KING TECHNOLOGYSolutions Group

Research · Industries

Driverless vehicles and the trust problem

Published October 2026 · Sources checked October 18, 2026

1. The question about the future

By around 2030, what will it take for the public to feel safe sharing roads with, and riding in, driverless vehicles, and what role will continuous observability (the ability to see, record, explain and act on how vehicles and fleets behave) play in earning that trust?

Trust in a driverless vehicle is not only about the vehicle. It is about whether someone can show, after the fact and in real time, that the system behaved as intended, that problems are noticed and handled, and that the people responsible are accountable. That is an observability problem as much as an engineering one.

A note on scope: this piece describes deployments neutrally from public reporting and makes no claims about any company's safety record.

2. The signals today

2.1 Public opinion: awareness is high, trust is low

AAA's annual survey on automated vehicles, released February 26, 2025 (fieldwork January 9-13, 2025), found that 13% of U.S. drivers would trust riding in a self-driving vehicle, up from 9% the year before, while about six in ten said they were afraid to. Of drivers, 74% were aware of robotaxis, but 53% said they would not choose to ride in one. 78% said improving vehicle safety systems mattered more to them than developing self-driving vehicles. The survey used a probability-based panel; 888 of 1,095 completed interviews qualified, with a margin of error of 4.4 percentage points. [S1][S2] AAA has run the survey annually since 2016. [S2]

The gap between awareness and willingness is the trust problem in one statistic.

2.2 Shared vocabulary: what "driverless" means

SAE J3016 defines six levels of driving automation. At Levels 0-2 a human remains responsible for some or all of the driving task; at Levels 3-5 the automated driving system (ADS) performs the entire driving task while engaged. Level 4 performs it within a defined operational design domain (ODD) with no expectation that a person will take over; Level 5 is not limited to an ODD. [S3] Texas law uses this taxonomy: its program defines an automated vehicle as one with an ADS capable of Level 4 or 5 under SAE J3016 (April 2021). [S4]

2.3 Federal signals

  • Crash reporting (Standing General Order). NHTSA requires named manufacturers and operators to report certain crashes involving ADS and Level 2 driver-assistance systems. The third amendment, issued April 24, 2025 and effective June 16, 2025, requires the most severe crashes to be reported within five days and less severe ones monthly, adds a property-damage threshold for less severe ADS crashes, and removes empty monthly reports. NHTSA publishes the reports as downloadable data (current files cover June 16, 2025 through September 15, 2026) and cautions that the data is not normalized for exposure such as miles traveled, so comparisons between companies are unreliable. [S5][S6]
  • AV Framework and exemptions. On April 24, 2025, the U.S. Department of Transportation announced NHTSA's AV Framework, including the amended reporting order and an expansion of the research exemption program to domestically built, non-commercial vehicles. [S6]
  • Rulemaking for vehicles without manual controls. On June 25, 2026, NHTSA proposed amending brake-system requirements (FMVSS 135) for ADS vehicles without manual controls. The comment period, first set to close July 27, 2026, was extended to August 26, 2026; no final rule had been issued at the time of writing. On the same day as the proposal it withdrew its January 2025 proposal for the voluntary AV STEP transparency program. The 2026 Unified Agenda lists further planned updates to standards for vehicles without traditional controls and an advance notice on ADS performance assessments. [S7]
  • First responders. On July 8, 2026, NHTSA's Administrator wrote to ADS developers about what the agency called a "clear pattern" of driverless vehicles entering active emergency scenes, blocking ambulances and fire apparatus, and failing to recognize cues such as flashing lights, flares and cones. The letter described the inability to detect and respond appropriately as a "functional insufficiency," and NHTSA said it would schedule meetings with developers by the end of that month; we found no public account of their outcome. The letter did not name companies or specific incidents. [S7][S8]

2.4 Safety and cybersecurity standards

  • ISO 26262 (functional safety). Addresses hazards caused by malfunctioning behavior of safety-related electrical and electronic systems in series-production road vehicles. The 2018 edition is marked for revision. [S9]
  • ISO 21448 (safety of the intended functionality, "SOTIF"). Published June 2022, it addresses "the absence of unreasonable risk due to a hazard caused by functional insufficiencies," including in sensors and algorithms, and covers reasonably foreseeable misuse and remote operation or back-office communication that can affect vehicle decisions. It is also under revision. [S10] NHTSA's July 2026 letter used the same "functional insufficiency" language. [S8]
  • UL 4600 (safety evaluation of autonomous products). Issued by UL Standards & Engagement, it takes a safety-case approach: a structured argument, backed by evidence, that a system is safe for a specific use and environment, maintained through the lifecycle. The first edition was published in 2020 and the third, which added autonomous trucking, in March 2023. [S11]
  • ISO/SAE 21434 (cybersecurity engineering). Published August 31, 2021, it covers cybersecurity risk management across the vehicle lifecycle, including operation, maintenance and decommissioning. It is marked for revision following a review that closed October 2, 2026. [S12]
  • UN Regulations 155 and 156. UN R155 requires vehicle manufacturers to operate a cybersecurity management system with processes "to monitor for, detect and respond to cyber-attacks, cyber threats and vulnerabilities," with monitoring continuing after first registration, vehicle-level "data forensic capability," and monitoring results reported to the approval authority at least once a year. [S13] UN R156 covers software-update management systems. Its UK type-approval authority lists application dates for different vehicle categories running from 2022 through 2029. [S14] These regulations apply in markets that adopt them, not in the United States, but they set a widely referenced benchmark.

2.5 Texas: law, rules and deployments

  • Statute. Senate Bill 2807 (89th Legislature, 2025), codified in Texas Transportation Code Chapter 545, Subchapter J, took effect September 1, 2025. Rules took effect February 27, 2026, and authorization became enforceable May 28, 2026. Companies operating driverless vehicles commercially must hold an authorization from the Texas Department of Motor Vehicles (TxDMV). [S4]
  • What operators attest to. Applicants acknowledge that each vehicle can comply with traffic laws, is equipped with a required recording device, has an ADS that complies with federal law, can reach a minimal risk condition if the ADS fails, and is properly registered and insured. They must also certify that the Department of Public Safety (DPS) has received an emergency-interaction plan describing how police, fire, medical and other emergency personnel should interact with the vehicle. [S4]
  • Oversight powers. TxDMV may suspend, revoke or restrict an authorization if operation has resulted, or is likely to result, in serious bodily injury, through a defined notice, correction and review process. [S4] TxDMV launched a public complaint process and an authorization lookup on June 2, 2026. [S15] Transportation Code §545.452 bars cities and other political subdivisions from imposing "a franchise or other regulation related to the operation of an automated motor vehicle," which is why, as reporting on a City of Austin update notes, cities cannot run their own AV permitting systems. [S4][S16]
  • Deployments (described neutrally). Public reporting describes Austin as an early testing ground for several robotaxi programs. [S16] Invitation-based driverless ride-hail service began in Dallas, Houston and San Antonio on February 24, 2026, and has since opened to the general public: in Dallas on August 4, Houston on August 21 and San Antonio on October 14, 2026. [S17] In April 2025, commercial driverless freight operations were announced on the roughly 200-mile I-45 corridor between Dallas and Houston. [S18]
  • City-level observability. The City of Austin launched an interactive dashboard in 2023 that logs AV safety concerns, near misses and collisions; as of July 2026 staff had documented more than 300 incidents. An interdepartmental working group, including public safety, emergency response, schools and the airport, flags problems to operators and follows up until fixes hold. [S19]
  • Research. Researchers at the University of Texas at Austin's Center for Autonomy have urged regulators to treat AV safety as something continuously maintained rather than certified once, calling for "adaptive, continuously updated standards"; news coverage reports proposals for staged rollouts, continuous monitoring and runtime safety checks. [S16][S20]

2.6 What the signals add up to: why observability is central to trust

Across the signals above, every trust mechanism depends on evidence that is captured continuously and can be explained later. The mapping below is our synthesis.

What the public needs to believeObservability capabilityWhere the signal comes from
"If something goes wrong, someone will know quickly."Fleet telemetry, incident detection, crash-report workflows with defined deadlinesFive-day and monthly federal reporting [S5]
"There is a record of what happened."On-vehicle data recording, retained and retrievableTexas recording-device attestation [S4]; UN R155 forensic capability [S13]
"The system did what it was designed to do."Runtime monitors, ODD-boundary tracking, evidence feeding a living safety caseUL 4600 safety case [S11]; SOTIF functional insufficiencies [S10]; UT Austin research [S16]
"A person is watching when the vehicle is unsure."Remote-operations monitoring: queue times, connectivity, handoff outcomesSOTIF includes remote operation and back-office communication [S10]
"It will get out of the way of an ambulance."Detection of emergency scenes, alerts to remote operators, first-responder contact channelsNHTSA letter [S8]; Texas emergency-interaction plans [S4]
"It can't be hacked without someone noticing."Vehicle and back-end security monitoring, detection and response, annual reportingUN R155 [S13]; ISO/SAE 21434 [S12]
"Updates don't quietly change how it drives."Software and configuration version tracking per vehicle, linked to behavior changesUN R156 [S14]
"My city can see what's happening on its streets."Municipal incident logging and data exchange with operatorsAustin dashboard [S19]

The common thread is that safety is demonstrated with evidence, continuously, rather than claimed once. That is what enterprise observability practice already does for critical IT services: define what "healthy" means, instrument for it, detect departures, respond, and keep the record.

3. Plausible outcomes to about 2030

These are scenarios, not predictions.

Scenario A: Evidence-led acceptance. Public trust rises gradually as operators and regulators publish more comparable, exposure-normalized data, safety cases are maintained continuously, and first-responder interaction improves measurably.

Signals that would tip toward A: federal ADS performance-assessment rulemaking progressing [S7]; crash data published with exposure context, addressing NHTSA's current caveat [S5]; AAA trust figures continuing upward from 13% [S1]; measurable progress after NHTSA's first-responder meetings. [S8]

Scenario B: Trust stalls; incidents drive the rules. Deployment continues, but public confidence stays flat. Rules tighten after individual events, with state and local tensions over preemption.

Signals that would tip toward B: first-responder conflicts persisting after NHTSA's call to action [S8]; use of Texas suspension or complaint powers in high-profile cases [S4][S15]; rising municipal incident counts without matching resolution [S19]; survey fear levels unchanged. [S1]

Scenario C: Corridors and zones first. Freight corridors and well-defined service areas expand steadily, where operating conditions are bounded and monitoring is easier, while general public comfort with riding lags.

Signals that would tip toward C: growth in corridor freight operations [S18]; ODD expansions announced area by area rather than citywide; continued gap between robotaxi awareness and willingness to ride. [S1]

Elements of all three may appear at once in different Texas cities and corridors.

4. What to do now

Each step below helps in all three scenarios.

For cities and transit agencies

  • Stand up an AV incident log if you don't have one. Austin's approach (a dashboard of concerns, near misses and collisions, plus a cross-department working group that follows issues to resolution) is a public, working model. [S19]
  • Get the emergency-interaction plans. Texas operators must file these with DPS. Make sure local police, fire and EMS have them, have trained on them and know the contact path for each operator. [S4]
  • Agree on data-exchange expectations with operators. Since cities in Texas can't run their own permitting [S16], work through cooperation: incident notification paths, contact points and timelines.
  • Use TxDMV's tools. Brief staff on the authorization lookup and complaint process. [S15]
  • Plan for transit integration with evidence in mind. If AV services connect to transit, define the operational data, availability and incident information you'll need before contracts are signed.

For fleet operators

  • Treat the safety case as a live system. Link runtime telemetry, incident records and software versions to the claims in your safety case, so evidence updates as the fleet operates. [S11][S16]
  • Instrument the boundaries. Monitor ODD exits, minimal-risk-condition events, remote-assistance requests and response times as first-class signals. [S4][S10]
  • Build reporting into the workflow. Map federal crash-reporting criteria and deadlines to automated triggers, so reports are complete and on time. [S5]
  • Monitor emergency-scene interactions explicitly. Given NHTSA's July 2026 letter, track encounters with emergency vehicles and scenes as a distinct, reviewed category. [S8]
  • Run security monitoring that would satisfy R155 even where it isn't required. Continuous detection and response, forensic data and periodic reporting are a strong benchmark. [S13]

For their IT and operations teams

  • Unify vehicle, fleet and back-end telemetry. Consistent identifiers for vehicle, software version, location zone and trip let incident responders and investigators join data quickly.
  • Design data retention and integrity in. Records that may be needed for regulators, insurers or courts need tamper-evident storage and clear retention rules. [S4][S13]
  • Monitor the remote-operations center like a critical service. Connectivity, staffing, queue depth and handoff outcomes should have defined service levels and alerting. [S10]
  • Track every software and configuration change per vehicle. Link updates to behavior changes, in line with software-update management expectations. [S14]
  • Practice joint incident response. Run exercises with first responders and city staff using realistic scenarios drawn from public reports. [S8][S19]

For CIOs and CFOs (agencies and operators)

  • Budget for evidence, not just vehicles. Data recording, monitoring, retention and reporting are ongoing costs that underpin authorization and public acceptance. [S4][S5]
  • Choose open, exportable data formats so evidence can be shared with regulators, cities and auditors without re-engineering.

Drafted with AI assistance; sources checked and interpreted by King Tech.

Planning for this in your organization?

King Tech collaborates on custom solution architecture: we work alongside your team to design an approach that fits your systems, your constraints and your timeline. Contact King Tech to start the conversation.

Book a consultation Ask a question

Sources