Why road safety messaging still blames the victim. Now AI is doing it too.
When a person who is walking is injured or killed on our roads or on the roadside, the first questions will too often focus on what the pedestrian did wrong. Were they distracted by their phone? Did they cross away from or without waiting for the green light? Did they ‘jaywalk’?
But the weight of evidence shows that these explanations are not the main cause of pedestrian crashes and fatalities. In fact, the real risks overwhelmingly come from driver behaviour, vehicle choice, and the way our streets are designed. Shifting the blame to people walking is not only misleading but it reflects a deeper bias in road safety thinking, which is now being repeated in how data and artificial intelligence interpret these crashes.
What the evidence shows
Distraction and jaywalking are minor factors
Yes, peer-reviewed studies show that pedestrians distracted by mobile phones are more likely to make risky crossing decisions: taking longer to cross, looking less often, or disobeying signals (Schwebel et al., 2012; Thompson et al., 2013; Useche et al., 2020). Jaywalking has also been linked to elevated risk at intersections (Wu et al., 2019).
But critically: these studies look at behavioural risk: not actual fatal crash outcomes. In Queensland and elsewhere, there is no crash dataset that attributes deaths directly to phone use or jaywalking. At most, these behaviours increase exposure, but they do not determine whether a crash will be survivable.
Distracted by distracted pedestrians?
Ralph, K; Girardeau, I (2020)
‘Practitioners were more concerned about distracted walking if they primarily use a car or spend little time in pedestrian-oriented areas (windshield bias), or if they work in engineering or public health (professional training). Most importantly, the distracted walking frame does indeed shape policy solutions.’
‘Despite relatively weak evidence about the harms of distracted walking, fully a third of the transportation practitioners we surveyed believe that distracted walking is a large problem, responsible for nearly 40% of pedestrian deaths. Among this group, some believe distraction that is an even bigger risk: 16% believe that the majority of pedestrian deaths are due to distracted walking and 9% believe that over three quarters are.’
‘An additional 50% of practitioners believe distracted walking is a “small problem”, responsible for roughly 15% of pedestrian deaths. Like their more concerned colleagues, these practitioners also prefer individual-level solutions (like pedestrian education) over systems level-solutions (reducing speeds via design).’ (Ralph & Girardeau, 2020).
The decisive factors are systemic
The factors that do explain pedestrian deaths are well established:
Speed: The risk of death when hit by a vehicle is about 10% at 30 km/h, but rises to 80% at 50 km/h (Rosén & Sander, 2009; WHO, 2017).
Vehicle size and design: SUVs and trucks significantly increase fatality risk due to higher fronts and weight thus raising pedestrian death risk by 20–30% compared to sedans (Hu & Cicchino, 2021).
Driver behaviour: Speeding, distraction, and impairment account for the vast majority of serious crashes (Oxley et al., 2018).
Road design: Streets designed for vehicle flow such as wide or multiple lanes, long crossing times or wait times, poor lighting, and missing or unformed footpaths requiring a pedestrian to walk on the road which all contribute to dramatically increasing pedestrian risk (Nieuwenhuijsen & Khreis, 2016).
This means that while pedestrians’ choices can raise their exposure to danger, the severity of outcomes is determined by how fast, how big, and how prioritised vehicles are in street design.
The bias problem in road safety
Despite this evidence, road safety reporting and policy have long carried an inherent bias:
Victim-blaming narratives: Media reports still call crashes “accidents” and emphasise pedestrian behaviour (“crossed illegally,” “was on their phone”, “was in the wrong place”, “under the influence”) rather than referring to the driver of the vehicle’s speed or the driver responsibility to operate a vehicle with due care.
Data gaps: Crash databases rarely capture “driver distraction” in detail but observational notes will highlight if a pedestrian was outside a crossing, reinforcing the perception of fault.
Engineering priorities: Road design standards have historically valued vehicle movement over pedestrian safety, embedding systemic bias in infrastructure and in our transport allocation budgets.
AI is learning these same biases
As governments and researchers begin using AI and machine learning to predict crash risk, these systems are trained on existing datasets and flawed narratives. That means:
If police crash reports emphasise “jaywalking,” AI may learn to treat pedestrians as the problem
If vehicle size or unsafe street design aren’t captured as variables, AI won’t identify them as causes
If use of “accident” language dominates, AI may normalise crashes as random events instead of preventable harms.
This creates a feedback loop of bias: flawed human assumptions → flawed data → flawed AI predictions → flawed policy responses.
A safer, fairer approach
If we want to reduce and prevent pedestrian deaths and injuries, the research points us clearly towards:
Safer streets and speed limits especially in high pedestrian areas (schools, shopping strips, retirement villages)
Drug and alcohol use and harm reduction when operating a vehicle or device
Street designs that prioritise walking: shorter crossing distances, safe crossings, slowed drivers and reducing traffic, better lighting.
Vehicle standards that reduce pedestrian harm, including front-end design improvements.
Data reform so that crash reports capture driver behaviour, vehicle type, and street or pathway context, not just pedestrian compliance.
This evidence-based approach reframes safety where it belongs: not on whether a person walked while using their phone, or walked home after dinner and wine, but on whether the street, the driving and the vehicle design around them made it possible to survive. That’s a Safe Systems design that is critical to road safety success.
Conclusion
The story we tell about pedestrian safety matters. When we focus on phones and ‘jaywalking’, we deflect attention from the real, systemic risks: speeding, drug or alcohol use, large domestic vehicles, and car-centred road design. These biases don’t just shape public opinion; they are now being encoded into the AI systems that will guide the future of transport safety.
If we want safer streets, we need to change the story. Stop blaming people walking, and start fixing the systems that put them in danger.
References
Hu, W., & Cicchino, J. B. (2021). An examination of the increased risk of pedestrian crashes with SUVs. Journal of Safety Research, 76, 272-278.
Nieuwenhuijsen, M. J., & Khreis, H. (2016). Car free cities: Pathway to healthy urban living. Environment International, 94, 251-262.
Oxley, J., et al. (2018). Pedestrian safety: Identifying risk factors in driver behaviour. Accident Analysis & Prevention, 115, 94-101.
Rosén, E., & Sander, U. (2009). Pedestrian fatality risk as a function of car impact speed. Accident Analysis & Prevention, 41(3), 536-542.
Schwebel, D. C., et al. (2012). Distraction and pedestrian safety. Accident Analysis & Prevention, 45, 266-271.
Thompson, L. L., et al. (2013). Impact of social and technological distraction on pedestrian crossing behaviour. Injury Prevention, 19(4), 232-237.
Useche, S. A., et al. (2020). Distraction of pedestrians and its role in road safety: A systematic review. Safety Science, 129, 104859.
World Health Organization. (2017). Managing speed: A policy framework for road safety. WHO.
Wu, Y., Xu, H., & Zhang, X. (2019). Effects of jaywalking on pedestrian safety at intersections. Cognition, Technology & Work, 21, 437-446.
