Written by: Glenn Havinoviski, Smart Mobility Market Leader for JMT Technology Group

While vehicle automation a decade ago was predicted to become a dominant if not universal force in transportation as soon as 2025, leading to unimagined levels of mobility and safety, the ensuing years have been a lesson in realism. Various automated shuttle, robotaxi and automated trucking services went through various demonstrations that frequently ended with a whimper (e.g., the automated shuttles at National Harbor in Maryland and Mosaic District in Virginia), deployment of what were essentially raw beta versions of some systems for public consumption (Cruise robotaxis), and finally, various business failures and consolidations. (Anyone remember ARGO AI, Local Motors, or Ghost Autonomy?) 

But the recently-concluded 2026 Automated Transportation Symposium in San Diego (the second year in which SAE has organized this event) has established that the autonomous vehicle (AV) is increasingly becoming a reality. From the sessions I attended, and conversations I had with the various policy, technology and engineering experts at this conference, below are four indicators that point to the way forward for vehicle automation and also five areas of caution and concern that could still limit its development. Last but not least, we will address how all this provides opportunities for JMT.

Four Indicators of the Way Forward

1. Robotaxis are maturing into a real mobility service. Waymo has moved beyond limited demonstrations into revenue service across eleven U.S. metros, with continued expansion planned in the Northeast and potential future service in the Washington, DC, region. Initial mapping is underway in markets such as Philadelphia and New York, with similar activity proposed for Baltimore.
   • Deployments still encounter visible edge cases—such as emergency-response conflicts, school-bus interactions, light-rail rights-of-way, and vehicles stopping in awkward locations—but the overall trajectory is improving. Recent safety studies have shown encouraging crash-rate comparisons, aided by conservative operating behavior, and investor interest in automated mobility has begun to return.
2. Automated Trucking Is Emerging as a Viable Business. Automated trucking is becoming more practical where routes, customers, and operating conditions are clearly defined. Early use cases emphasize repeatable corridors, logistics hubs, ports, industrial sites, and middle-mile operations rather than universal driverless trucking.
   • Texas has become a proving ground because it combines long freight corridors, logistics hubs, favorable conditions, and a permissive deployment environment. The Atlas Energy Solutions/Kodiak AI frac-sand operation in the Permian Basin illustrates the near-term business case: high-volume, repeatable industrial freight movements in which driver availability, utilization, safety exposure, and cost are persistent concerns. California remains important for technology development, capital, testing, safety cases, and OEM partnerships, even though fully driverless heavy-truck deployment is more constrained there. Across the U.S., firms such as Aurora, Kodiak, Gatik, Waabi, PlusAI, and Torc are pursuing targeted freight applications where routes, facilities, and support needs can be managed.
3. The Federal Government is Becoming More Active in Supporting Safe AV Operations. NHTSA Administrator Jonathan Morrison’s Thursday keynote at the ATS underscored a changing federal role: the agency is moving beyond broad guidance and case-by-case caution toward a more active framework that both enables deployment and tightens oversight. The most visible example was NHTSA’s approval of a temporary commercial exemption allowing Zoox to deploy purpose-built robotaxis in Las Vegas without traditional manual controls (e.g., steering wheel, hand, or foot controls), subject to conditions and enhanced reporting. At the same time, NHTSA announced it would work with the SAE Industry Technologies Consortia on A2SCEND, a new effort intended to gather data and accelerate the development of national AV performance standards. The message was that the federal role is shifting toward scalable safety governance: clearer exemption pathways, more structured monitoring, updated guidance on areas such as emergency responder interaction and remote assistance, and an eventual national framework to reduce the current patchwork of state-by-state rules.
4. Increased Public Acceptance. Public exposure to automation is growing through three channels: Automated Driver Assistance Systems (ADAS) in personal and fleet vehicles, expanding robotaxi services, and smaller automated shuttle or microtransit pilots. Acceptance varies by application, with ADAS becoming mainstream, robotaxis gaining adoption in select markets, and fixed-route circulators and shuttles seeing more limited adoption.

• • ADAS: For most consumers, automation is first experienced through Level 1 and 2 driver-assistance features such as automatic emergency braking (AEB), pedestrian detection, lane-keeping assistance, blind-spot warning, adaptive cruise control, and lane centering. These systems do not make a vehicle autonomous, but they normalize automated support for perception, braking, and steering. ADAS has also moved from optional technology toward a safety baseline. By 2023, joint NHTSA/automaker studies found that at least 10 ADAS features exceeded 50 percent market penetration, with five—including forward collision warning, automatic emergency braking, pedestrian warning/AEB, and lane departure warning—reaching 91 to 94 percent. NHTSA will require AEB, pedestrian AEB, and forward collision warning on new light vehicles by September 1, 2029.
  • Ride-Hailing Services: Robotaxis are being used where available, but trust remains uneven. Waymo reported 500,000 paid weekly rides across 10 U.S. cities in March 2026, up from 50,000 in May 2024, while AAA’s 2025 survey found that only 13 percent of U.S. drivers would trust a self-driving ride and 53 percent would not choose a robotaxi. For agencies, acceptance will depend on reliability, coverage, safety performance, emergency-response coordination, labor concerns, and community benefit.

Five Areas of Caution and Concern

1. Connectivity. Several ATS sessions addressed challenges in AV interactions with first responders, weather, and work zones. AVs rely heavily on onboard perception, maps, remote monitoring, and fleet operations, while public agencies manage real-time conditions through TMCs, incident data, work-zone information, emergency response, and third-party feeds. The challenge is turning that information into trusted, standardized, machine-readable data AVs can safely use. Key gaps include latency, reliability, cybersecurity, privacy, liability, and responses to wrong or missing data. First responders add operational complexity, as AVs must respond to emergency vehicles, police direction, crash scenes, fire hoses, blocked lanes, and detours that may not appear in digital feeds. V2X and cooperative automation can help by sharing signal timing, work-zone status, lane changes, emergency-vehicle presence, vehicle intent, and planned maneuvers; SAE J3216, AVSC guidance, USDOT’s V2X plan, and FHWA’s CARMA program support this direction. Still, AVs need safe fallback behavior when connectivity fails, and agencies must address governance, cybersecurity, procurement, maintenance, emergency procedures, and traffic-operations roles.
2. Sharply Different Levels of AV Deployment Across the U.S. As was discussed in several sessions at ATS, each State has taken different approaches to overseeing or regulating AV deployment activities. For example, California is trying to manage risk through formal state regulation; Texas is trying to enable deployment while shaping oversight through coordination and industry engagement; and states such as Arizona, Florida, and Georgia generally allow broader deployment with more localized operational control. Maryland and Virginia, on the other hand, represent the majority of states with limited or no legislation on the deployment of driverless vehicles. In fact, states such as New York have allowed AV testing only with safety drivers and under stringent permitting requirements. Should the District of Columbia approve the deployment of robotaxis, it may well force the hand of the Maryland and Virginia legislatures. The result is a patchwork national environment in which AV companies choose states not only based on markets and road conditions, but also on regulatory predictability, reporting obligations, and the ability to scale operations.
3. Workforce. Workforce issues in vehicle automation extend beyond potential driver displacement. They include how automation changes job tasks, skill requirements, staffing models, labor agreements, training needs, safety responsibilities, and emergency-response protocols. Impacts will differ by use case: ADAS may change how drivers supervise vehicles; robotaxis may reduce in-vehicle driving roles while increasing fleet operations and remote assistance needs; and automated trucking may shift some work from long-haul driving to local delivery, terminal operations, maintenance, dispatch, and oversight. At the same time, automation creates new workforce demands in remote operations, fleet management, sensor calibration, software support, cybersecurity, data analysis, incident response, V2X/traffic systems operations, and AV maintenance. Public agencies should therefore treat AV deployment as a workforce transition issue, requiring impact assessments, labor engagement, vendor workforce plans, retraining partnerships, updated job classifications, emergency-response training, and procurement requirements for staffing, remote operations, incident response, and local workforce participation.
4. Safety and Security. AV safety is improving, but agencies and the public still need confidence that systems can handle edge cases within their operational design domains, including emergency scenes, construction zones, unusual pedestrian behavior, weather, degraded sensors, and confusing roadway geometry. Safety evaluation is therefore shifting toward safety cases, performance metrics, event reporting, remote assistance, post-crash behavior, and continuous monitoring. Security risks are equally important because AVs depend on sensors, maps, software, V2X, cloud systems, fleet operations, remote assistance, and over-the-air updates. Key vulnerabilities include spoofing, jamming, GPS manipulation, false messages, software supply-chain attacks, unauthorized access, privacy risks, and denial-of-service attacks. The emphasis should be on resilience: layered cybersecurity, authentication, secure updates, redundancy, graceful degradation, safe fallback behavior, incident response, and recovery without unsafe roadway conditions.
5. Does Vehicle Automation Actually Help Mobility?  Automation, alongside shared rides, electrification, and connectivity, was long considered a key element of future mobility. However, whether it is an essential part of mobility is often questioned. Furthermore, several people in various ATS sessions acknowledged the challenges in much of the U.S. regarding willingness to use shared-ride services, let alone ride a bus.

• • Traffic Impacts of Single-Ride Robotaxis: Single-ride robotaxis may improve personal mobility without improving system mobility. Empty repositioning trips, depot travel for charging or cleaning, curb waiting, pickups/drop-offs, and fleet balancing can add vehicle miles and curb conflicts. Agencies should manage robotaxis as a street and curb-use system, not just as a private mobility app.
  • Ridesharing: Automated ridepooling may offer a stronger mobility case than single-occupant robotaxis if it raises vehicle occupancy and complements transit. MOIA’s Hamburg ridepooling service and ALIKE autonomous shuttle project illustrate the model: app-based booking, shared rides, virtual stops, fleet management, and public-private coordination. The key is using automation to support shared, networked mobility rather than simply replacing a human driver in a private ride. 
  • Shuttles / Microtransit: Automated shuttles have seen more limited acceptance than ADAS or robotaxis because most operate as low-speed, geofenced circulators with small vehicles, fixed routes, limited hours, onboard attendants, and pilot-oriented service designs. Florida shows the range: Lake Nona’s Move Nona service has been one of the longer-running examples, while Jacksonville’s Ultimate Urban Circulator/NAVI has faced low-ridership concerns, including reports of roughly 76 to 100 riders per day for a costly downtown deployment.
  • Other pilots show the same challenge. Fairfax County’s Relay shuttle in the Mosaic District ended after a two-year pilot and 356 reported trips; Maryland’s Olli demonstration at National Harbor remained a proof of concept; and Local Motors later shut down. Microtransit has stronger prospects when it solves specific access problems, but agencies should require clear trip purposes, realistic ridership forecasts, useful hours, transit integration, and a path from pilot to sustainable operation.
  • Fixed-Route Transit Services. Automation has seen limited application for buses beyond use of ADAS technologies as well as legacy applications such as precision docking at dedicated stations along bus rapid transit corridors. During the poster session at ATS, I spoke with some transit-focused staff from USDOT’s Volpe Center, about a prior planned demonstration programmed for Level 4 operations along an exclusive busway in the Hartford, CT region. The concept was to retrofit a standard transit bus with vehicle automation capabilities. They mentioned that the automation provider (based in Maryland) was at odds with the operator over liability in the event of incidents and withdrew from the project.