Research report: Pre-implementation public attitudes toward the Czech legal framework for SAE level 3 automated driving

When the Czech Republic’s new legal framework for SAE Level 3 automated driving entered into force on 1 January 2026, most of the population had apparently never heard of it. A national survey of 1,003 Czech adults, conducted between 28 August and 10 September 2025, found that only 9.7 percent of respondents had previously encountered information about the forthcoming rules, 10.8 percent were unsure, and 79.6 percent reported no prior exposure at all. The survey, carried out by the external research agency STEM/MARK using computer-assisted web interviewing, was analysed by researchers at the Transport Research Centre and published in European Transport Research Review alongside a detailed legal analysis of the new legislation.

The law in question, Act No. 130/2025 Coll., amended two cornerstone statutes: Act No. 361/2000 Coll. on Road Traffic and Act No. 56/2001 Coll. on Conditions of Operation of Vehicles on Roads. Together, these targeted amendments establish national operational rules for using approved SAE Level 3 automated driving functions on public roads. Unlike broader regulatory architectures adopted elsewhere in Europe, the Czech approach relies on surgical changes to existing road-traffic and vehicle-registration legislation rather than a separate regulatory code, and it deliberately leaves civil, criminal, product-liability and insurance consequences to other legal regimes.

The legal core of the framework is a redefinition of who counts as a driver. Under the amended Section 2(d) of the Road Traffic Act, a driver is also a person sitting in the driver’s seat of an automated vehicle, even when the machine is doing the driving. The person in the seat remains a legally relevant fallback actor throughout SAE Level 3 operation, but the scope of their duties depends on whether the automated function is active and whether the conditions of the new Section 79b are satisfied.

Section 79b creates a functional, task-specific exemption. While an automated vehicle is being driven by the vehicle itself, road rules and driving duties whose compliance is ensured by the automated system no longer apply to the human driver. Duties unrelated to the system’s performance, including licensing, sobriety, seat-belt use, vehicle fitness and the safety of passengers and cargo, remain firmly with the person in the seat. The provision reallocates the performance of defined driving tasks during active automated mode, but it does not itself decide who bears civil, criminal, product or insurance liability after an incident.

The fallback role comes with teeth. Under Section 79b(2), the driver must be ready to take over control and, upon the vehicle’s request, must do so immediately; failing an immediate takeover ends the exemption. Notably, the statutory wording requires an immediate response but does not itself set a numerical takeover interval, which instead depends on the approved automated driving function and applicable technical requirements. A further provision obliges the driver to let police or military police officers verify whether the vehicle was actually driving itself at a given moment; if verification is refused, it is presumed that the automated mode was not active. The amendment also gives the vehicle operator a specific defence against administrative offences: under Section 125f(7), the operator is not liable if they prove that compliance with the breached duty was ensured by the automated vehicle. Finally, the road vehicle register must now record that a vehicle is an automated vehicle.

Technically, the Czech framework rests on Regulation (EU) 2019/2144 read together with UN Regulation No. 157 on Automated Lane Keeping Systems. The 01 series of amendments to that UN regulation, in force since January 2023, extended the maximum operational speed of automated lane keeping to 130 km/h and added provisions on automated lane changes, making motorway-type Level 3 operation technically approvable. By contrast, Commission Implementing Regulation (EU) 2022/1426 covers fully automated vehicles without a human fallback, a use case the amended Czech rules, which retain the person in the driver’s seat, do not accommodate.

In a functional comparison of five European jurisdictions, the study identifies Germany as the closest direct analogue to the Czech approach to conditional automation with a human fallback driver. France and the United Kingdom provide broader partial comparators: France built its framework on a 2019 mobility law and a 2021 ordinance on criminal liability, while the United Kingdom’s Automated Vehicles Act 2024 centres on an authorized self-driving entity. The UK act, however, is predominantly prospective; only a limited set of provisions took effect on 1 January 2026, the authorization regime itself is not yet operative, and full implementation is targeted for 2027. Sweden and the Netherlands serve primarily as contrasts, based on permit- and exemption-based experimental regimes rather than regular operation.

The empirical half of the study reveals a public largely unprepared for these changes. Declared knowledge of autonomous mobility was generally thin: 32 percent of respondents reported no knowledge, 39 percent a low level, 25 percent a medium level, and only 4 percent a high or very high level. Reactions to the regulatory scenario presented in the questionnaire varied across four analytical age cohorts, but the association was small (Cramer’s V = 0.12). Positive responses declined from 23.4 percent among respondents aged 18 to 28 to 8.4 percent among those aged 61 and above, while negative responses rose from 34.1 percent to 56.6 percent. Neutral answers were the largest single category in every cohort, ranging from 35.0 to 42.6 percent. Prior awareness followed a similar, similarly weak age gradient, falling from 17.7 percent in the youngest cohort to 6.8 percent among the oldest (Cramer’s V = 0.11).

Perceived benefits and concerns showed consistent patterns. Increased driver comfort was the most frequently selected benefit in every age cohort, while the share selecting “no benefits” climbed from 14.2 percent in the youngest cohort to 40.5 percent among the oldest. The most frequently selected concern in every cohort was deterioration of driving skills and over-reliance on technology, rising from 41.8 percent among 18-to-28-year-olds to 50.5 percent among those aged 61 and above. Takeover-related concerns varied only slightly across cohorts, between 31.3 and 35.6 percent, while ethical and environmental concerns were more often selected by younger respondents. Regional differences in mean acceptance scores, from roughly 2.76 to 2.90 in the highest-scoring regions down to 2.27 to 2.34 in the lowest, were judged exploratory, since regional subsamples differed substantially in size and the analysis did not control for compositional factors.

The authors are careful about what these numbers do and do not show. The survey vignette described SAE Level 3 operation in simplified terms, stating that while the automated system was active, responsibility for driving rested with the vehicle manufacturer, and that the driver had to take over within ten seconds of a request. Both statements diverge from the actual law, which conditionally disapplies selected driver duties and creates an operator’s administrative defence without transferring legal responsibility to the manufacturer, and which demands immediate rather than timed takeover. The acceptance variable therefore measures reactions to a stated manufacturer-responsibility scenario, not informed acceptance of the Czech amendment or of any specific liability regime. Because the survey used a non-probability quota sample without weighting, no post-stratification, and no available recruitment para-data, the findings describe the achieved sample rather than design-based estimates for the Czech adult population, and the cross-sectional design cannot separate age effects from cohort effects. The study’s practical conclusion is correspondingly modest: implementation is occurring in an environment of limited prior awareness and substantial uncertainty, so public communication must precisely distinguish the duties that remain with the driver, the meaning of a takeover request and the evidentiary role of stored operational data, and attitudes should be monitored after implementation, when they may begin to reflect real experience rather than an abstract description.

Subject of Research: Public awareness and attitudes toward the Czech legal framework for SAE Level 3 automated driving before its implementation

Article Title: Pre-implementation public attitudes toward the Czech legal framework for SAE level 3 automated driving

Article References: Skokan, A., Kukla-Młynarczyk, W., Blažková, M., Šimonová, L., & Valihrach, D. (2026). Pre-implementation public attitudes toward the Czech legal framework for SAE level 3 automated driving. European Transport Research Review, 18(1), Article 95. https://doi.org/10.1186/s12544-026-00853-x

Image Credits: AI Generated

DOI: 10.1186/s12544-026-00853-x

Keywords: SAE Level 3, automated driving, Czech Republic, road traffic law, public acceptance, survey, UN Regulation No. 157, vehicle operator liability, takeover request, Germany comparison, quota sample, driver duties