverable is clear on the overall centre of gravity: CCAM workforce transformation is predominantly linked to highly qualified profiles, with ISCO skill levels 3 and 4 accounting for more than 70% of the job families. But levels 1 and 2 persist in support and customer-facing roles, travel attendants, charging station attendants, and warehouse assistants, and these workers will need reskilling in digital literacy, safety protocols and safe interaction with automated systems to stay relevant.

The skills that will define readiness

Across the literature and the 60 analysed job advertisements, the same priority skill clusters surfaced again and again:

• AI and data analytics

• Cybersecurity and privacy

• Functional safety: standards such as ISO 26262 and SOTIF

• V2X interoperability and connectivity

• Remote supervision and telematics

• Human factors and UX / accessibility design

Just as important, transversal skills, communication, adaptability, and stakeholder engagement remain essential across every profile. The job-advertisement analysis also confirmed a strong current bias toward highly skilled technical roles concentrated in European R&D hubs (Belgium, the Netherlands, Germany), with blue-collar CCAM vacancies still scarce. RESKILLING reads this not as reassurance but as a warning: the window to prepare vocational and operational profiles is now, before demand peaks.

A taxonomy that ties skills to automation levels

Perhaps the most practical output is a CCAM-specific skills taxonomy, built by combining insights from the Drive2theFuture project with the European ESCO classification, and mapped against the six SAE levels of automation (0 to 5).

For each job family, the taxonomy lays out the relevant skills and, using a simple green / orange / red colour code, shows whether each skill is fully needed, partially needed, or no longer needed at each automation level. The result is a visual roadmap of how a profession's skill requirements evolve as automation deepens.

The taxonomy surfaces four insights that matter for anyone planning training:

• Qualified professions (e.g. Engineers) don't disappear; they upskill. Their core competencies persist across automation levels, but require continuous adaptation.

• Lower-skill profiles (e.g. Freight & Warehouse workers) see many tasks fall away as automation advances, pushing toward reskilling into higher-qualified, technology-linked roles.

• Technology-focused professions (e.g. Data Science & Analytics) grow in importance and must themselves keep evolving.

• Emerging professions demand genuinely new skill sets, from AM (additive manufacturing) process engineering to CCAM-specific risk and liability expertise.

Read the Deliverable now!

How large AI's footprint has become

The report situates AI within a wider finding: roughly 1 in 10 job postings in advanced economies now lists at least one new skill, about half that rate in emerging market economies. AI-related skills are a growing share of that total, now accounting for nearly one-third of all new IT skills.

Figure 3 tracks the trajectory. In the United States, the country with the longest posting series ,AI skills appeared in fewer than 1 per cent of postings before 2015 and reached almost 5 percent by 2025. Generative AI recorded the single largest absolute increase in demand in 2024 across all four advanced economies in the sample. Citing Cazzaniga and others (2024), the note reports that 40 percent of global employment is potentially exposed to AI: 60 percent in advanced economies, 40 percent in emerging markets, and 28 percent in low-income countries.

Cross-country, Figure 3 panel 2 shows Denmark and the United States with the highest prevalence of AI-related postings, while Brazil and South Africa sit below 2 per cent.

The AI skill split: users versus developers

The note separates AI-user skills (using generative-AI tools such as ChatGPT or GitHub Copilot) from AI-developer skills (building and deploying models). Within U.S. AI-skill demand in 2024, roughly half of postings mention only AI-user skills, one-fourth only AI-developer skills, and one-fourth both (Figure 3, panel 1). AI-user skills are distributed broadly across fields of study, while AI-developer skills are concentrated among ICT and STEM graduates, who account for nearly 60 percent of all AI-developer skills (Figure 11, panel 3).

Wage returns to AI skills

Figure 5, panel 4 reports the wage premiums associated with AI skills, and they differ by country and skill type. In the United Kingdom, both AI-developer and AI-user skills carry posted-wage premiums of roughly 7.5 to 8 percent within occupations. In the United States, the premium above 8 percent is concentrated in AI-developer skills, while AI-user skills show a smaller premium near 2 percent. For comparison, other non-AI new skills in the United States carry a premium of about 2.5 percent. (For any single new skill of any type, the baseline premium is about 3 to 3.4 percent, per Figure 5, panel 1.)

AI and employment: the exposure-complementarity divide

This is the report's most consequential AI finding, and it separates AI skills from new skills generally. On average, the demand for new AI skills has so far not boosted overall employment in U.S. local labour markets, whereas demand for non-AI new skills is associated with significantly higher overall employment (Figure 9, panel 4).

The effect turns on whether an occupation is complemented or substituted by AI. Figure 9 disaggregates by exposure and complementarity:

• High-exposure, low-complementarity occupations (about 30 percent of total employment): employment levels run 6.3 percent lower than other regions five years after AI-related skills appear, a predicted decline of about 3.6 percent given the observed increase in AI-skill postings (Figure 9, panel 1). Workers in these occupations also see modestly lower wages two to three years after AI-skill entry.

• High-exposure, high-complementarity occupations: effects are statistically insignificant (Figure 9, panel 2).

• Low-exposure occupations: employment is about 4 to 5 percent lower in years one and three, possibly reflecting spillovers, but insignificant otherwise (Figure 9, panel 3).

Supporting this, Figure 8 panel 2 (drawing on Pizzinelli and others 2023) finds that a one-standard-deviation higher AI adoption at the commuting-zone level, about 0.18 percentage point — is associated with a 0.4 percentage point lower vacancy share and 2.5 percent lower growth of high-exposure, low-complementarity vacancies, relative to high-exposure, high-complementarity occupations.

Young workers and entry-level roles

Figure 8, panel 1 shows that young workers, and college-educated young workers in particular, are more concentrated in high-exposure, low-complementarity occupations than prime-age or older workers. The note reports evidence (Brynjolfsson, Chandar, and Chen 2025) that since the release of ChatGPT, early-career workers aged 22 to 25 in the most AI-exposed occupations experienced a 13 percent relative decline in employment, while employment held stable or grew for less-exposed fields and more experienced workers in the same occupations.

Where transport occupations sit

The note contains no AI-specific figures isolated to the transport sector, but transport appears at two points in the data.

First, in the occupational dynamics of Figure 1, drivers and freight/cargo handlers are named among the fastest-declining occupations, "nonprofessional and routine-intensive occupations… highly vulnerable to the latest wave of automation." Freight and cargo handling is plotted explicitly in the decreasing group in the U.S. data.

Second, on the skills side, "security and transport" ranks among the highest fields of study for workers reporting new skills, driven especially by cybersecurity, and it also sits near the top for AI-related new skills (Figure 11, panels 2 and 3). In Lightcast's taxonomy, transportation, supply chain, and logistics is treated as a distinct skill field (Annex Table 1.4).

Read against the exposure-complementarity finding, the transport data divide along the same line the report draws economy-wide: routine driving and material-moving roles fall in the declining, automation-exposed category, while the transport-adjacent work reporting new and AI-related skills is concentrated in security, systems, and logistics.

It is important to note at the outset that this is a United States study. The authors, Shubham Agrawal and colleagues at Clemson University and Michigan State University, draw on a nationwide US survey of 922 professional drivers (274 truck, 341 gig, and 307 taxi drivers) conducted in 2022, under a National Science Foundation project titled "Preparing the Future Workforce for the Era of Automated Vehicles." The sample was predominantly male, without a bachelor's degree, and relatively young, which is characteristic of these US occupations but not necessarily of their European counterparts.

Three types of driver

Using latent profile analysis, a method that groups people by patterns across several attitudes at once, the researchers identified three distinct profiles among drivers.

Reluctant Acceptors, the largest group at 45.6 per cent, held the least positive views of automated vehicles. They expected automation to shrink their industry's workforce over time and were willing to stay in their jobs only while automation remained partial; their attachment fell sharply as more of the driving task was taken over.

Enthusiasts, 31.5 per cent, held markedly more positive views. They expected automation to take over routine driving tasks, anticipated staying in their jobs, and, notably, expected their industry's workforce to grow rather than shrink. They also reported higher incomes than the other groups.

Cautious Endorsers, 23.0 per cent, were mildly positive about automated vehicles but expected adoption to arrive later and then spread quickly. They were less confident than Enthusiasts about remaining in their jobs and tended to have lower incomes.

Income was the only demographic characteristic that differed significantly across the three groups; age, gender, race, education, and driver type did not. The authors also note that no strongly anti-automation profile emerged, suggesting outright resistance is a minority position among current drivers, though they acknowledge an online survey may under-reach the most technology-sceptical workers.

The decisive factor: who pays

The study's most policy-relevant result concerns how training willingness varies with the funding model. Drivers were asked about three scenarios: paid on-the-job training, free training outside the company, and training they would have to pay for themselves.

Self-paid training was the least popular option across all three groups, and least popular of all among Reluctant Acceptors. Enthusiasts, the wealthier group, were the most willing to pay for their own training. Reluctant Acceptors, by contrast, were more willing than Enthusiasts to undertake employer-funded, on-the-job training. In short, the workers least positive about automation are not unwilling to train, but they are far more likely to do so when their employer bears the cost and the training happens on paid time.

What the authors recommend

The paper's recommendations are addressed to US workforce policy, and this is where the difference in context matters most for European readers. The authors suggest, for example, that federal programmes under the US Workforce Innovation and Opportunity Act could reimburse employers for on-the-job training in regions where Reluctant Acceptors predominate, while workers with Enthusiast attitudes might reasonably be expected to invest in their own training. They also argue that assessing a single attitude is not enough: employers and regional agencies should profile their workforce more fully before designing training or career-transition support, and that because willingness to train is generally high, relatively modest public investment may suffice to encourage it.

What about Europe?

European labour markets differ in ways that bear directly on the findings: driver demographics are not identical, vocational and sectoral training systems are more established in several member states, social dialogue and works councils play a larger role, and automated mobility in Europe is expected to enter first through public transport and shared services rather than private robotaxis.

What does appear likely to transfer is the underlying lesson: a driver workforce is not uniform in its readiness to retrain, and the willingness to participate is shaped by both attitude and the terms on which training is offered. Programmes that assume a single, homogeneous response, or that place the cost of retraining on the workers least able or willing to bear it, risk reaching precisely the wrong people. Segmenting the workforce, and matching the funding model to each group, emerges as a principle worth testing in European conditions rather than a conclusion to be imported wholesale.

Published in the Journal of Urban Mobility in 2026, "Barriers to autonomous vehicles adoption in Europe: Insights from literature and interviews," by Giuseppe de Leo and Giovanni Miragliotta of Politecnico di Milano, combines a review of 38 academic articles with 21 semi-structured interviews with European AV experts. From this evidence the authors distil six interrelated barriers: fragmented regulation, unresolved ethical and legal liabilities, technological and cybersecurity limits, inadequate infrastructure and V2X integration, uncertain economic models, and low societal trust.

What is striking, for anyone focused on skills, is what is not on that list. "Skills," "workforce," and "training" are nowhere among the six headline barriers. Instead, they surface indirectly, folded into other categories, raised in passing by interviewees, and deferred to the study's agenda for future research. That positioning is itself the story.

The clearest skills finding: regulators who can't keep up

The paper's sharpest observation about skills concerns not drivers but the public sector. Discussing regulatory bottlenecks, the authors report an interviewee's account of autonomous shuttles sitting idle in an Italian city while awaiting a national ministry's authorisation, and, more pointedly, the same expert's warning that "public authorities often lack the technical expertise to evaluate and approve AV pilots, resulting in delays."

The authors flag this as a practitioner insight rarely captured in academic work, which tends to assume regulators will eventually adapt. It is, in effect, a public-sector skills gap: the people charged with approving and overseeing automated mobility may not have the engineering and data literacy to assess it. Another interviewee's call for policymakers to be "involved from the development phase onward to craft laws in parallel with technology" points to the same need — regulators who can engage with the technology on its own terms rather than react to it.

Workforce displacement: acknowledged, then set aside

Where skills touch the wider labour force, the paper is notably brief. Potential job losses for drivers appear in the study's summary of prior research. for instance in the barrier lists drawn from Raj and colleagues and from Ullah and colleagues, where "employment shifts (job losses for drivers)" are recorded as a societal and economic impact. But displacement is never developed into a barrier in its own right.

It returns only at the very end, in the recommendations, where the authors call for "further socio-economic research into autonomous vehicle deployment… to mitigate negative impacts like potential job losses or urban sprawl and amplify positive ones like improved accessibility." In other words, the workforce question is explicitly designated as unfinished business , a gap for future study rather than a finding of this one.

Interestingly, the labour angle also appears in reverse. In the economic discussion, the authors note that driver shortages could actually help justify automation sooner in specific niches, particularly long-haul trucking, a reminder that automation interacts with the workforce not only by removing jobs but by filling gaps the workforce can no longer supply.

The other half: teaching the public to use the technology

A third strand of the paper's implicit skills story concerns users rather than workers. Across the societal-acceptance section, education and familiarity emerge as the main levers for building trust: pilot programmes that let people ride in autonomous shuttles, transparent communication about safety records, and outreach campaigns to demystify the technology. The barrier lists the authors review echo this, citing "learning and usability challenges for users" and the importance of awareness and education in shaping adoption. The study's own interview guide asks experts directly how consumers can be "educated for using this technology."

This reframes skills as a two-sided challenge: not only the professional competencies needed to build, regulate, and operate AVs, but the broader public literacy needed for society to accept them.

Why the framing matters

Rounding out the picture, the authors' recommendations lean on capabilities that are fundamentally about people and know-how, interdisciplinary research clusters pairing engineers with ethicists and legal scholars, industry sharing technical expertise with lawmakers, and universities acting as neutral hubs for data and public dialogue. Skills, in that sense, run quietly through the paper's proposed solutions even though they never headline its problems.

For a research field increasingly attentive to the human-capital side of automation, that is the useful takeaway. This is a comprehensive, carefully evidenced account of Europe's AV barriers, and it still treats skills and workforce as secondary to regulation, technology, and economics, scattered across categories and largely postponed to future work. If even a thorough barriers study situates skills at the edges, the gap between how central the workforce transition will be and how much attention it currently receives is precisely where the next conversation needs to happen.

A training centre built to rehearse the future

The training is taking place in a purpose-built centre at the Villaverde depot, designed as a hands-on test space that closely replicates real-world operating conditions. Its centrepiece is a replica simulator, identical to the driving desk hidden inside the new automatic trains being built in Navarra. Although these trains run in automatic mode and have no conventional cab, they conceal a small driving console that Metro staff can use if needed, and the simulator lets crews rehearse exactly that.

Around it, the centre is equipped with five pre-training stations, an instructor position, and an observation station for monitoring sessions. The set-up deliberately blends immersive driving environments with railway operation and management scenarios, so that teams can train for supervision, incident handling, and passenger service, not just for the rare moments when manual intervention is required. Roughly 120 workers will pass through, entering the automated environment through realistic scenarios before the system goes live. Meanwhile, the first automatic train for the line has already begun test running in Corella, Navarra.

Three new roles — and a workforce that keeps its jobs

The most instructive detail is what the 117 are being trained for. Automation of the Circular line does not simply remove drivers; it creates three new professional profiles, all filled by existing Metro staff who passed a prior competency-based selection process:

• 27 managers (GELAR) who supervise the service, coordinate personnel, and handle operational incidents, from managing large passenger flows to overseeing critical equipment such as platform doors.

• 25 controllers (CLAR) working from the Central Control Post, responsible for centrally organising operations, attending directly to passengers on the trains, managing disruptions, and coordinating emergency response.

• 65 operators (OLAR) stationed along the line itself, assisting travellers, supporting circulation, and supervising the installations tied to operation and driving when necessary.

Read together, these roles describe a familiar pattern in well-managed automation: the job of "driving" dissolves into a wider set of supervisory, control, and customer-facing functions. The skills required shift from manual operation toward monitoring automated systems, interpreting data, managing exceptions, and looking after passengers, but the people are retained and retrained rather than displaced. Regional transport minister Jorge Rodrigo framed the programme as helping workers "adapt to the new operating model and acquire the knowledge needed to face this technological change," while positioning it as a commitment to innovation, safety, and service excellence.

Part of a much larger transformation

The training is one strand of a complex modernisation of La Circular, a line that carries around 400,000 passengers a day. Beyond the 48 new automatic trains, the project has meant renovating the track superstructure, raising the power supply from 600 to 1,500 volts, installing platform doors, and deploying renewed command-and-control systems. Automation, in other words, arrives bundled with a deep upgrade of physical and digital infrastructure, and the workforce has to be ready for all of it at once.

This article was drafted based on the press notes from the Madrid Metro's website (1, 2).

For anyone thinking about the human side of this transition, the competencies, roles, and skills that cities will need, the Charter is more relevant than its focus on governance might first suggest. Automation does not just change vehicles. It changes what public authorities must know how to do, and it reshapes the workforce that keeps mobility moving.

Why cities and regions must lead

The Charter's starting point is that cities and regions are not bystanders to automation but its natural orchestrators. As managers of public space and local mobility systems, and as procurers and authorisers of publicly operated services, they already set the conditions under which new services succeed or fail. Their responsibilities span safety, the balancing of competing uses of public space, the integration of new services into existing networks, and the maintenance of public trust.

That is why the Charter calls on EU and national decision-makers to guarantee that local authorities are adequately involved in developing and adapting legislation. Crucially, it frames this not only as a question of authority but of capability: deployment must be supported by "the necessary conditions and competencies for cities and regions to manage it effectively." In other words, giving cities a role is meaningless unless they also have the skills and capacity to exercise it.

Ten principles, one throughline

The Charter's ten principles range across public policy direction, liveability and safety, public transport integration, sustainable economic models, governance, infrastructure, societal impact, data access, Europe's public-transport strengths, and the regulation of private AVs and robotaxis. Running through them is a consistent priority: automation should reinforce shared and public transport rather than compete with it.

The Charter is explicit that priority should go to shared use cases, automated feeder services, pooled demand-responsive roboshuttles, and the automation of large regular buses — which can extend coverage and help address operational challenges such as driver shortages. By contrast, privately owned AVs and non-pooled robotaxis risk increasing congestion and total vehicle-kilometres travelled. This is a mobility argument, but it is also a workforce one: the use cases the Charter favours are precisely those that reshape existing transport jobs rather than simply removing them.

The skills dimension: a just transition, and new competencies

Where the Charter speaks most directly to skills is in its principle that societal impacts must be placed at the centre, recognising that automation affects workers, users, and communities across multiple sectors beyond mobility.

Here the Charter asks cities and regions to act as facilitators of social dialogue and managers of transition. It calls for the early and continuous involvement of workers and citizens, transparent communication to build trust, and, most pointedly for a skills agenda, the capacity, resources, and policy frameworks needed to anticipate skills and employment shifts, promote fair working conditions, and support a just, inclusive, and widely accepted transition.

That single sentence carries a substantial to-do list. Anticipating skills shifts means understanding, in advance, which roles will change: professional drivers moving toward supervision, fleet-monitoring, and passenger-service roles; traffic managers working alongside automated systems; procurement teams learning to specify and evaluate AI-enabled services they did not design.

AI competence is now a public-sector competence

The Charter also makes clear that automation cannot develop as a standalone ecosystem bolted onto the city. CCAM services must be integrated into existing transport management, public transport networks, traffic operations, and digital mobility platforms, which requires sustained investment in the digitalisation of infrastructure and mobility management.

This is where the skills question becomes an AI-skills question. To manage automated services safely and fairly, public authorities need to access and interpret operational data from service providers, while addressing data protection, cybersecurity, system integrity, and dependencies on external technology providers. Those are not tasks that can be fully outsourced. They demand in-house fluency in data governance, digital infrastructure, and the workings of AI-driven systems — competencies that many local and regional administrations are only beginning to build.

The Charter reinforces this by grounding deployment in Europe's strengths in integrated public transport, multimodal planning, and public governance, while acknowledging that cities may need to use non-European technologies in the short term to gain experience. Turning that borrowed experience into lasting European capability is fundamentally a matter of skills and knowledge retention inside public institutions.

When the European Parliament's Scientific Foresight Unit published its May 2026 study on the Expected impact of the deployment of Automated Vehicles in the EU, most of the headlines it invited were about regulation, industrial sovereignty, and the race with the United States and China. But threaded through the report is a quieter, arguably more consequential story: what automated vehicles (AVs) will do to European jobs, and which skills the continent needs to protect the value it creates from the technology.

The study never devotes a standalone chapter to the labour market. Instead, its workforce message emerges across its scenario analysis, and the picture it paints is that the employment outcome is not fixed by the technology itself, but decided by the strategic choices Europe makes.

A double-edged economic promise

The report opens by acknowledging both sides of the ledger. Reviewing 49 review articles on AV effects, it notes advantages "on the economic level regarding potential cost reductions and productivity gains," while flagging high capital costs, labour displacement, and regulatory uncertainty as real risks. Its conclusion is deliberately conditional: the sustainability of AVs, it argues, depends largely on deployment strategies. In other words, whether automation lifts or hollows out the European workforce is a policy question, not a technological inevitability.

Three futures, three labour markets

The heart of the study is a set of three scenarios for the next 10–15 years, and each carries a distinct workforce signature.

In the most optimistic future, European Leadership, employment shifts toward highly skilled roles in engineering, AI research, cybersecurity, system integration, and Mobility-as-a-Service. The report is candid that some traditional driving professions become less important, but it frames this as a managed transition: the change is "accompanied by training and retraining programmes," leaving the overall labour-market impact stable or even positive. The key word is accompanied — the good outcome is contingent on active reskilling, not automatic.

The scenario the study's interviewees consider most plausible, Selective Strengths and Dependencies, produces a mixed result. Europe still creates high-skill jobs in engineering, systems integration, and transport, but the most advanced innovation work, large AI models and advanced chip design, stays concentrated outside Europe. Here, Europe captures some of the value chain's skilled employment while ceding its frontier tiers to global partners.

The bleakest future, External Dependence, is where the workforce warning becomes sharpest. High-skill jobs in AI, chip design, and advanced vehicle software decline, while only operational and maintenance-oriented roles grow, and then only moderately. In this world Europe becomes an integrator and operator of foreign systems rather than a creator of them, a labour market skewed toward lower-margin activity, with data sovereignty and vendor dependency emerging as pressing social concerns.

Europe's real skills advantage

If there is a strategic through-line, it is this: the study locates Europe's genuine human-capital strength not in raw AI scale but in a specific engineering culture. Across its expert interviews, Europe is described as strong in safety engineering, verification and validation (V&V), regulatory compliance, and system integration.

Crucially, the report argues these are not soft advantages to be cashed in passively. Regulatory leadership in complex, AI-driven mobility systems, it insists, requires hands-on engineering competence and continuous engagement at the technological frontier. Without that engagement, Europe risks becoming "a rule-maker without sufficient technological grounding" , able to write standards it can no longer credibly assess. One interviewee put the underlying point bluntly: technological development capacity is an essential pillar of strategic autonomy, and regulatory leadership cannot be exercised independently of engineering expertise. Skills, in this framing, are the precondition for sovereignty, not merely a byproduct of it.

Building the talent pipeline

The report's proposed answer to the skills question lives largely in its treatment of research and development. Its favoured model, echoed by most experts consulted, is a hybrid: a limited number of European Centres of Excellence concentrating high-end expertise, supported by a distributed network of research institutions that keeps participation and knowledge exchange broad across Member States.

This is, in effect, a talent-development architecture. Centres of Excellence provide the critical mass and the deep specialists; the distributed network ensures skills and knowledge do not pool in a few regions while others fall behind. The study pairs this with a call for targeted public funding and public–private partnerships to sustain the deployment-oriented research where much practical engineering skill is built, the demanding work of validation, type approval, and certification. One expert usefully qualified the model: the case for Centres of Excellence is strongest at higher technology-readiness levels, while earlier-stage innovation benefits from broader, distributed effort.

The workforce also on the other side of the windscreen

Finally, the study's social lens reaches beyond people who build or operate AVs to those who ride in them. It emphasises improved transport access for elderly people and people with disabilities, and the potential for autonomous shuttles to supplement thin public-transport networks in rural areas, connecting people to shops, hospitals, and jobs. A workforce story is embedded here too: better mobility widens the labour pool that can physically reach employment.

The takeaway

The clearest message the report sends on skills and workforce is that displacement of driving jobs is real but not the whole story, and that the size and quality of the jobs created in its place depend on choices Europe has not yet fully made. Build the engineering, safety, and integration talent, pair automation with retraining, and Europe captures skilled, durable employment. Fail to invest, and it inherits the maintenance work while the high-value roles migrate elsewhere. The vehicles may be automated; the workforce outcome is not.

The transition towards CCAM is expected to reshape job profiles, skills requirements, working conditions, and training needs across the sector. Understanding these changes is essential to ensure that workers, employers, training providers, and policymakers are prepared for a fair and inclusive transition.

We invite Forum members to contribute their perspectives on:

🔹 emerging skills needs and training requirements
🔹 opportunities for new and evolving job profiles
🔹 potential risks of labour market disruption
🔹 barriers to participation in reskilling and upskilling activities
🔹 policy and organisational measures that can support workforce adaptation

Your input will directly contribute to RESKILLING's analysis of future skills needs and the development of recommendations for a socially sustainable deployment of CCAM across Europe.

COMPLETE THE SURVEY HERE: https://docs.google.com/forms/d/e/1FAIpQLSeRiMEq1Td95dAqFkELFeHXZoquJekL3Rf6YhcZzdFmha2DfA/viewform

While CCAM can create new opportunities, it may also present challenges for some workers, particularly those who face barriers to training, digitalisation, or labour market participation.

We are particularly interested in perspectives relating to:

🔹 women working in transport
🔹 older workers
🔹 young people entering the labour market
🔹 workers with low digital skills
🔹 migrant workers
🔹 persons with disabilities
🔹 platform, temporary, or self-employed workers
🔹 workers in rural, remote, or underserved areas
🔹 caregivers and people facing multiple barriers

We encourage workers, worker representatives, trade unions, civil society organisations, employers, training providers, researchers, and public authorities to share their views.

Your feedback will help identify risks, opportunities, reskilling needs, and policy measures that can support an inclusive and equitable transition to automated mobility.

COMPLETE THE SURVEY HERE: https://docs.google.com/forms/d/e/1FAIpQLSeRiMEq1Td95dAqFkELFeHXZoquJekL3Rf6YhcZzdFmha2DfA/viewform

EUCAD symposia are the bi-annual EUCAD events that alternate with the European Conferences on Connected and Automated Driving conferences. The 2026 edition is the continuation of EUCAD 2018, EUCAD 2020 and EUCAD 2024 symposia.

The EUCAD 2026 Symposium is organised by the CCAMbassador project and supported by the European Commission Directorate General for Research & Innovation (RTD) as well as the CCAM Partnership.

The event aims to delve deeper into specific key topics and challenges for CCAM Research & Innovation and deployment, in Europe and beyond, namely this year: regulations, AI, digital infrastructure and CCAM in cities. It is targeted at public and private stakeholders interested in exchanging knowledge and views on the latest progress and future actions required to accelerate transition from innovation to implementation.

Participants will have the opportunity to learn more about European Research & Innovation activities on CCAM, engage with experts and stakeholders from the EU, international, and Swiss CCAM ecosystem, and experience an automated public transport service in operation in Geneva, as well as other connected vehicle demonstrators.

The Symposium is targeted, though not limited to, European and international stakeholders from the industry and research sectors, policy makers, regulators, representatives from the European institutions, national and local public authorities, road authorities / operators, transport operators and users.

Participation in the event is free of charge, but registration is compulsory.

More information on the role of our project in the Conference and the posibilities for the stakeholders to attend will be soon provided.

For further enquiries, please contact eucad@connectedautomateddriving.eu