Enabling Advancing and Possibilities

Although the possibilities of the promising technologies described herein appear endless, enabling progress in the near term requires selecting and prioritizing public-sector initiatives and investments. Federal, state, and local transportation agencies, along with their partners in universities and private industry, will face a multitude of decisions about where and when to advance or invest in these technologies and what regulatory precautions to take. Whereas some proponents assert that these technologies will lead to immediate and radical transformations in human society, others argue that they are “normal” technologies. If the latter is the case, the diffusion of these technologies is likely to occur over a time scale of decades, with attendant uncertainty about the scale and timing of their impact on transportation.¹⁰⁹ Although it is beyond the scope of this publication to weigh in on specific investment priorities, this concluding section offers observations about ways to advance promising candidates.

Transportation agencies have already begun the work of developing initiatives and strategies focused on AI, digitalization, and advanced automation and autonomy. The U.S. Department of Transportation (USDOT) released its AI strategy in October 2025, which presents use cases and its maturity assessment.¹¹⁰ State DOTs, transit agencies, and other transportation agencies have begun to do likewise; Box 3 summarizes a recent effort to produce an AI research roadmap for state and local DOTs.¹¹¹ Significant work still needs to be done. A 2025 AASHTO survey revealed that a majority of its members did not think

that the transportation operations sector was even somewhat prepared to deploy AI systems responsibly at scale.¹¹² Survey respondents also identified the biggest challenge (76%) to be data quality and availability.¹¹³

Data Needs, Digitalization, and AI

Digitalization of the elements of the transportation system is a necessary step, including establishing data standards as appropriate. Training machine learning models requires enormous amounts of data; however, once a model is trained, it can be evaluated for how well it generalizes to a similar domain with little or no retraining or fine-tuning. The transformative power of these technologies accelerates when formerly isolated knowledge bases and data sources can be integrated, requiring data that can be accessed by numerous applications and platforms. Providently, these technologies can be used to automate the process of creating high-quality digital representations of transportation systems. For example, an accurate streetlight inventory is critical for road safety research and interventions. Image analytics on a database of geolocated street images can be used to cost-effectively build such an inventory by identifying their presence and, just as importantly, their absence.¹¹⁴

Transportation agencies have been working on digitizing and improving the quality of their data for decades. Major efforts run the gamut from

roadway elements¹¹⁵ to work zones¹¹⁶ to traffic crash reports.¹¹⁷ State DOT–sponsored research is under way that tackles knowledge management for this new age of AI.¹¹⁸ While advanced automation and digital infrastructure offer the potential to produce new sources of data to feed AI applications, data collection and digitalization efforts can still be expensive, and clarity about their benefits and cost-effectiveness may only be achieved in the intermediate or longer term.¹¹⁹

Because the private sector will continue to be a source of innovation for these technologies, it is important that private entities have ready access to data, when and where appropriate. Data standards developed by transportation agencies can also help maintain a competitive environment among existing and new technology vendors. The development of the General Transit Feed Specification (GTFS) in the 2000s is an example of the transformative potential of data standards. GTFS—the “G” originally stood for Google—smoothed the way for today’s ubiquitous trip-planning and real-time arrival apps that make it easier for transit riders to navigate familiar and new destinations. GTFS is maintained by the nonprofit organization MobilityData.¹²⁰

In addition, data collection, storage, maintenance, and sharing will continue to require special care to address privacy, liability, surveillance, security, and cybersecurity concerns. Automated systems that integrate image, text, sound, and real-time data and AI-enabled analysis of such data also raise concerns about privacy, bias, and liability. Additional issues, such as the use of data beyond its original use case, will present themselves. In addition, autonomous vehicles, aircraft, and aerial drones raise complex, unresolved security¹²¹ and privacy issues¹²² that require further research, analysis, and policy resolution. New policies, procedures, and regulations may well be required.¹²³

Supporting and Spurring the Innovation Process

Although transportation is increasingly benefiting from a plethora of innovations that leverage AI, digitalization, and advanced automation, the power and potential of these technologies has scarcely been tapped. TRB has prepared this publication both to inspire transportation practitioners to envisage the possibilities and to encourage the research community to help realize them.

To be sure, entrepreneurial forces will continue to drive the innovation process, but at a pace that may be constrained by the mixed public- and private-sector nature of the transportation enterprise. Reducing these constraints will require concerted efforts to support and spur innovation, such as through priority setting, regulatory flexibility, coordination across federal and state transportation R&D programs, standards development, and funding commitments. The rapid pace of technological development will also require concerted efforts to engage agency decision makers, who may need to be made aware of the potential uses and benefits of emerging innovations and who may operate in a risk-averse and budget-constrained environment that can complicate procurement.

Although most of the innovations in publicly owned and operated infrastructure will be delivered by state and local agencies, broader supportive and collaborative efforts will be needed. With USDOT assuming a national leadership role, TRB can play its traditional role of bringing together the many private- and public-sector parties

engaged in transportation R&D—a role that can be instrumental in furthering innovation through the sharing of understanding and experiences. Advanced tools and exploratory applications that are within the domains of TRB’s sponsors will require further R&D before they can be introduced, a process that is already under way. As detailed above, TRB’s cooperative research programs, federally funded University Transportation Centers, and the research offices of USDOT and its modal agencies are pursuing such research, but within a dispersed environment that can hinder strategic coordination and integration to accelerate the innovation process.

Given the serendipitous and unpredictable nature of innovation, there is no way to predict all the innovations that will emerge from AI, digitalization, and advanced automation to the benefit of transportation, much less to reliably predict those that will be transformative. However, it is possible to identify barriers to innovation and to take actions to overcome them—and thus, to help expand and accelerate the pace of innovation. While R&D underpins innovation, TRB has prepared this publication with recognition of the importance of creating and sustaining a background environment that encourages and values innovation. Research can help inform the development of such an environment conducive to innovations and their use.

The following is an example set of topics that matter for capitalizing on the possibilities of AI, digitization, and advanced automation. Some present fertile ground for research, including policy research, while others are likely to require policy initiative, strategic coordination, and cultural changes so that the research and innovation processes can bear fruit.

1. Regulation. Current motor vehicle and road safety regulations were written largely with human-operated vehicles in mind. Advances in autonomy and their deployment are being constrained by such well-intentioned regulations. Allowing innovation that could enhance safety requires alternate approaches. The safety case mentioned above may be one. A comprehensive review of regulatory constraints could identify other options for preserving and enhancing safety. New regulatory approaches may also have to take into consideration liability for failures.
2. Knowledge and skill gaps. The staff of government agencies at all levels who would need to evaluate and implement potential innovations will require increasing amounts of guidance and training on the benefits, performance, and limits of the technologies described above. Agency workforces may well need bolstering with individuals who have the expertise needed. Many agencies are struggling to compete with private industry in attracting and retaining employees with the skills and abilities needed and need better strategies for doing so.
3. Digitalization data standards and data sharing. As noted above, AI requires data for training and the digitalization process itself requires greater data standardization to allow for deployment at scale. Even though state transportation agencies have been working on this issue for many years, much work remains to be done. Greater federal leadership and support could accelerate progress.
4. Disruption risks. Digitalization introduces risks of disruption from natural disasters and malevolent actors, as well as from service outages (e.g., cloud and web) and technology

support phase-outs (deprecation). These risks will need to be effectively mitigated to keep systems operating and to protect privacy and sensitive information. Further R&D to enhance cybersecurity is a clear priority.

5. Agency and public trust. There will be natural risk aversion and reticence to allow unfamiliar innovations to be introduced, particularly those based on generative AI using reasoning processes that appear opaque, without some form of independent verification and validation. Universities and other independent organizations could serve this function, providing verification, validation, and transparency.
6. Building an evidentiary base for guidance. Many decisions about infrastructure design, operations, and safety are driven by formal guidance documents and decision tools. While these resources are evidence based to the extent possible, many also rely on professional judgment and rules of thumb when data and research are lacking. The technologies described above provide opportunities to accelerate data gathering and research to develop sound evidentiary bases, but the opportunities will need to be exploited by parties responsible for updating guidance and decision tools.
7. Monitoring and communicating progress. The implementation of innovations depends on potential users being made aware of deployment opportunities and the potential benefits. Awareness and knowledge are gained through sharing and learning processes.¹²⁴ However, traditional processes may be overwhelmed by the expanse and rapid pace of technological development. AI may provide ways to track and categorize innovations to keep communities of interest better informed and able to share experiences.

As is often the case, marked progress in addressing many of the topics identified above to support and spur innovation will require a commitment of funding, particularly by agency implementers. Public agency budgets are finite and can provide little room for investments to obtain, operate, and maintain innovative technologies. For example, digital infrastructure requires not only capital investment but also ongoing spending on operations and maintenance. Likewise, applications that use AI can require large annual expenses for data storage, computing power, and software licenses. When additional public-sector funding for technologies is not forthcoming, transportation agencies may need to be encouraged and given more latitude to partner with the private sector to test and implement innovations, perhaps by building off the precedents of existing public–private partnerships that have been used successfully for the supply, operation, and maintenance of transportation infrastructure.