Bus Drivers, Transit and Intercity
53-3052.00Median wage $59,050/yr159,240 employed (US)Rank #325 of 923 scored · top 35% by substitution
Drive bus or motor coach, including regular route operations, charters, and private carriage. May assist passengers with baggage. May collect fares or tickets.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution33
Exposure31
Augmentation38
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
14 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
21%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%32
panel mean rating 2.3/5 → substitution pressure 32/100
Technical feasibility todayw 20%29
panel mean rating 2.1/5 → substitution pressure 29/100
Cost vs. human wagew 15%34
panel mean rating 2.4/5 → substitution pressure 34/100
Adoption barriersw 20%inverted — strong barriers lower the score40
panel mean rating 3.4/5 (barrier strength) → substitution pressure 40/100
Sector adoption velocityw 10%28
panel mean rating 2.1/5 → substitution pressure 28/100
Task breakdown (14 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Announce stops to passengers.
93CI 86–100 · exposure 92 · augmentation 38 · importance 4.5/5 · click for rater detail
Announce stops to passengers.
93| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Transit agencies worldwide have rapidly adopted automated stop announcement systems, particularly in urban and intercity bus operations; this is now standard practice in digitized transit infrastructure. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Public transit agencies have broadly adopted automated stop announcement systems for years, driven by accessibility regulations and cost savings. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Automated announcements assist drivers by reducing verbal workload and improving consistency, though the driver's role in this task is already minimal and augmentation potential is limited once automation is deployed. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Since this specific sub-task is already fully automated rather than human-performed with AI assistance, there is little ongoing augmentation of a human driver doing this action. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can reliably generate and deliver stop announcements via text-to-speech synthesis with high accuracy and >50% time savings compared to manual verbal delivery. The task is straightforward and repetitive, requiring no complex judgment. |
| Task automatability | claude-sonnet-5 | 5/5 | Announcing stops is a simple, repetitive verbal/audio task already fully automatable with pre-recorded or GPS-triggered automated announcement systems, meeting the time-saving threshold trivially. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or liability barriers exist for automated announcements, though some transit agencies may prefer human drivers to retain the task for customer experience or union agreements, creating moderate organizational friction. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory barrier requires a human to verbally announce stops; automated systems are widely accepted and often mandated for ADA compliance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The cost of automated announcement systems (one-time infrastructure plus minimal inference) is orders of magnitude cheaper than paying a human driver to verbally announce every stop for multiple daily routes. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Once installed, automated announcement systems cost negligible per-trip compared to any human labor allocated to this sub-task, and hardware/software costs are amortized over years of use. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Production transit systems already deploy automated stop announcement systems (audio playback, digital displays, and TTS) at scale worldwide; this is a mature, reliably functioning technology in real operations. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated stop announcement systems are standard equipment on transit buses in most major cities today, operating reliably at scale via GPS-linked audio and visual displays. |
Record information, such as cash receipts and ticket fares, and maintain log book.
84CI 76–92 · exposure 87 · augmentation 50 · importance 4.1/5 · click for rater detail
Record information, such as cash receipts and ticket fares, and maintain log book.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Transit agencies globally have widely adopted automated fare collection, ticketing systems, and digital logging over the past 15+ years; this is mainstream infrastructure in digitized transportation sectors, though some smaller or regional systems remain partially manual. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Many transit systems have modernized fare collection with smart cards and apps, but numerous agencies, especially smaller or intercity bus operators, still use manual or semi-manual logs. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Digital logging and automated fare systems assist drivers by reducing manual paperwork burden and providing real-time records, though the human role becomes primarily oversight and exception handling rather than active task participation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where automated systems aren't fully deployed, mobile apps and digital forms can still assist drivers in reducing time spent on manual recording and reconciliation. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Recording cash receipts, ticket fares, and maintaining logbooks involves structured data entry and documentation that can be fully automated through point-of-sale integration, automated fare collection systems, and digital logging—easily achieving 50% time savings at equal or better quality with current technology. |
| Task automatability | claude-sonnet-5 | 4/5 | Recording cash receipts and fares and maintaining logs is a structured data-entry task easily handled by digital fare systems, automated ticketing, and simple software, meeting the time-saving threshold with off-the-shelf tools. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some transit agencies have legacy systems or regulatory requirements around auditable records, there are no fundamental legal barriers preventing automation—only organizational inertia and established procurement patterns that create modest friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement ties this specific record-keeping subtask to the driver; agencies just need capital investment and integration with existing fare/payment infrastructure. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated fare collection and digital logging systems cost far less to operate per transaction than employing humans to manually record fares and maintain paper or manual digital logs, with negligible ongoing inference costs compared to labor. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Electronic fare collection and automated logging systems cost far less per transaction than manual driver record-keeping once implemented at scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, deployed products for transit fare collection, automated ticketing systems, and digital logbook maintenance exist and operate reliably at scale across major transit agencies worldwide, reliably capturing and recording transaction data. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Automated fare collection systems, smart cards, and digital logging are already deployed at scale in many transit agencies, though some smaller or older systems still rely on manual logs. |
Read maps to plan bus routes.
84CI 79–89 · exposure 80 · augmentation 88 · importance 4.2/5 · click for rater detail
Read maps to plan bus routes.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Transit agencies and intercity bus operators widely use routing and fleet management software in production; adoption is fast in digitized transport sectors. Most medium to large operators have already integrated automated route planning tools. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Transit and logistics industries have broadly adopted GPS and routing software, though full route design authority may still involve human planners at the agency level. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI mapping and routing tools substantially augment driver productivity by automating route selection, highlighting turn-by-turn guidance, and flagging traffic disruptions, while drivers retain control over final decisions and adapt in real time. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI-based mapping tools dramatically speed up and improve accuracy of route planning, and are routinely used alongside human oversight for schedule and stop adjustments. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current AI systems can reliably extract route information, optimize paths, and interpret map data with high accuracy. Autonomous route planning using mapping APIs, geospatial data, and routing algorithms is well-established, reducing the task time by well over 50% compared to manual planning. |
| Task automatability | claude-sonnet-5 | 4/5 | Modern GPS and route-planning software (Google Maps, transit routing systems) can automatically generate optimal routes, largely replacing manual map-reading for route planning purposes. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers prevent automation of route planning itself; the task is purely analytical. The primary friction is organizational (preference for driver input, union concerns about deskilling) rather than hard compliance requirement. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal requirement mandates a human read paper maps; software-assisted route planning is standard practice with no regulatory obstacle. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated route planning via mapping APIs costs pennies to dollars per plan, while a human driver's hourly wage (often $20–40/hour with benefits) makes AI roughly an order of magnitude cheaper for equivalent output. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Digital mapping and routing software cost pennies per use compared to the time a human would spend manually planning routes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (Google Maps, routing engines, fleet management software) perform route optimization and map reading reliably in production across transportation firms. Minor edge cases around manual adjustments or client-specific constraints may still require human input, but core task delivery is mature and production-ready. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | GPS navigation and route-optimization software are mature, widely deployed products used daily by transit agencies and drivers alike. |
Regulate heating, lighting, and ventilating systems for passenger comfort.
61CI 55–67 · exposure 66 · augmentation 63 · importance 4.1/5 · click for rater detail
Regulate heating, lighting, and ventilating systems for passenger comfort.
61| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Transit agencies adopt climate automation slowly due to aging fleet infrastructure, budget constraints, and union agreements that protect driver roles. Adoption is concentrated in newer urban systems and intercity coaches, while many local transit buses retain manual controls. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | placeholder |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Automated climate systems substantially enhance driver productivity by removing the need for continuous manual adjustment, allowing the driver to focus on navigation and passenger safety. Real-time feedback systems further assist decision-making when conditions need tweaking. |
| Augmentation potential | claude-sonnet-5 | 3/5 | placeholder |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Modern buses increasingly feature climate control systems that can automatically adjust heating, lighting, and ventilation based on sensor inputs and preset comfort parameters. While some manual oversight and adjustment may remain, AI-integrated HVAC systems can handle 70–80% of this task autonomously, meeting the ≥50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 4/5 | Adjusting HVAC and lighting is a simple control task that automated climate systems and sensors can already handle without human intervention on many modern vehicles.'},'feasibility':{'rating':3, |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory requirements for driver presence, passenger safety oversight, and liability for malfunctions create significant barriers. Labor agreements and customer expectations also require human drivers to remain responsible for environmental conditions, even if systems are automated. |
| Adoption barriers | claude-sonnet-5 | 2/5 | placeholder |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Sensor-based automated climate and lighting systems have low marginal inference costs once installed; ongoing maintenance and oversight are minimal compared to the continuous labor of a human operator adjusting these systems throughout a shift. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | placeholder |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed climate control and smart lighting systems exist in modern transit buses and are production-ready, but integration with AI agents for end-to-end autonomous regulation remains partial and variable across transit systems. Many systems still require human monitoring and manual intervention. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | placeholder |
Report delays or accidents.
34CI 25–43 · exposure 33 · augmentation 63 · importance 4.2/5 · click for rater detail
Report delays or accidents.
34| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Transit agencies are risk-averse and digitization of reporting is slow; most still rely on radio/manual logs despite telematics deployments. Pilot programs exist but production adoption of autonomous incident reporting remains minimal across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Public transit is a moderately digitized but often resource-constrained and unionized sector, with adoption of automated reporting tools proceeding slowly and unevenly across agencies. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI dashcam analysis and automatic alert systems can assist drivers by flagging potential incidents in real time, reducing missed reports and improving documentation, but the driver remains the decision-maker on whether and how to report. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enabled dispatch and communication tools (e.g., automatic incident logging, voice-to-text transcription, GPS-triggered alerts) can significantly speed up and standardize how drivers report delays or accidents. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Detecting and recognizing delays or accidents requires real-time perception of road conditions, traffic, and incident classification—tasks where AI vision systems show promise but still struggle with edge cases, ambiguity in severity assessment, and integration into existing reporting systems. Current systems cannot reliably end-to-end replace a driver's judgment about reportable incidents without significant human oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | Reporting delays/accidents involves communicating structured information (time, location, nature of incident) that AI voice/text systems could capture and log, but the initial detection and judgment of severity still require human input.imin.g. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety reporting is often legally mandated and may require a qualified human witness or certified operator to file reports; liability asymmetry means false or missed reports carry serious consequences. Regulatory frameworks in transit require documented human accountability for incident reporting. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Accident reports often require accurate factual documentation for insurance, liability, and regulatory purposes, creating moderate barriers to full automation of driver-side reporting, though not a licensing requirement per se. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI-assisted incident detection systems (hardware, cloud processing, integration) cost roughly as much or more than the marginal cost of a driver performing the reporting task manually, especially when factoring in false-positive handling and human oversight overhead. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated logging/dispatch systems are relatively cheap to run once integrated, but the human driver still must trigger and describe the incident, so cost savings are only partial compared to fully manual reporting. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with incident detection (dashcam footage analysis, telematics alerts), no deployed product reliably performs autonomous incident reporting at scale without human verification. Existing systems flag events but require a human to validate and formally report them into dispatch/safety systems. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some transit fleets use telematics and automated incident-detection systems that auto-generate delay reports, but driver-initiated verbal/written accident reporting is still largely manual in most transit agencies today. |
Collect tickets or cash fares from passengers.
33CI 20–46 · exposure 25 · augmentation 38 · importance 4.3/5 · click for rater detail
Collect tickets or cash fares from passengers.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of automated payment systems (like contactless readers) is slowly increasing in some transit agencies, but driver-based fare collection remains the norm in many regions; adoption is lagging and piecemeal rather than rapid and comprehensive. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Public transit agencies have adopted automated/contactless fare systems at a moderate pace, with wide variation by city and funding, reflecting middling public-sector digitization speed. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Automated fare systems (contactless cards, mobile payments) can assist drivers by reducing manual cash handling and speeding up boarding, but the augmentation is partial and depends on passenger adoption and infrastructure that not all systems have deployed. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI/automation reduces the driver's need to handle fares directly but offers limited direct productivity augmentation to the driver's core task of driving and passenger management. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI systems cannot reliably perform end-to-end fare collection, which requires physical cash/ticket handling, dynamic passenger interaction, and real-time judgment about fare types and exceptions. Limited automation potential exists without significant physical infrastructure changes. |
| Task automatability | claude-sonnet-5 | 2/5 | Fare collection has been partly automated via card readers, mobile payment, and automated fare gates, but the physical driving task and human presence remain, limiting full end-to-end automation of this specific sub-task by 'AI' per se (much of this is hardware/payment automation, not AI-driven judgment).</br>Overall time savings are moderate and often already realized via non-AI automation.4Not primarily an AI automatability story.rating2rationale). |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: transit agencies operate under public service contracts, regulatory frameworks often require visible human staff for passenger safety and assistance, and liability concerns around unattended fare collection create organizational and legal friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement ties fare collection to the driver; existing automated fare systems already bypass drivers in many agencies, though cash-based riders and accessibility needs create some friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Contactless or automated payment systems have significant upfront capital costs (card readers, integration, maintenance) that typically exceed the loaded wage of a single bus driver performing this task, making widespread deployment economically marginal. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated fare systems have upfront infrastructure costs but lower marginal cost per transaction than driver-mediated cash handling, though implementation and maintenance costs keep the ratio moderate rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI systems today perform this task reliably in production transit settings. Contactless payment systems exist but require passenger cooperation and don't replace the driver's fare-collection role; fully autonomous collection remains research-stage. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Automated fare collection systems (smart cards, contactless payment, ticket kiosks) are deployed widely in transit systems today, reliably handling most fare transactions without driver involvement. |
Inspect vehicles and check gas, oil, and water levels prior to departure.
21CI 16–25 · exposure 20 · augmentation 38 · importance 4.5/5 · click for rater detail
Inspect vehicles and check gas, oil, and water levels prior to departure.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Transit and intercity bus operations remain relatively low-tech adoption environments; while fleet telematics are growing, pre-trip fluid inspection is still predominantly manual, with limited investment in robotics or autonomous verification systems in this sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Transit and transportation sectors have been slower to adopt AI-driven automation for physical inspection tasks compared to information-based industries, though some fleets use telematics dashboards.: |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist via sensor dashboards flagging low fluid alerts from vehicle CAN-bus data, but the task is simple and frequent enough that drivers gain minimal productivity benefit; augmentation potential is modest compared to more complex diagnostic or decision-making tasks. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Sensor dashboards, predictive maintenance alerts, and digital checklists can help drivers quickly identify issues and streamline the inspection process, improving efficiency without replacing the physical check.: |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI-powered vision systems could potentially identify some fluid levels via camera inspection, the task requires physical access to multiple check points (gas cap, oil dipstick, radiator), hands-on verification, and judgment about acceptable levels—capabilities that current autonomous systems lack at reliable scale without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a physical inspection task requiring presence at the vehicle to visually check fluid levels and physical condition; current AI cannot perform the physical checking, though sensors could report some data.: |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory and liability barriers are substantial: drivers remain legally responsible for vehicle safety before departure, and fleet operators face DOT/FMVSS compliance; mechanically certified personnel typically must sign off on safety inspections, limiting pure automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Pre-trip vehicle inspections are often legally mandated (e.g., FMCSA/DOT requirements) requiring a qualified driver to physically verify vehicle safety before departure, creating strong regulatory barriers.: |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The infrastructure cost to automate fluid checks (robotic arms, specialized sensors, integration) would far exceed the wage cost of a driver spending 5–10 minutes per shift on manual inspection, especially given low task frequency and high capital overhead. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Sensor-based monitoring systems have upfront hardware and integration costs; while cheap per-check once installed, they don't fully replace the human inspection required by regulation.: |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs end-to-end vehicle fluid inspection without human verification. While computer vision for damage detection exists, fluid-level checking requires robotic manipulation or human confirmation that isn't standard in production systems today. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some vehicles have telematics/sensor systems that report fluid levels and diagnostic codes, but full physical pre-trip inspections including visual checks are not replaced by deployed AI products.: |
Maintain cleanliness of bus or motor coach.
15CI 15–15 · exposure 0 · augmentation 0 · importance 4.0/5 · click for rater detail
Maintain cleanliness of bus or motor coach.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Transit and intercity bus operations are traditionally conservative, capital-constrained sectors with low digitization of support tasks; no measurable displacement of cleaning staff by automation is evident in the industry. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Transit and physical fleet maintenance is a low-digitization, low-AI-adoption sector; there is no evidence of AI/robotic cleaning solutions being adopted for buses at any pace. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | Current AI offers no meaningful assistance to human bus cleaners in performing this task; the work remains manual and unaugmented by available technologies. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance for the physical act of cleaning a bus interior or exterior; this remains a purely manual task with no meaningful software or AI augmentation pathway. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Bus and coach cleaning involves physical manipulation of diverse surfaces, debris removal, and judgment about cleanliness standards in unstructured environments. Current AI systems cannot perform this embodied, multi-step physical task end-to-end without human intervention. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical cleaning of a bus interior/exterior requires manual manipulation and dexterity that no current off-the-shelf AI system can perform end-to-end.6 This is a physical labor task, not a cognitive/information task, so AI language or vision models offer no direct automation path. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While no explicit licensing barrier exists, organizational friction is moderate: fleet operators have established cleaning contracts, staff routines, and preference for predictable human labor over experimental automation in a safety-sensitive environment. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human specifically clean the bus, but practical/organizational barriers (fleet maintenance depots, existing janitorial staff/contracts) create some friction against automation, mostly moot given lack of robotic solutions. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Autonomous cleaning robots capable of this work remain expensive to purchase, deploy, and maintain; contracted human cleaners cost far less than the capital and operational overhead of specialised hardware, making this economically unattractive. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI-driven robotic solution for this task at comparable cost; specialized cleaning robots (where they exist) are more expensive and less flexible than a human driver doing basic upkeep. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed autonomous systems reliably clean buses or coaches in production. Robotic cleaning in highly controlled environments exists, but no commercial product handles the variability, clutter, and surface types encountered in real transit operations. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs autonomous bus interior/exterior cleaning at any meaningful scale; robotic cleaning systems for vehicles remain research-stage or limited to narrow industrial contexts. |
Drive vehicles over specified routes or to specified destinations according to time schedules, complying with traffic regulations to ensure that passengers have a smooth and safe ride.
13CI 0–25 · exposure 13 · augmentation 38 · importance 4.7/5 · click for rater detail
Drive vehicles over specified routes or to specified destinations according to time schedules, complying with traffic regulations to ensure that passengers have a smooth and safe ride.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of autonomous buses is still in pilot stage in most jurisdictions; widespread production deployment remains rare, with most transit agencies retaining human drivers and proceeding cautiously on autonomous trials. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Public transit and intercity bus sectors are slow-moving, heavily regulated, and have seen negligible autonomous deployment at scale compared to information/finance sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist drivers through route optimization, real-time traffic alerts, collision avoidance warnings, and passenger information systems, improving safety and efficiency while the driver remains in control. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Some ADAS features (lane-keeping, collision warnings, route optimization software) assist drivers marginally, but they do not transform the core driving task's productivity. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While autonomous vehicles exist in research and limited deployment, current AI cannot reliably handle the full complexity of driving safely in diverse traffic, weather, and urban conditions at the 50% time-saving threshold in general use. The task also requires passenger interaction and safety judgment beyond current autonomous system capabilities. |
| Task automatability | claude-sonnet-5 | 1/5 | Full autonomous transit/intercity bus driving with passengers, in mixed traffic, at scale is not achievable end-to-end with off-the-shelf systems today; no general time-saving deployment exists for this exact task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory barriers are substantial—most jurisdictions require a licensed human driver or heavily restrict autonomous operation; liability frameworks favor human accountability; union and labor protections also create contractual friction against full substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Commercial driving requires licensure (CDL), passenger safety liability, and regulatory frameworks that mandate a qualified human operator; autonomous passenger transit faces heavy regulatory and legal barriers to replace this role. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current autonomous vehicle systems remain capital-intensive and operationally expensive when accounting for infrastructure, insurance, and fail-safe oversight, making them costlier than human drivers in most transit contexts today. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Current autonomous vehicle systems for buses require expensive sensor suites, mapping, monitoring staff, and regulatory compliance costs that exceed a bus driver's wage for equivalent service today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Autonomous vehicle products exist but remain in pilot phases with significant limitations; no mature production system reliably performs long-distance or city transit driving at scale without human oversight or frequent interventions. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | Autonomous buses exist only in limited pilot/geofenced trials (e.g., shuttle demonstrations) with safety drivers present; no product reliably performs full-route transit or intercity driving in production without human oversight. |
Advise passengers to be seated and orderly while on vehicles.
8CI 0–16 · exposure 8 · augmentation 25 · importance 4.2/5 · click for rater detail
Advise passengers to be seated and orderly while on vehicles.
8| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Transit sectors remain heavily reliant on human drivers for in-cabin oversight; despite autonomous vehicle research, on-vehicle behavioral management remains unautomated and is not a focus of current AI deployment in public transit. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Transit is a physical, unionized, safety-regulated sector with slow AI adoption for driver-facing interpersonal tasks; automation efforts focus on vehicle operation, not passenger management. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI systems could assist a driver by flagging passenger violations on a display or logging incidents for review, but current systems offer minimal productivity gain for this primarily interpersonal task. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Automated announcements or signage can supplement but do not meaningfully enhance a driver's ability to manage passenger behavior and order in real time. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time presence, interpersonal judgment, and authority to manage passenger behavior—capabilities that are fundamentally unavailable in current autonomous systems deployed on buses today. AI cannot reliably assess complex social dynamics, enforce compliance through presence, or handle the unpredictability of in-vehicle passenger management. |
| Task automatability | claude-sonnet-5 | 2/5 | Verbal passenger management requires real-time presence and physical authority on the vehicle; current AI cannot replicate a driver's in-person crowd control role.This is a minor communicative sub-task but not separable and automatable end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong regulatory and human-contact barriers exist: transit agencies maintain driver requirements for safety and liability, passengers expect human interaction and authority for behavioral guidance, and legal responsibility for passenger safety typically rests on a human operator. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Transit driving requires licensed, physically present personnel responsible for passenger safety and order; regulatory and safety requirements strongly favor human presence. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Deploying automated systems (cameras, speakers, oversight staff) to monitor and enforce seating/order would likely cost more per trip than the marginal cost of a driver already present on the vehicle. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this specific in-person task, so no meaningful cost comparison favors AI; the human driver already performs it as part of the job. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs passenger behavior management on transit vehicles. While computer vision systems exist, they cannot issue commands, interpret context, or manage the social/behavioral aspects central to the task. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product has drivers or robots verbally managing passenger behavior and seating on transit vehicles; this remains firmly a human function. |
Load and unload baggage in baggage compartments.
7CI 0–15 · exposure 0 · augmentation 13 · importance 3.7/5 · click for rater detail
Load and unload baggage in baggage compartments.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Transit agencies operate in capital-constrained, heavily unionized sectors with slow digitalization. Baggage handling automation adoption remains negligible in practice, with no measurable displacement in the industry. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Transit and ground transportation sectors show minimal automation of physical baggage handling; this is a low-digitization, physical-labor task with no evidence of robotic adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal augmentation for manual baggage loading; computer vision might marginally assist with inventory tracking, but current systems do not meaningfully enhance the core physical task itself. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no meaningful assistance to a bus driver in the physical act of loading and unloading baggage. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Loading and unloading baggage requires physical manipulation of variable-sized, fragile items in confined compartment spaces with dynamic stacking constraints. Current AI systems lack the embodied manipulation capability, sensorimotor coordination, and real-time spatial reasoning needed to perform this end-to-end without human intervention. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical loading and unloading of baggage into bus compartments requires manual dexterity and mobility that no deployed AI or robotic system currently performs in this context.rolling luggage of varying shapes/weights into tight compartments is a manipulation task beyond current robotics deployment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Physical safety requirements, liability for damaged cargo, passenger expectations of human service, and union labor agreements in transit collectively create strong structural barriers against automation. Human presence remains legally and operationally expected. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement specifically for baggage handling, but physical infrastructure and safety practicalities (uneven items, compartment access, liability for damage) create some friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Robotic systems capable of baggage handling remain capital-intensive and require significant infrastructure investment, maintenance, and oversight—making total cost per task far exceed the loaded wage of a bus driver or baggage handler. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | No viable AI/robotic solution exists for this physical task, so the comparison to human labor cost is moot—any hypothetical robotic solution would be far more expensive than a driver doing it manually. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system reliably performs baggage loading/unloading in production transit operations today. While robotics research exists for manipulation tasks, no mature product is in operational use at transit agencies for this purpose. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | There are no commercial products deployed to load/unload passenger baggage on buses; this remains firmly a human physical labor task. |
Park vehicles at loading areas so that passengers can board.
4CI 0–9 · exposure 8 · augmentation 25 · importance 4.5/5 · click for rater detail
Park vehicles at loading areas so that passengers can board.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Transit agencies have shown minimal real-world deployment of autonomous buses for passenger boarding, with most pilots remaining experimental; adoption velocity in this sector remains very slow. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Transit and intercity bus sectors have very limited automation adoption for physical driving maneuvers; this remains a slow-moving, capital-intensive, heavily regulated industry. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could offer minor assistance through collision-warning systems or automated guidance to the parking spot, but the driver must retain full control; the augmentation potential is limited by safety-critical nature and regulatory constraints. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Some driver-assist technologies (parking sensors, cameras, lane-keeping) offer minor assistance, but they do not substantially transform the productivity of this specific docking/parking action. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Parking at designated loading areas requires precise spatial maneuvering, real-time environmental perception, and safe interaction with pedestrians—tasks where current autonomous vehicle systems remain immature and unreliable in unstructured real-world conditions. |
| Task automatability | claude-sonnet-5 | 1/5 | Precision parking of large passenger vehicles at loading zones amid pedestrians, traffic, and variable curb conditions requires real-world physical maneuvering that current AI systems cannot perform end-to-end without embodiment in a fully autonomous vehicle. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Transit operations are heavily regulated; driver licensing, liability frameworks, passenger safety requirements, and insurance mandates that a qualified human operator control the vehicle during loading, creating hard legal and regulatory barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Commercial driving requires licensed, certified human drivers, DOT/transit regulations, insurance liability, and passenger safety oversight, making autonomous substitution for this specific action highly regulated and restricted. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current autonomous vehicle systems suitable for this task remain expensive to acquire, maintain, and insure; the total cost per operation far exceeds a bus driver's loaded hourly wage. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Autonomous driving hardware, sensors, safety redundancy, and remote oversight for this task currently cost far more than a driver's marginal time performing this action as part of their normal route duties. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | While autonomous parking exists in controlled settings, no deployed product reliably performs safe passenger-loading-area parking across typical transit environments with the safety margins required for frequent daily operations. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No widely deployed transit or intercity bus product autonomously parks buses at loading areas in commercial service; autonomous bus pilots remain limited, geofenced, and non-scaled. |
Assist passengers, such as elderly or individuals with disabilities, on and off bus, ensure they are seated properly, help carry baggage, and answer questions about bus schedules or routes.
3CI 0–5 · exposure 0 · augmentation 25 · importance 4.3/5 · click for rater detail
Assist passengers, such as elderly or individuals with disabilities, on and off bus, ensure they are seated properly, help carry baggage, and answer questions about bus schedules or routes.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Transit agencies are slow to adopt automation; driver labor is heavily unionized; accessibility services remain human-centric due to regulatory requirements and public expectation. No evidence of material AI/robotic adoption for passenger assistance in this sector. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Public transit is a low-digitization, physical-labor sector with minimal AI-driven displacement of driver assistance duties. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist with schedule/route questions (chatbots or in-vehicle systems), but cannot augment the critical physical assistance component that defines this task. Limited augmentation value for the core human-contact and safety-critical elements. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help answer scheduling/route questions via apps or voice assistants, offering modest assistance, but cannot aid the physical passenger-assistance components. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical assistance (helping elderly/disabled passengers on/off the bus, carrying baggage) and real-time interpersonal interaction that current robots cannot reliably perform in uncontrolled, dynamic transit environments. No current AI system can autonomously provide safe, dignified physical support or adaptive assistance to diverse passengers. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation, mobility assistance, and real-world judgment about passenger safety that current AI systems cannot perform without embodiment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and safety barriers exist: transit agencies have duty-of-care obligations toward vulnerable passengers; physical assistance requires human licensure/certification; liability is asymmetric (injuries to elderly/disabled passengers carry significant legal exposure); many jurisdictions mandate human attendants on accessible transit. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Safety liability, ADA-related accessibility obligations, and the need for physical human contact create strong practical barriers to automation of the physical-assistance parts. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Humanoid robots or AI-driven systems capable of safe physical assistance are prohibitively expensive (hundreds of thousands to millions per unit) compared to the loaded hourly wage of a bus driver, with high maintenance and liability overhead. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for the physical assistance component, so any AI cost is irrelevant relative to the human wage for full task completion. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI or robotic system reliably performs physical passenger assistance and personalized interpersonal support in transit environments at production scale. Research prototypes exist but are far from reliable real-world deployment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product physically assists passengers on/off buses or handles baggage; this remains purely a human physical-care task. |
Handle passenger emergencies or disruptions.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail
Handle passenger emergencies or disruptions.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | The transit industry has not adopted AI for emergency or disruption handling; buses require human drivers who remain legally responsible for passenger safety. No meaningful automation or AI deployment in this domain is occurring. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Transit driving is a physical, safety-critical, heavily regulated sector with minimal AI adoption for real-time emergency response tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist with limited tasks like alerting authorities or logging incidents, but current systems offer minimal real-time support for the core activities of de-escalation, medical response, or safety assessment during emergencies. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could provide minor support such as alerting dispatch or providing emergency protocol reminders, but it offers little direct assistance during the actual event as it unfolds. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Handling passenger emergencies or disruptions requires real-time judgment, physical intervention, de-escalation, and situational awareness that current AI cannot replicate end-to-end. This task involves unpredictable human behavior and safety-critical decisions that demand human presence and authority. |
| Task automatability | claude-sonnet-5 | 1/5 | Handling unpredictable in-cabin emergencies (medical events, altercations, security threats) requires real-time physical intervention, judgment, and communication that no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong legal and regulatory barriers exist: bus drivers are required by law and regulation to be physically present and responsible for passenger safety. Liability for emergencies rests with a licensed, accountable human operator, and no AI system can assume this legal duty. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Safety regulations, liability, and legal requirements mandate a trained, licensed human driver present to manage emergencies, and physical intervention capability is a hard requirement AI cannot meet. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot yet perform this task at all, making a cost-benefit comparison infeasible. A human driver is currently the only viable option for emergency handling, and no AI alternative exists to reduce costs. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI alternative to compare costs against; a human driver/attendant is required, making AI substitution cost irrelevant or effectively infinite. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can autonomously manage passenger emergencies or disruptions on a bus. Current AI cannot assess crises in real time, make judgment calls on safety, or provide the authoritative human presence required to de-escalate or manage passengers. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product manages passenger emergencies or disruptions on transit vehicles; this remains entirely a human responsibility with no AI substitute in production. |
Related occupations — Transportation & Material Moving
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.