Product1 publisher2 min readPublished
A Gothenburg workshop asks vehicle designers to prevent sex in driverless cabins
Pavlo Bazilinskyy's research turned up 297 amateur clips tagged to public transport, and he and his co-hosts now want engineers proposing transparent bodies and blinking lights for robotaxi interiors, with the paper to come after the workshop.
The Product Desk · Product desk
What happened
- Pavlo Bazilinskyy, an assistant professor at Eindhoven University of Technology, searched "public transport" in the amateur section of PornHub as part of his research and counted 297 hits.
- He and academic colleagues are hosting a four-hour workshop starting at 8:30 am, inside a Gothenburg industry conference on how people interact with vehicles and mobility systems.
- Attendees from academia and industry will be prompted to propose physical and design interventions that discourage people from having sex inside self-driving vehicles.
- Gothenburg is home to Volvo Cars and the tech supplier SmartEye, and BMW and Google are among the conference's sponsors.
Compiled by The Product DeskSomething wrong?How this is made
Why it matters
- decision The brief is physical design, so whatever comes out of it is hardware, committed when a fleet is ordered and not shipped later as an app update.
- exposure A cabin that blinks at behavior it disapproves of puts a classifier's judgement onto a paying rider, and there is no staff member on board to overrule it.
- constraint The session is scoped to consensual behavior, so nonconsensual incidents in an unstaffed cabin sit outside whatever this work ends up recommending.
- precedent The organizers plan to publish, which means four hours of brainstorming becomes a citable reference for in-cabin monitoring at a point when there is little else to cite.
In a taxi, what stops a rider is a stranger's eyes in the mirror. No product team specified that control; it came with the driver. Robotaxis with nobody behind the wheel are already picking up passengers in a handful of countries, including the US [4].
It is pitched as a session on how to stop people having sex in self-driving cars [5]. The decision it actually makes is about what an unstaffed shared cabin senses, and what it does about what it senses.
The evidence available so far is one number. The term Bazilinskyy typed was "public transport", so the 297 hits describe conventional transit, not a driverless cabin [2][19]. Wired reports no incident count from any robotaxi operator [19]. Bazilinskyy calls sex in automated vehicles "the elephant in the room" for people studying how humans and machines interact, and says it is badly under-studied [16].
Rouchitsas offered two ideas off the top of his head: transparent vehicles, and a light that blinks when the technology senses behavior it disapproves of [11]. Those cost different amounts. Glass is a body decision, priced once at build, and it stores nothing afterwards. A light that reacts to behavior needs something in the cabin deciding what the behavior is, and that means sensors, a retention rule and an off-vehicle support path.
The organizers' stated goal is the awkward part. "We want to see automated mobility solutions accepted, trusted by the general public," Rouchitsas said [13]. He also said the upside is largest for elderly people, children and people with disabilities, who have a harder time getting around [15]. The route to that trust runs through watching those same riders inside the vehicle. Rouchitsas is candid about the posture: the hosts are "taking the side of the parents, the elderly", he said, and "we are not the cool people here" [14].
For anyone running a fleet, the useful split is per behavior. First, whether the vehicle can tell the thing is happening without storing a recording of a person. Second, whether the response can be automated or needs a human who is not on board. Yes and yes is cheap product work, a seat sensor or a door that will not unlock. Yes and no buys a queue somewhere else and the staffing that goes with it. No and no leaves a complaint from the next rider and nothing to check it against.
Bazilinskyy's argument for using sex as the example is that the design work transfers to robbery, bullying and homicide [17]. "It could be helpful for the industry and academics to use this example to design other things," he said [18].
What to watch
- The paper the organizers said they will publish, and whether it names specific sensor or body-design requirements.
- Whether any robotaxi operator publishes in-cabin incident numbers that would size the problem this work is designing against.
- Whether the conference's sponsors or Gothenburg's local suppliers turn any of the workshop's suggestions into a product spec.