Chuhee LeeIndependent research
All publications
Cover art: three bars to scale, hours a year per person riding as a car passenger, lost to congestion, and flying.
Sizing
United States
Concepts
US · Europe · China · Korea · Japan
Lens
Driver and passengers
Basis
Official statistics first
OpportunitiesMobilityOctober 2026 · Opportunity brief

One cabin, several people

Several people share many car trips, but the car is built around one of them. This brief sizes that problem against issues the industry already treats as priorities, looks at what has been tried, and invites solutions.

Research questionIs the shared car cabin, where a driver and passengers ride together, a problem large enough to be worth solving, and where could a solution start?

By Chuhee Lee≈16 min readData as of 17 Oct 2026One-time brief, not refreshedChange log

With research assistance from Claude (Anthropic); figures cross-checked with other AI tools and primary sources · Views are the author’s own

Cite asChuhee Lee, One cabin, several people, opportunity brief, October 2026 (data as of 17 October 2026), https://blueplanet-ai.github.io/shared-cabin/, CC BY-NC 4.0.

The brief in 60 seconds

Three takeaways · each with its source
  1. 1.5 Fact

    The car usually carries more than its driver, and family vehicles carry the most

    Household vehicles carried an average of 1.5 people in 2022; vans carried 2.1 and SUVs 1.7.

    Source: US Department of Energy, from NHTS 2022 [2]
  2. 35% Fact

    Passenger screens, as built, mostly go unused

    35% of owners with a passenger screen have never used it, only 6% use it on every drive, and passenger screens draw 21.3 problems per 100 vehicles.

    Source: J.D. Power 2026 Tech Experience Index [8]
  3. Uncounted Hypothesis

    Large, unmeasured, unsolved

    Congestion ($269 billion a year), distraction (3,208 deaths in 2024) and charging (46% of EV rejecters) each have a cost and an owner. The shared cabin has neither, which may be one reason it stays fragmented, and why it may be an opening.

    Source: TTI [5]; NHTSA [6]; J.D. Power [10]; conclusion is the author’s

Fact: stated by the cited source. Interpretation: the author’s reading of the cited evidence. Hypothesis: expected from the evidence but not yet observed. Figures inside every takeaway are sourced.

At a glance

One-page summary

Headline findingThe car is designed around one person, the driver. Everyone else in it is a guest using their own phone.

Summary

The short version

The car is designed around one person, the driver. Everyone else in it is a guest using their own phone. Measured in hours, the time Americans spend riding as passengers is comparable in scale to the time they lose to congestion (the two measures differ) and about ten times the time they spend flying. Yet unlike congestion, distraction or charging, this brief found no widely used measure of its cost, and decades of attempts have not produced a good answer. This brief does not propose a product. It sets out the evidence and invites solutions.

Framing · author’s hypothesisThe shared cabin may be a coordination problem more than a missing gadget. Phones, voice assistants, seat and audio controls each serve one person or one device. In the sources reviewed, none coordinated the passengers, their phones, the car’s state and the driver’s attention as one experience. This is the author’s interpretation, not a finding; chapter 6 offers illustrations and the questions a solver would need to answer.

Shared trips are common, and passenger time is large.

Household vehicles carried an average of 1.5 people in 2022; vans carried 2.1 and SUVs 1.7.[2] Together they carried about 0.79 trillion passenger-miles that year, roughly 70 to 104 hours a year per person (author’s arithmetic). That is comparable in scale to the 63 hours Americans lost to congestion in 2024, although riding time and delay are different measures.[1][5]

Passenger screens, as built, mostly go unused.

35% of owners with a passenger screen have never used it, only 6% use it on every drive, and passenger screens draw 21.3 problems per 100 vehicles.[8]

Passengers, especially children, are the ones who feel unwell.

University of Michigan engineers estimate that about a third of adults and half of children are affected by motion sickness, and that passengers suffer it far more than drivers. In a UK survey, 34% of drivers with children said their children get car sick.[12][13]

The pieces exist, mostly in China; no whole answer turned up in the sources reviewed.

Seat-zoned voice, rear-seat entertainment and motion-sickness drive modes are already in production. No concept found combines knowing who is asking, what the driver’s attention allows, and how passengers’ own phones fit in (author’s analysis).[24][30]

Large, unmeasured, unsolved.

Congestion, driver distraction, phone integration and EV charging each have a cost figure and an owner. The shared cabin has neither. That gap, more than the technology, is why the problem persists (author’s interpretation).

Passenger time is comparable in scale to congestion delay

Hours per person per year · United States

Car passengerauthor’s arithmetic, NHTS 2022
70–104
Congestion delayTTI, 2024
63
Flyingauthor’s arithmetic, BTS
7–9

Bars to scale (125 hours = full width). Solid bars show the low end; dashed outlines show the range, which depends on an assumed average speed (see Method). Denominators differ: the passenger and flying figures are per resident, while TTI reports congestion delay per traveler. Sources: NHTS 2022 [1], TTI [5], BTS via FRED [11].

Chapter 1 of 7
Problem

A cabin built for one

The car knows one person, the driver; everyone else is a guest using their own phone, and anything offered to passengers must not cost the driver attention.

Shared trips are common. Household vehicles carried an average of 1.5 people in 2022; vans carried 2.1 and SUVs 1.7.[2] Weighted by distance, household vehicles carried about 1.43 people per vehicle-mile (author’s arithmetic from NHTS: 2.64 trillion person-miles over 1.85 trillion vehicle-miles).[1] FHWA’s own mileage-weighted occupancy factor is 1.52, which would imply about 0.96 trillion passenger-miles; this brief uses the lower figure.[42]

Passengers cannot always use the time. University of Michigan engineers estimate that motion sickness affects about a third of adults and half of children, and that passengers suffer it far more than drivers.[12] This is an attributed estimate, not a measured population figure. In a UK survey of 2,021 drivers, 34% of those with children said their children have been car sick; 27% of those avoid car travel with them altogether, and 41% of drivers with children dread long journeys with them.[13]

Passenger screens, as built, mostly go unused. In J.D. Power’s 2026 Tech Experience Index (68,084 owners), 35% of owners with a passenger screen had never used it and 6% used it on every drive; the screens drew 21.3 problems per 100 vehicles.[8] This shows that passenger screens as built are not earning their place; it does not show whether passengers would use or pay for a better-integrated experience. One truck maker’s engineer reportedly found that front passengers prefer their own phone and more storage, while rear passengers are often children without phones.[36]

Every passenger request competes for the driver’s attention. 3,208 people died in distraction-affected crashes in the US in 2024.[6] That count covers every kind of distraction, not only passenger requests. Of owners who reported a distraction-related problem in J.D. Power’s 2026 quality study, 46% pointed to the infotainment or touchscreen.[9] Crash risk rises when teen drivers carry teen passengers.[7]

Problem statement · author’sSeveral people share the trip, but the cabin serves one of them. Passengers’ time goes unused or turns unpleasant, children get sick or bored, and anything the car offers them must not cost the driver attention.
Chapter 2 of 7
Problem size

How big is it?

Measured against problems the industry already treats as priorities, the shared cabin is largest on exposure, moderate on pain, and the only one with no widely used measure of its cost.

The comparison uses four yardsticks: how many people and hours are exposed, how many are affected, the cost that is already recognized, and whether the problem is being worked on. Figures are US unless stated.

IssueExposureHow many are affectedRecognized costStatus
Shared cabin (this brief)About 0.79 trillion passenger-miles; 21 to 32 billion passenger-hours a year; 70 to 104 hours per person at an assumed 25 to 37 mph (author’s arithmetic)[1]About a third of adults prone to motion sickness (University of Michigan estimate); 34% of UK parents report car-sick children; 35% of passenger screens never used[12][13][8]Not measured. This brief found no study that puts a cost on unserved passenger time; congestion’s cost is measured at $269 billion a yearUnsolved for decades
Traffic congestion63 hours lost per person, 2024Urban travelers$269 billion a year[5]Recognized; a public-policy target
Driver distractionEvery trip3,208 deaths, 2024[6]Regulated (EU distraction warning in all new cars from July 2026)[38]Recognized; regulated
Phone–car integrationMost new carsCarPlay and Android Auto connectivity is the largest single driver of change in infotainment problems, the only quality category that got worse in 2026[9]Warranty and satisfaction cost (not public)Recognized; worked on
EV charging anxietyEV shoppers46% of EV rejecters cite charging availability; 44% cite charging time[10]Lost EV sales; public charging programsRecognized; funded
In-flight entertainment (comparator)1.14 trillion US airline passenger-miles; 2.3 to 2.9 billion passenger-hours; 7 to 9 hours per person (author’s arithmetic)[11]All flyersAirlines treat it as a product categoryServed
Reading the table. On exposure, time spent as a car passenger is roughly 8 to 14 times the time Americans spend flying, and per person it is comparable to congestion delay (the measures differ). On intensity it is weaker and less well measured than distraction or charging: most passenger time is not painful, it is unserved; motion sickness and family trips are where it hurts. On recognition it stands alone: the other issues each have a cost figure and an owner. Denominators differ between rows, so the comparison holds in direction, not as a like-for-like ranking. Passenger hours show exposure, not unmet need or willingness to pay.
Chapter 3 of 7
Timing

Why now

The hardware is already in the cabin, assistants can hold a conversation, and regulation now draws the line around the driver’s attention.

Seat awareness is becoming software. Analog Devices says its personal sound zone technology runs on speakers already fitted and could be sold as a paid feature after purchase.[18] Hyundai’s Gleo AI applies “turn on my heated seat” to the seat the request came from.[25] Garmin’s Unified Cabin routes audio and video to individual seats.[23]

Conversation is no longer the hard part. Every assistant in this site’s benchmark already understands natural language; the spread is in what it can operate.[40] The missing pieces are knowing who is asking and what the driver can afford at that moment.

Regulation sets the driver’s boundary. The EU’s distraction warning is required in all new cars from July 2026, which makes the rules for passenger features clearer.[38]

Other markets already treat the rear seat as a customer. In China, close to 900,000 new cars with rear entertainment screens were sold in 2024, up 122%, mostly premium SUVs and MPVs.[37]

A way to charge for it exists. A volume carmaker includes in-car gaming in a connected-services tier priced at $149 a year after the free period.[14]

Chapter 4 of 7
Landscape

Who is already there

Partial answers exist in the US and Europe and more in China, Korea and Japan; none covers the whole problem, and none publishes measured outcomes for passengers.

Listed to show what has been tried, not ranked. No company named here sponsored, reviewed or endorsed this brief.

Startups and suppliers

CompanyWhat it does for the shared cabinStage and evidenceWhat it shows
AirConsole (N-Dream, owned by KPIT)Multiplayer games on the car’s screens, with passengers’ phones as controllersIn models from several European brands, a US volume carmaker (six 2027 models) and Tata[14][15]
Company claim
Phone and cabin working together; exit to a supplier
KardomeIdentifies who is speaking and from which seat, on existing microphonesSeed led by Hyundai; $10m Series A, November 2024[17][16]Seat awareness without cameras
Analog DevicesPersonal sound zones as software on existing audio hardwareAnnounced August 2026[18]Suppliers can ship it as a feature
SilentiumActive noise cancellation and a “personal sound bubble” per seatNoise cancellation available on Cadence DSPs; sound-bubble production status not found[19]Per-seat audio from a specialist
AudioscenicHead-tracked 3D audio for individual listenersAbout £1.5M raised (2020) with automotive as a target; no automotive customer found[20]Early and small
Motion SyncSeat cues that warn the body of upcoming turns and braking, against motion sicknessUniversity spin-out; moved from an automaker pilot to a consumer retrofit product[21]The aftermarket route
HolorideMotion-synced VR for passengersInsolvent in 2024 after a funding round failed; later restarted[22]The cautionary case

Concepts to learn from: China, Korea and Japan

Part of the problemConceptMarket and statusWhat it shows
Who is askingLi Auto: four voice zones (driver, front passenger, rear left, rear right); a “sound lock” so the owner can ignore rear-seat commands; a passenger’s “me too” copies the driver’s setting[24]China; shipped by over-the-air update, September 2021
Shipped
Seat-zoned voice has worked for five years
Who is askingHyundai Gleo AI: knows which seat a command comes from[25]Korea, May 2026, then global; Pleos Connect, which includes Gleo AI, targeted for about 20 million vehicles by 2030
Target
Seat awareness reaching volume brands
Per-seat entertainmentHuawei HarmonySpace 6 on the Yijing X9 six-seat SUV: rear-seat projection screens, karaoke without microphones from any seat, reminders for a forgotten child or pet[26]China, August 2026
Announced
The rear seat treated as a main customer
Per-seat entertainmentSony Honda AFEELA 1 (canceled 2026): displays for each seat, spatial sound, rear entertainment on the top trim[27]Planned for California; Sony Honda Mobility canceled the car and a second model in March 2026, before any delivery.[43] On the prototype, how apps behave while moving was still being studied[28]
Canceled
Screens per seat; separation from the driver unresolved
Per-seat entertainmentBYD: karaoke in new premium models[29]China, 2026
Reported
The cabin as a social space
Per-seat entertainmentChrysler Pacifica: available seatback touchscreens for second-row passengers with Amazon Fire TV built in[46]United States; 2026 model year, optional
Shipped
Rear screens in a family vehicle; separation from the driver and from passengers’ own phones is not described
Motion sicknessXiaomi YU7 motion sickness relief mode: tunes power delivery, regenerative braking and suspension; company tests show nausea frequency down 51% and onset 16% later[30]China, July 2025
Company claim
Software on the drivetrain, no new hardware
Motion sicknessHonda: acceleration mapping that smooths low-speed EV torque[31]Japan; in the e:Ny1, 2023
Reported
The same idea from a Japanese carmaker
Motion sicknessNissan Serena e-POWER sold as the “anti-motion-sickness MPV”: motor-only drive, suspension, large windows, supportive seats[32]Japanese model sold in Malaysia, April 2026; no test data published
Company claim
Motion sickness as a family-car selling point
Motion sicknessHuawei patent: on-screen images of the car’s motion against the road for rear passengers[33]China; published September 2023
Patent
The phone’s motion-cue idea on cabin screens
Motion sicknessApple Vehicle Motion Cues: animated dots on the edges of an iPhone or iPad screen that follow the car’s motion, switched on automatically when the phone senses it is in a car[45]United States and worldwide; iOS 18, fall 2024
Shipped
A phone-side answer already ships; the cabin does not use it
Motion sicknessGAC and the First Affiliated Hospital of Sun Yat-sen University: joint research and standards[34]China; agreement signed July 2026
Announced
Moving toward clinical evidence
ContrastLG AI Cabin Platform: cabin cameras and context-aware voice aimed at the driver[35]Korea; CES 2026 private demonstration
Announced
New “AI cabin” platforms still center on the driver

Li Auto’s own data, from 2021, supports the problem: it reported that more than 66.7% of trips carried several people, and that front passengers used the voice assistant more than 72% of the time and rear passengers more than 44%.[44] Passengers are not a side case in a car built for families, so the assistant has to know who is speaking. (The English trade report of the same release gave 65% passenger-seat occupancy and attributed the 72% to the driver’s seat; the Chinese original reads as above.)[24] Most entries in this table are company claims or trade press. Labels: Shipped is on sale; Reported is a trade-press report of a product; Company claim is a result or claim by the maker; Announced is planned or demonstrated, not on sale; Patent is a filing, not a product.

Chapter 5 of 7
The phone test

Why not just the phone?

The phone keeps each passenger’s content and identity; the car adds what no single phone can do, so the likely answer orchestrates both.

CapabilityPhoneCarEdge
Each passenger’s own content, accounts and paymentStrongWeak unless linked to the passenger’s accountPhone
Shared screens and a shared experience for several peopleOne person per deviceBuilt in (AirConsole uses phones as controllers on the car’s screens)[15]Car
Separate sound per seat without headphonesNot possiblePersonal sound zones on existing speakers[18]Car
Knowing which seat is speakingNot possibleSeat-zoned voice (Li Auto, Hyundai, Kardome)[24][25]Car
Knowing the car’s next maneuver (for motion sickness)Infers from its own sensors (Apple’s motion cues)Native vehicle state; can also change how the car drives (Xiaomi, Honda)[30][31]Car
Knowing what the driver’s attention allows right nowNot possibleDriver monitoring and driving state are in the carCar

Author’s analysis based on the evidence cited. Products that fight the phone repeat the passenger-screen result; a promising design hypothesis uses passengers’ phones and the cabin together.

Chapter 6 of 7
Invitation

Business model and entry wedges

Three routes to revenue exist today; four starting points look open to a solver.

Ways to get paid

Through the carmaker

  • License or per-car software fee, usually through a tier-one supplier.
  • Long design cycles; suppliers can bundle the same feature.
  • An idea, with no example found: content or game platforms paying carmakers or software providers for built-in delivery to several passengers.

As a paid tier or upgrade

  • Connected-services tier or a feature unlocked after purchase.
  • The only public price found is a volume carmaker’s $149 a year tier[14]; take rates are not public.

Aftermarket and retrofit

  • Products for the cars already on the road.
  • Motion Sync’s route[21]; Holoride shows the risk[22].

Why no whole answer yet: the author’s hypotheses

If most of the pieces exist, why has no one joined them? These are possible reasons to investigate, not findings of this brief.

  1. Split ownership. Voice, audio, screens, seats and climate usually come from different suppliers and teams.
  2. Closed vehicle data. Access to seats, microphones and vehicle state differs by carmaker, and a phone cannot reach most of it.
  3. Long design cycles. A feature agreed now reaches cars years later.
  4. Hard to know who is asking. Seat awareness and permission rules are needed, and this brief did not find them settled.
  5. Driver safety limits. Anything shown or said to passengers must not reach the driver’s view or attention.
  6. Unproven payment. Owners often do not use passenger screens, and no source reviewed shows passengers paying for a better experience.
  7. No shared standard found. This brief did not find a common way to link passengers’ phones with cabin systems; it did not search for one in depth.

Entry wedges, as invitations

WedgeWhat it would doEvidence that it matters
Family long-trip modeMotion-aware, seat-aware entertainment for children that never reaches the driver’s view, possibly paired with gentler drive modes34% of UK parents report car-sick children; 41% dread long trips with them[13]
“Who’s asking” layerSeat identification plus driver-attention rules for any assistantLi Auto reported front-passenger voice use above 72% and rear above 44%, with over 66.7% of trips carrying several people[44]
Sound per seatSeparate audio per seat as a software upgrade on existing speakersSupplier says it can be sold after purchase[18]
Motion-sickness mitigationUses the car’s knowledge of its next maneuver, in the drivetrain, seat or screensXiaomi reports 51% fewer nausea episodes in its own tests[30]

Illustrations to spark your own ideas

These are examples of the kind of answer a solver might explore. They are the author’s illustrations, not recommendations and not validated: each builds on the evidence above and would need testing with real passengers, drivers and carmakers.

A coordination layer between phones and the cabin

  • Each passenger’s phone acts as the controller, while the car supplies seat position, audio and vehicle state, a rule for what the driver can safely be shown, and a rule for who may ask for what.
  • Pieces exist: phones as game controllers[14] and seat-zoned voice[44]. Open question: who owns the layer, the carmaker, a supplier or a new company?
Author’s illustration

Comfort for passengers in electric cars

  • Motion sickness may matter more in electric cars, where regenerative braking and quick acceleration change how a ride feels. The trade-press piece cited here describes the link, and Xiaomi and Honda tune their drivetrains for it.[30][31]
  • A cabin-level answer could combine drivetrain tuning with seat cues and screens. Open question: how much relief, for whom, and who pays?
Author’s illustration

Passenger features that run in the car, not the cloud

  • Seat-aware requests, per-seat audio and comfort settings handled on the car’s own computer could work without a connection and keep voices and audio off the cloud.
  • Components already run on existing hardware: seat-aware voice on existing microphones[16] and sound zones on existing speakers[18]. Open question: what can the car’s computer carry safely?
Author’s illustration

Questions any solver would need to answer

  1. Which passenger problem comes first? Motion sickness, boredom, fights over the screen or audio, or requests that distract the driver.
  2. Who feels it most? Families on long trips, the daily school run, or groups of adults.
  3. What do passengers do today? Phones, headphones, a tablet, or nothing.
  4. Who is the user, who benefits, who buys and who distributes? The passenger has the problem, but the owner, the carmaker or a supplier may control the purchase.
  5. Who may ask the car for what? A rear passenger changing the temperature is different from one changing the route.
  6. Can it work across cars and stay safe for the driver? Access to seats, audio, microphones and vehicle data differs by carmaker.
  7. Who captures the value? A carmaker’s own team, a supplier or a new company.

What would show the opportunity is smaller than it looks

A good brief also says what would weaken its case. These tests are the author’s suggestion for any solver, not results.

TestEvidence that would support itEvidence that would weaken it
Passengers have a real unmet needInterviews and observed trips show recurring, consequential problemsMost passengers are content with phones, headphones and existing cabin features
Coordination beats separate appsUsers repeatedly prefer a coordinated experience on real tripsUsers prefer separate devices and do not value coordination
The car adds something a phone cannotSeat identity, audio and vehicle state enable useful features phones cannot matchPhone-only solutions give comparable results at lower cost
A buyer existsA named buyer pays for, or funds a pilot of, the solutionInterest is positive but no one accepts the cost or the integration work
It scales across carsOne integration approach works across several platformsEach carmaker needs bespoke engineering and its own negotiation
Chapter 7 of 7
Judgment

Risks, potential and complexity

Exposure is large and solid; demand, payment and the supplier squeeze are the open questions.

Demand is unproven. Owners don’t use passenger screens, and stated interest often doesn’t become payment.[8]

Suppliers move fast. Audio, voice and cabin suppliers can add seat awareness as a feature, which squeezes startups; the best-known passenger startup so far, AirConsole, now belongs to a supplier.[15]

Privacy. Reusing driver-monitoring cameras for passenger features faces the EU’s closed-loop rules; voice and audio approaches avoid cameras.[38] How the rules apply to passenger features has not been checked against the official texts for this brief.

Measurement. No study measures the cost of unserved passenger time directly. The exposure figure is solid; the pain is shown through proxies.

Potential vs complexity · author’s viewPotential: high on exposure, unproven on demand and payment. Complexity: medium to high. Passenger hours show exposure, not proof of need. Exposure per person is comparable in scale to congestion delay and about ten times flying in hours. Most hardware already exists; the hard parts are orchestration across phone and car, rules for the driver’s attention, and carmaker integration cycles.

Method, limits and flags

Read before citing

Question. Is the shared car cabin, where a driver and passengers ride together, a problem large enough to be worth solving, and where could a solution start?

Scope. Problem sizing is for the United States. Concepts are drawn from the US, Europe, China, Korea and Japan. Placed under Opportunities / Mobility on this site. This is an invitation to solve, not a recommendation to any company.

Evidence base. Builds on this site’s Mobility Product Outlook, Physical AI Outlook 2026–2031 and In-Car AI Assistant Benchmark, with Agentic AI and the Consumer as supporting evidence.[39][38][40][41]

Sources. Official statistics first (FHWA, BTS, NHTSA, USDOT), then benchmark studies (J.D. Power, TTI), then company releases and trade press. Each reference carries its type and date.

Arithmetic (author’s). Household-vehicle person-miles = 23.69 daily miles per person × 305.561 million people × 365 = 2.64 trillion; minus 1.848 trillion household vehicle-miles = 0.794 trillion passenger-miles (NHTS 2022).[1] Passenger-hours = passenger-miles ÷ an assumed 25 to 37 mph (the upper speed is what DOE’s reported driving time and distance imply) = 21.5 to 31.8 billion hours, or 70 to 104 hours per person. Illustrative value of time (not a loss, a market size or willingness to pay) = hours × $21.80 per person-hour (USDOT all-purpose rate, 2024 dollars) = $0.47 to $0.69 trillion; at USDOT’s $20.10 personal-travel rate, $0.43 to $0.64 trillion.[4] Flying = 1.14 trillion revenue passenger-miles (12 months to June 2026) ÷ an assumed 400 to 500 mph = 2.3 to 2.9 billion hours, or 7.5 to 9.3 hours per person.[11]

What this brief is not. The problem statement, the comparison reading, the phone table, the framing, the reasons no whole answer exists yet, the entry wedges, the illustrations, the tests and the potential-versus-complexity view are the author’s analysis, not sourced figures.

Flags and open questions

  1. The hours figure depends on an assumed speed. The speed range of 25 to 37 mph is the author’s assumption. DOE reports household vehicles driven 64.6 minutes a day and 39.7 vehicle-miles a day; if both describe the same unit, that implies about 37 mph, which sets the top of the speed range and the bottom of the hours range. The units behind those two figures were not confirmed, and a different speed would move the hours.[3]
  2. Denominators differ. TTI’s 63 hours is reported per traveler; the passenger and flying figures are per resident. National congestion delay in hours was not found in the 2025 release.
  3. Motion sickness prevalence. “A third of adults and half of children” is the University of Michigan’s own estimate in an engineering news release, with no sampling method given, not a peer-reviewed prevalence study. The UK child figures are from an insurer’s survey.
  4. Passengers as a distraction source. NHTSA’s 2024 summary gives no passenger figure; the IIHS evidence covers teen drivers with teen passengers only.
  5. Company claims not independently verified: Xiaomi’s 51% figure, Hyundai’s 2030 target, Nissan’s comfort claims. Li Auto’s usage figures date from 2021.
  6. Not covered: EU problem sizing; take rates for paid connected tiers; how EU closed-loop rules apply to passenger features; Chinese demand data.
  7. One-time brief. This brief will not be refreshed; figures are as of 17 October 2026.

Change log

  • October 2026
  • Added
    First opportunity brief, data as of 17 October 2026.

Related reports on this site

Physical AI Outlook 2026–2031 (physical AI across sectors) · Mobility Product Outlook 2026–2031 (passenger screens and customer expectations) · In-Car AI Assistant Benchmark (seat awareness and vehicle control) · Agentic AI and the Consumer Outlook 2026–2031 (trust and control)

About this brief

Personal research by Chuhee Lee. Research, source verification and drafting were carried out with Claude (Anthropic); figures were then cross-checked with other AI assistants and against primary sources. Every factual claim is sourced to the references listed below. Topic selection, framing and conclusions are the author’s. These are personal research projects; the views are the author’s own and do not represent any employer or organization. No company named sponsors, reviews or endorses this brief, and product names and trademarks belong to their owners. Licensed under CC BY-NC 4.0.

References

Continue down the map

ForcesShiftsSignalsOpportunities
Follow new editions

Each report is refreshed on a schedule. Know when it changes.

Next refresh dates are listed on the home page.