Taxi Drivers
53-3054.00Median wage $42,100/yr41,050 employed (US)Rank #388 of 923 scored · top 42% by substitution
Drive a motor vehicle to transport passengers on an unplanned basis and charge a fare, usually based on a meter.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution30
Exposure25
Augmentation33
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
16 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
6%
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%26
panel mean rating 2.0/5 → substitution pressure 26/100
Technical feasibility todayw 20%24
panel mean rating 2.0/5 → substitution pressure 24/100
Cost vs. human wagew 15%27
panel mean rating 2.1/5 → substitution pressure 27/100
Adoption barriersw 20%inverted — strong barriers lower the score49
panel mean rating 3.0/5 (barrier strength) → substitution pressure 49/100
Sector adoption velocityw 10%23
panel mean rating 1.9/5 → substitution pressure 23/100
Task breakdown (16 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.
Provide passengers with information or advice about the local area, points of interest, hotels, or restaurants.
74CI 65–84 · exposure 62 · augmentation 63 · click for rater detail
Provide passengers with information or advice about the local area, points of interest, hotels, or restaurants.
74| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Major rideshare platforms (Uber, Lyft) and vehicle manufacturers are rapidly integrating AI-powered local information and recommendation systems into passenger interfaces. Early and accelerating deployment in digitized transportation and hospitality sectors. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Consumers already widely use smartphone maps, review apps, and AI chat assistants for local recommendations, reflecting fast, deep adoption of this information task outside the taxi industry itself. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI dramatically augments taxi drivers' ability to provide current, accurate, comprehensive local information in real time, enabling drivers to serve passengers better without requiring extensive local knowledge updates or manual research. |
| Augmentation potential | claude-sonnet-5 | 2/5 | While passengers can augment their own experience via personal apps, this does not meaningfully improve the driver's own performance of this incidental task. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems (LLMs, mapping APIs, review aggregators) can reliably provide information about local points of interest, hotels, restaurants, and directions with minimal human intervention. Current systems can answer 80%+ of common queries about local areas and recommendations at substantially lower time cost than a human driver would require. |
| Task automatability | claude-sonnet-5 | 3/5 | Providing local recommendations is information retrieval/conversational, something AI assistants (voice or app-based) already do well, though the in-person, contextual delivery during a ride is not fully replicated by standalone AI today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal legal or regulatory barriers exist; passengers may prefer human interaction and local knowledge, but nothing legally requires a human to provide this information. Most friction is social preference rather than hard authorization or liability requirements. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory barrier restricts passengers from using AI apps for local recommendations instead of asking the driver. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI inference costs for providing local information and recommendations are negligible (fractions of a cent per query), while a human driver's loaded wage for the same advisory time is $25–50+ per hour. The cost ratio heavily favors automation. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Query-based local information lookup via AI/search apps costs fractions of a cent compared to any human time allocated to this sub-task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (Google Maps, OpenAI plugins, hospitality recommendation systems, voice assistants in vehicles) already perform this task reliably in production. Rideshare platforms increasingly integrate AI-powered local information systems, though some edge cases and nuanced queries still benefit from human judgment. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Smartphone assistants, GPS apps, and chatbots already give reliable local recommendations at scale, but they are not integrated into taxi service delivery as a substitute for driver conversation. |
Determine fares based on trip distances and times, using taximeters and fee schedules, and announce fares to passengers.
70CI 40–100 · exposure 67 · augmentation 38 · click for rater detail
Determine fares based on trip distances and times, using taximeters and fee schedules, and announce fares to passengers.
70| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Ride-sharing platforms have adopted automated fare calculation widely, but traditional taxi fleets (medallion-based in many cities) have adopted more slowly due to regulatory conservatism and entrenched labor agreements. Adoption is uneven: digital-first services move faster, legacy taxi companies lag. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | This sub-task is already near-universally automated across taxi and rideshare markets globally, representing full, mature adoption rather than a pilot phase. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | GPS-integrated fare estimation and real-time fee-schedule lookup assist drivers in transparency and compliance, but the core task of reading a meter and announcing a fare is simple enough that augmentation provides modest productivity gains over the status quo. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Since the task is already automated by meters/apps, there's little remaining human task to 'augment'; the driver simply reads/announces the output. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Taximeter fare calculation is algorithmically straightforward and could be fully automated, but the task includes announcing fares to passengers, which involves real-time human interaction in a moving vehicle. AI cannot safely or reliably handle the full end-to-end workflow (navigating traffic, interacting with passengers, handling exceptions) at 50% time savings today. |
| Task automatability | claude-sonnet-5 | 5/5 | Calculating fares from distance/time using a taximeter and fee schedule is a deterministic computation already fully automated by standard taximeter hardware and ride-hailing apps. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Taxicab operations are heavily regulated by municipal authorities, which often mandate specific fare structures, driver licensing, and metering devices. Regulatory approval of AI-driven fare systems varies by jurisdiction, and some regions require human operator sign-off, creating material legal and compliance friction. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or human-judgment requirement blocks automated fare calculation; it's already standard practice with regulatory acceptance of certified meters. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated fare calculation (via app or integrated dispatch system) has near-zero marginal cost per trip compared to a human driver's loaded wage (~$15–25/hour). Integration and oversight are minimal once systems are deployed, making AI substantially cheaper for the pure fare-calculation component. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated fare calculation via meter/app is essentially free marginal cost compared to any human time spent calculating or announcing fares. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Ride-sharing apps (Uber, Lyft) have deployed automated fare calculation for years, but this operates in a narrower context than traditional taxi dispatch. Traditional taxi operations still rely heavily on human drivers using taximeters; no production system fully replaces the driver's role in determining and announcing fares in real-world taxi operations. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Taximeters and app-based fare calculators (Uber, Lyft, standard taxi meters) reliably compute and display fares in production at massive scale today. |
Communicate with dispatchers by radio, telephone, or computer to exchange information and receive requests for passenger service.
39CI 10–69 · exposure 34 · augmentation 50 · click for rater detail
Communicate with dispatchers by radio, telephone, or computer to exchange information and receive requests for passenger service.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi fleets are fragmented, small-scale, low-digitization sectors with strong unions and regulatory oversight. Even large platforms with sophisticated AI still route dispatch through app interfaces that drivers actively confirm rather than automated voice agents replacing radio communication. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Ride-hailing platforms have already displaced traditional radio dispatch across most urban markets globally, representing one of the fastest, deepest AI-driven adoption patterns in transportation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by transcribing dispatcher messages or suggesting route information, but the core task—receiving real-time vocal dispatch requests and confirming receipt—relies on the driver's active participation and does not benefit substantially from AI assistance today. |
| Augmentation potential | claude-sonnet-5 | 4/5 | GPS-based apps, automated matching, and real-time notifications substantially augment a driver's ability to receive and respond to service requests efficiently. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task is fundamentally about real-time two-way voice/radio communication with human dispatchers to coordinate passenger pickups. Current AI cannot reliably conduct spontaneous two-way radio conversations with human operators who may use colloquialisms, abbreviations, or informal speech in a live dispatch environment. |
| Task automatability | claude-sonnet-5 | 3/5 | The communication exchange itself (dispatch requests, status updates) can be automated via app-based dispatch systems, but this is bundled with the driving task and requires the human to physically respond and drive. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory requirement that a licensed, responsible human (the driver) receive and confirm dispatch instructions; liability if the AI misinterprets or loses communication; customer safety and service guarantees legally rest on authenticated human acknowledgment of assignments. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates human dispatch communication; the main friction is organizational (legacy taxi companies, driver habits) rather than legal or safety-critical. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The infrastructure and reliability requirements to replace human dispatcher communication (low-latency voice systems, fallback protocols, legal liability for missed rides) are expensive relative to the modest wage for this narrow subtask. Integration costs would exceed labor savings. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | App-based dispatch systems cost fractions of a cent per match compared to human dispatcher labor, though this doesn't cover the full taxi driver task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While AI can parse structured dispatch data and send basic text replies, no deployed product reliably handles the full back-and-forth vocal or real-time messaging exchange that characterizes dispatcher communication in actual taxi operations. Some rideshare apps automate order receipt, but human drivers still communicate directly with human dispatchers in most taxi fleets. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Ride-hailing apps (Uber, Lyft) already automate dispatch communication at scale, replacing radio/telephone dispatch with algorithmic matching in production for millions of trips daily. |
Perform routine vehicle maintenance, such as regulating tire pressure and adding gasoline, oil, and water.
39CI 15–64 · exposure 41 · augmentation 25 · click for rater detail
Perform routine vehicle maintenance, such as regulating tire pressure and adding gasoline, oil, and water.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi operators remain largely small, independent, or informal businesses with low digitization and capital constraints; adoption of automated maintenance infrastructure is nascent and concentrated only in large fleet operators in advanced markets. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Taxi driving and vehicle upkeep sit in a low-digitization, physical-labor sector with minimal AI adoption for manual maintenance tasks specifically. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI/robotic systems offer limited augmentation for humans performing these routine tasks, since the bottleneck is physical labor and mechanical precision rather than decision-making or information synthesis where AI assistance would meaningfully boost productivity. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could provide reminders or diagnostic alerts (e.g., tire pressure sensors, maintenance apps) but does not meaningfully assist in the physical execution of the task. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Routine vehicle maintenance tasks like tire pressure adjustment, refueling, and fluid top-ups are mechanical and well-defined procedures that modern autonomous vehicle systems can execute end-to-end with specialized hardware (robotic arms, sensors) in controlled environments, easily meeting the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manual task requiring hands-on manipulation of vehicle components; no current AI system can perform physical maintenance actions like checking tire pressure or adding fluids. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Barriers are modest: no strict licensing requirement for routine maintenance itself, though liability concerns exist around automated systems and some jurisdictions may require human verification; organizational friction is low since this is already outsourced by many operators. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barrier prevents automation of vehicle maintenance itself, but the physical nature of the task (not a software/cognitive task) creates a structural barrier to AI substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current robotic maintenance systems require significant capital investment, specialized infrastructure, and integration costs that currently exceed the cost of human maintenance labor, particularly for small taxi operators with modest fleets. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI system that can substitute for this physical task, so cost comparison favors the human doing it directly since AI cannot perform it at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Robotic fueling and tire-pressure systems exist in research/limited deployment (e.g., automated fueling stations, tire-service robots), but reliable end-to-end autonomous maintenance across diverse vehicle types and conditions remains inconsistent in production settings outside specialized depots. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical vehicle maintenance tasks; this remains purely a human physical activity, not something even research-stage robotics has solved for general deployment. |
Collect fares or vouchers from passengers, and make change or issue receipts as necessary.
37CI 16–57 · exposure 22 · augmentation 50 · click for rater detail
Collect fares or vouchers from passengers, and make change or issue receipts as necessary.
37| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Taxi services show slow adoption of fully automated payment systems; ride-sharing apps have deployed digital payments more rapidly, but traditional taxi operators lag due to infrastructure limitations, cost, and workforce resistance in small-operator segments. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Ride-hailing and modernized taxi fleets have rapidly adopted cashless, app-based payment systems, though traditional independent taxi operators lag. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Digital payment terminals, card readers, and receipt printers do assist drivers by reducing manual cash handling and paperwork, but the augmentation is modest—these are mature, widely-adopted technologies that drivers already use rather than transformative AI-driven assistance. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Payment apps and point-of-sale devices assist drivers by automating calculations and receipt generation, reducing errors and speeding transactions, though drivers still handle exceptions and cash. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Collecting cash/card payments and issuing receipts can be partially automated (card readers, receipt printers), but the task also requires human judgment for dispute resolution, handling special requests, and managing passenger interaction. Current AI cannot reliably handle the full end-to-end task with 50% time saving at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Digital payment apps and integrated ride-hailing platforms have automated much of fare collection, but cash handling, receipt issuance in traditional taxis, and voucher processing still require human involvement in many contexts, especially outside app-based systems. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: payment handling involves liability and fraud concerns, regulatory compliance for receipts and tax documentation, customer preference for transparency in transactions, and the requirement for a human (the driver) to be present for legal and safety accountability in the ride. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing or legal requirement mandates human fare collection; payment automation is already normalized and unregulated as a barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Mobile payment processors and digital receipt systems have costs comparable to or slightly higher than manual cash/receipt handling when integration and oversight are factored in. Automated payment infrastructure is expensive relative to simple human-operated alternatives in low-margin taxi operations. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated payment processing (card readers, apps) is very cheap per transaction relative to driver time spent handling cash and making change, though integration costs exist for legacy fleets. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Mobile payment systems and digital receipts exist and are deployed, but taxi drivers still require human oversight for payment disputes, manual cash handling, and customer service. No AI system today reliably performs the complete payment collection task without significant human supervision. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Payment terminals, mobile apps, and automated fare systems are deployed at scale in ride-hailing and many taxi fleets, but traditional taxis and cash transactions in many markets still rely on driver-managed collection. |
Notify dispatchers or company mechanics of vehicle problems.
33CI 13–52 · exposure 25 · augmentation 50 · click for rater detail
Notify dispatchers or company mechanics of vehicle problems.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Taxi and ride-hailing fleets are adopting telematics and automated diagnostics, but these typically supplement rather than replace driver notification; full automation of the notification workflow remains uncommon in production systems. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Taxi and rideshare fleets are a low-to-moderate digitization sector; large rideshare companies use some vehicle diagnostics but traditional taxi operations still rely heavily on manual driver reporting. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered vehicle diagnostics and alert systems can assist drivers by flagging potential problems automatically, reducing the cognitive load of identifying and reporting issues, though the driver still controls the final notification decision. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Apps and diagnostic dashboards can help drivers flag and communicate issues faster and more accurately, but this is incremental assistance rather than a transformative capability. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time verbal or written communication to a human dispatcher or mechanic about vehicle-specific problems, which demands human judgment about severity, urgency, and context that current AI cannot autonomously execute end-to-end. |
| Task automatability | claude-sonnet-5 | 3/5 | The core action—reporting a mechanical issue—could be automated via vehicle telematics/sensor systems that auto-detect and transmit diagnostics, but human driver observation and communication of nuanced issues still adds value.rendere |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | High barriers exist: dispatchers and mechanics are human workers who receive and act on notifications, creating organizational and operational dependencies; liability for missed critical vehicle problems rests on the reporting chain; and fleet operators depend on human judgment about which problems warrant immediate attention. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing or legal requirement that a human personally report vehicle problems; automated systems can and do transmit this information without regulatory obstruction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current AI systems (diagnostics, notification templates) still require human oversight and decision-making, making the total cost of automation comparable to or higher than having a driver notify via voice or text. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Telematics hardware and automated alert software have upfront and maintenance costs comparable to or sometimes cheaper than driver time spent reporting, but retrofitting older taxi fleets adds cost, making the ratio only moderately favorable. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While diagnostic sensors and telematics can detect some vehicle problems automatically, the task specifically requires notification *by the driver to dispatchers/mechanics*, which still depends on human initiation and communication in deployed taxi systems. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Fleet telematics and OBD-based diagnostic alert systems exist and are deployed in some commercial fleets, but for typical taxi drivers this remains largely a manual verbal/radio/app notification process rather than a mature automated pipeline. |
Complete accident reports when necessary.
31CI 23–40 · exposure 33 · augmentation 50 · click for rater detail
Complete accident reports when necessary.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi and rideshare companies operate in traditional regulatory environments with slow digitization of accident workflows; adoption of AI accident-report automation is minimal and confined to pilot projects, not production deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Taxi driving is a low-digitization, physical-world occupation with minimal AI tool adoption for administrative subtasks like this. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could meaningfully assist by auto-populating driver and vehicle info, suggesting relevant insurance details, or drafting narrative sections from images and prompts, thereby reducing report-writing time while the driver retains responsibility for accuracy and liability. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI dictation and form-completion apps can meaningfully speed up writing an accident report, though the driver must still supply and verify the facts. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI could assist with data entry and document formatting, but cannot reliably assess accident liability, determine fault, or capture context-dependent details that insurance and legal systems require. End-to-end completion meeting the 50% time-saving threshold is not achievable without significant human review. |
| Task automatability | claude-sonnet-5 | 3/5 | Drafting an accident report from dictated facts is well within current AI capability (transcription, summarization, form-filling), though gathering the underlying facts, photos, and witness info still requires a human on-scene. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Insurance companies and legal jurisdictions typically require reports signed by the driver or witness; automated or AI-generated accident reports face liability and regulatory acceptance barriers. Human accountability for accuracy is legally expected. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Accident reports often require the driver's signed, first-person account for insurance and legal purposes, creating some requirement for human authorship even if AI drafts the text. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI infrastructure, human oversight to verify accuracy, and liability exposure for errors likely exceeds the wage cost of a driver spending 30–60 minutes completing a report, especially given low task frequency. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Basic transcription/report generation tools are cheap, but the overall cost is dominated by human involvement (statement gathering, insurance follow-up), keeping total cost comparable to doing it manually. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably complete accident reports independently. Vision systems can extract some scene details, but integrating liability assessment, regulatory compliance, and insurance-required statements into production systems is not standard practice in commercial products. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Voice-to-text and template-filling apps exist and could assist, but no widely deployed product specifically automates taxi accident report completion end-to-end today. |
Vacuum and clean interiors and wash and polish exteriors of automobiles.
29CI 19–40 · exposure 20 · augmentation 13 · click for rater detail
Vacuum and clean interiors and wash and polish exteriors of automobiles.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Taxi and rideshare fleets operate in diverse, cost-sensitive markets where custom robotic systems are rare. Few production deployments exist; most cleaning remains manual despite high labor volumes, indicating slow actual adoption rather than technological readiness. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Taxi driving and vehicle maintenance are low-digitization, physical-labor sectors with minimal AI/robotics adoption for cleaning tasks specifically.4 |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | Automated pressure washers and some vacuum assists exist, but most taxi cleaning remains manual labor with minimal AI augmentation. The task is routine and labor-intensive rather than knowledge-intensive, limiting meaningful AI productivity enhancement while a driver remains involved. |
| Augmentation potential | claude-sonnet-5 | 1/5 | Current AI tools offer no meaningful assistance for the physical act of vacuuming, washing, or polishing a vehicle.4 |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current robots can handle basic vacuuming and washing in controlled environments, but taxi interiors vary widely in configuration, fabric types, and debris. Full end-to-end cleaning with human-equivalent quality and speed across diverse vehicles remains infeasible without substantial human intervention. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a physical manual cleaning task requiring dexterity and adaptability to vehicle interiors/exteriors that current general-purpose robots cannot perform economically or reliably.4 |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No hard licensing barriers exist for automated cleaning, but taxi fleet operators face organizational friction around equipment reliability, liability for damage, and customer expectations for human-touched final quality. Some regulatory regimes may require human sign-off on passenger-facing cleanliness. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, regulatory, or liability barriers prevent automating or outsourcing vehicle cleaning; it's a purely physical, unregulated task.4 |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Robotic vacuum and washing systems remain capital-intensive with high maintenance costs. Per-vehicle cleaning cost via current robots exceeds a taxi driver's labor cost for the same task when amortized over typical fleet utilization. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Robotic solutions for full interior/exterior detailing don't exist at scale, so any hypothetical automation would require expensive specialized hardware far costlier than a human doing this quick manual task.4 |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Prototype robotic cleaners exist for specific surfaces (e.g., automated car wash tunnels), but none reliably handle the full interior-and-exterior job autonomously in production taxi fleets. Deployed systems require human oversight and struggle with variable geometry and debris types. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI or robotic product autonomously vacuums, washes, and polishes taxi vehicles in real-world operational settings; automated car washes exist but don't handle interior detailing or judgment-based cleaning.4 |
Pick up passengers at prearranged locations, at taxi stands, or by cruising streets in high-traffic areas.
28CI 25–30 · exposure 25 · augmentation 38 · click for rater detail
Pick up passengers at prearranged locations, at taxi stands, or by cruising streets in high-traffic areas.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Despite decades of development, autonomous taxi adoption remains concentrated in limited pilot cities and special zones. Mainstream fleet adoption is slow outside controlled environments; most taxi services remain human-driven, indicating laggard sectoral adoption relative to AI capability hype. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Adoption is real but narrow and slow, confined to a few pilot cities with gradual regulatory expansion rather than broad sector-wide deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | GPS navigation, traffic prediction, and ride-matching systems already assist human drivers in improving pickup efficiency and reducing time to passenger. AI-powered route optimization and demand forecasting provide measurable productivity gains for human taxi operators. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-assisted dispatch, navigation, and demand-prediction tools help drivers find passengers more efficiently, but the core physical pickup task itself isn't meaningfully augmented for human drivers. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Finding and navigating to pickup locations is automatable, but identifying and physically picking up the correct passenger reliably requires human judgment and real-world interaction that current AI cannot replicate end-to-end at the required quality and speed. The task involves route optimization (automatable) but passenger recognition and communication (not yet reliably automatable) are core. |
| Task automatability | claude-sonnet-5 | 2/5 | Autonomous vehicle robotaxis exist but remain geographically limited and require substantial infrastructure and remote oversight, so this cannot be automated end-to-end at scale today across the full occupation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Regulatory requirements for driverless operation, liability frameworks, and customer preference for human drivers in certain contexts create meaningful friction. However, these are evolving barriers rather than hard legal prohibitions in all jurisdictions, and pilot programs are advancing in some regions. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Heavy regulatory approval, permitting, insurance/liability frameworks, and municipal restrictions govern where and how driverless pickup can operate, creating strong barriers to broad substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current autonomous vehicle fleets are expensive to deploy and maintain, with high sensor and compute costs. Deployed driverless services remain significantly more costly per ride than human taxi drivers in most markets, offsetting any labor savings. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Current autonomous vehicle fleets require expensive sensor suites, mapping, remote monitoring, and safety operations, making per-ride costs still comparable to or higher than human drivers in most markets. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Autonomous vehicles can navigate to predetermined locations, but fully driverless taxi pickup in unstructured urban environments with passenger verification, communication, and dynamic route-finding remains in pilot stage with notable failure rates. No production system reliably handles the full end-to-end passenger pickup workflow at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Robotaxi services (Waymo, limited Cruise/Baidu deployments) operate commercially but only in a handful of geofenced cities with heavy support infrastructure, far from general reliability. |
Drive taxicabs or privately owned vehicles to transport passengers.
26CI 23–30 · exposure 25 · augmentation 13 · click for rater detail
Drive taxicabs or privately owned vehicles to transport passengers.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi driving is concentrated in small/medium firms, urban services, and traditional labor markets with slow digital adoption. Even in tech-forward cities, autonomous taxi penetration remains a tiny fraction of total taxi operations, and displacement has been minimal to date. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Ride-hailing and taxi sectors are adopting robotaxis slowly and only in a handful of pilot cities; most of the industry remains human-driven. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI offers minimal assistance to a taxi driver performing their core task of safe vehicle operation and passenger transport. Navigation systems pre-date modern AI; safety and route augmentation are incremental and do not materially transform driver productivity on the task itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Navigation apps and ride-hailing dispatch software assist drivers with routing and matching, but this offers modest rather than transformative productivity gains for the core driving task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While autonomous vehicle technology exists, current AI-driven taxi systems have not achieved the >50% time-saving threshold at equal quality in most real-world environments due to edge cases, safety concerns, weather handling, and passenger interaction complexity. Full end-to-end automation remains research/pilot-stage rather than production-ready. |
| Task automatability | claude-sonnet-5 | 2/5 | Autonomous driving in complex urban environments with fare handling and passenger interaction is not yet reliably automatable at scale; robotaxis exist only in limited geofenced pilots.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant regulatory, liability, and safety barriers protect taxi driving. Jurisdictions require human licensure and passenger-safety oversight; insurance and accident liability frameworks still favor human accountability. Regulatory frameworks are actively tightening around autonomous operation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Regulatory approval for autonomous vehicle operation is required in most jurisdictions, creating moderate barriers, though this is regulation of the automation itself rather than a licensing requirement for a human driver. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Autonomous vehicle infrastructure (sensors, compute, fleet management, insurance, liability) remains expensive relative to human driver wages in most markets. Capital and operational costs have not yet undercut human-driven taxi economics at scale. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Autonomous vehicle fleets require expensive sensor suites, mapping, remote monitoring and safety infrastructure that currently make them costly relative to a human driver's wage, though costs are falling in mature deployment cities. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Autonomous taxi services exist in narrow, controlled geographies (e.g., Waymo in select US cities, robotaxi pilots), but deployment is limited to specific routes, weather conditions, and controlled environments. These are not reliable, scalable production systems across general taxi-driving contexts today. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Waymo and a few others operate robotaxis commercially but only in specific cities and conditions, far from general taxi driving capability. |
Report to taxicab services or garages to receive vehicle assignments.
18CI 0–35 · exposure 13 · augmentation 38 · click for rater detail
Report to taxicab services or garages to receive vehicle assignments.
18| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | The taxi industry, particularly traditional cab services, operates in legacy, fragmented, and often low-digitization environments. Modern dispatch apps (e.g., Uber) do automate assignment, but traditional garages have not adopted AI at scale. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Taxi/ride-hail fleet management is only partially digitized in traditional taxi garages, with slower adoption compared to app-based ride-hail platforms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | Minimal augmentation opportunity; AI might help a driver check assignment status or vehicle status remotely before arriving, but the core task is inherently a synchronous, in-person coordination that offers little room for AI assistance to raise productivity materially. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Digital dispatch and fleet management systems already streamline assignment scheduling and communication, improving efficiency around this task even though the physical reporting step remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task is fundamentally about human-to-human communication and assignment dispatch, which requires real-time coordination with a garage or service. While AI could theoretically automate parts of assignment logic, the core action—a driver physically reporting and receiving a vehicle—cannot be automated by AI without human participation. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a brief physical/administrative check-in step tied to receiving a vehicle; current AI could handle scheduling/dispatch logic but the physical act of reporting and receiving a vehicle assignment still requires a human presence at a garage. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong human-contact requirement: a driver must physically appear and interact with garage staff or dispatch to receive vehicle assignments and verify vehicle condition. This is embedded in operational and insurance procedures. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement blocks automating the assignment process itself, though organizational habits and the need for physical vehicle transfer create mild friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | There is no meaningful cost advantage to automating this task with AI, since the task inherently requires a human driver to be present and communicative. Any AI system would add overhead rather than replace the driver's unavoidable participation. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Dispatch software is cheap, but it doesn't eliminate the need for the driver's time or a physical vehicle handoff process, so overall cost savings versus human coordination are modest. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No current AI system performs this task end-to-end. Dispatch systems exist, but they still require human drivers to acknowledge assignments and show up; AI cannot replace the driver's physical presence or verbal/in-person check-in with garage staff. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Fleet management software already automates the assignment logic, but the human driver still must physically show up and coordinate with staff, so no deployed product fully replaces this task end-to-end. |
Turn the taximeter on when passengers enter the cab, and turn it off when they reach the final destination.
18CI 9–26 · exposure 16 · augmentation 13 · click for rater detail
Turn the taximeter on when passengers enter the cab, and turn it off when they reach the final destination.
18| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi industries remain fragmented, with many small operators and low digitization outside major urban centers. Even in cities with modern fleets, regulatory conservatism and driver skepticism have resulted in minimal AI adoption for meter automation specifically. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Taxi/rideshare is a physical, moderately digitized sector; autonomous vehicle pilots exist in a few cities but broad displacement of human taxi driving remains slow and limited to specific markets. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could theoretically provide alerts reminding drivers to toggle the meter or flag missed meter events, offering minimal augmentation. However, the simplicity and near-zero cognitive load of the task itself limits the practical value of such assistance. |
| Augmentation potential | claude-sonnet-5 | 1/5 | This is a simple manual toggle action with no cognitive or drafting component that AI assistance could meaningfully enhance. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task involves physical manipulation of equipment and requires accurate detection of passenger entry/exit events. While computer vision could theoretically identify passengers entering/exiting, integrating this with physical taximeter activation in an uncontrolled cab environment presents significant reliability challenges that current AI falls short of automating end-to-end at 50% time savings. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a trivial physical action but requires a human physically operating a device inside a standard taxi; current AI has no way to perform this discrete manual step unless the vehicle itself is autonomous, which is a separate technology, not an AI 'task performance' issue for the driver role. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Taxi metering is heavily regulated in most jurisdictions, with legal requirements that drivers must manually activate regulated meters to ensure fare compliance and consumer protection. Regulators typically mandate human operator control over metering to prevent fraud and ensure accountability, creating hard legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing rule requires a human specifically to toggle a meter, but it is embedded in the broader taxi driving job which does have licensing and regulatory requirements around fare metering and consumer protection. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Installing and maintaining AI vision systems, sensors, and integration infrastructure to automate a task that takes seconds and costs pennies in human labor would be far more expensive than manual operation. The cost of hardware, integration, and oversight would exceed the minimal loaded wage cost of a driver toggling a meter. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute cost to compare since this action is bundled into simply being a driver in the vehicle; replacing it requires full vehicle autonomy, an entirely different and far more expensive capital investment than a human driver's marginal action. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed products reliably automate taximeter operation in production taxi fleets today. While research prototypes exist, commercial deployment of autonomous taxi meter management across diverse vehicle types and contexts remains absent from real-world operations. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs this specific micro-task for human-driven taxis; it only becomes moot in fully autonomous vehicle fleets, which are narrow, geofenced deployments, not general taxi service. |
Follow relevant safety regulations and state laws governing vehicle operation, and ensure that passengers follow safety regulations.
13CI 0–25 · exposure 13 · augmentation 38 · click for rater detail
Follow relevant safety regulations and state laws governing vehicle operation, and ensure that passengers follow safety regulations.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Taxi and rideshare operators have adopted telematics and monitoring tools slowly; most driver safety compliance is still manual. Large-scale AI-driven enforcement of passenger behavior remains rare in production, with adoption limited to pilots. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Taxi/ride-hail driving is a physical, low-digitization task with only nascent, geographically limited autonomous vehicle deployment; broad sector adoption remains minimal. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered dashcams, alerts for speeding or unsafe driving, and real-time compliance monitoring can assist drivers in awareness and safety compliance, but the task of ensuring passenger behavior still relies heavily on human communication and judgment. |
| Augmentation potential | claude-sonnet-5 | 2/5 | GPS navigation and dispatch apps assist route planning, but there is little AI assistance specifically for monitoring or enforcing passenger safety compliance. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI systems can monitor some safety compliance parameters (seatbelts, speed limits) and flag violations, but enforcing passenger safety behavior requires judgment, communication, and occasional physical intervention that current systems cannot perform end-to-end. No single AI intervention achieves 50% time savings at equal quality. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical vehicle control, situational judgment, and interpersonal enforcement of passenger behavior, none of which current AI systems can perform end-to-end in a taxi context. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Taxi regulations and safety laws typically require a licensed driver to be present, responsible, and capable of intervening in passenger safety—a hard legal requirement in most jurisdictions. Liability for safety violations also rests with the driver, creating a legal barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Vehicle operation is heavily regulated with licensing, insurance, and legal liability requirements; autonomous replacement requires extensive regulatory approval in each jurisdiction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Safety monitoring dashcams, telematics, and compliance systems add cost; the infrastructure needed to automate passenger-facing enforcement through AI would likely exceed savings from reduced driver attentiveness, and human drivers remain necessary for the full task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Full self-driving hardware/software stacks with safety redundancy are far more capital-intensive per ride than a human driver in most markets today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed products exist for monitoring vehicle telemetry and compliance logging, but no production system reliably ensures passenger compliance with safety regulations or makes real-time enforcement decisions. Current systems are narrow and require human oversight. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | Autonomous vehicle products exist only in limited geofenced pilots (e.g., Waymo) and do not generally handle passenger safety compliance enforcement as a taxi driver would; not deployed at scale for this task. |
Test vehicle equipment, such as lights, brakes, horns, or windshield wipers, to ensure proper operation.
12CI 5–19 · exposure 8 · augmentation 25 · click for rater detail
Test vehicle equipment, such as lights, brakes, horns, or windshield wipers, to ensure proper operation.
12| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi operations remain largely traditional and low-digitization; even larger fleet operators have not widely adopted automated vehicle inspection systems at scale, with adoption remaining in pilot or niche applications only. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Taxi/ride-hail is a physical, lower-digitization sector, and this specific physical inspection subtask sees essentially no AI adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | While diagnostic scanners can assist by reading electronic codes, they provide minimal augmentation for the core task of physically testing brakes, horns, lights, and wipers—human judgment and direct hands-on testing remain central and largely unenhanced by current AI. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Some vehicles include automated dashboard alerts or diagnostic apps that can flag issues, offering minor assistance, but this doesn't meaningfully transform the manual inspection process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While some vehicle diagnostics can be partially automated (e.g., electronic fault codes), the physical inspection and testing of lights, brakes, horns, and wipers requires hands-on interaction that current AI systems cannot perform autonomously. Visual inspection and tactile feedback remain necessary, and no single automated solution achieves 50% time savings at equal quality for the full task. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence, manual manipulation of vehicle controls, and sensory checks (visual, auditory) inside a real vehicle; no current AI system can perform this physical inspection end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal liability and insurance requirements typically mandate that vehicle operators themselves or certified mechanics verify safety equipment before operation; liability asymmetry is high if an automated system fails to detect a fault that causes an accident. Regulatory frameworks for vehicle safety place responsibility on the driver. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not licensed per se, safety regulations and liability for vehicle roadworthiness create strong incentive for human verification, and physical presence is inherently required. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of automated vehicle testing systems (hardware, integration, maintenance) exceeds the modest loaded wage of a taxi driver performing a quick pre-shift equipment check, especially when the human can accomplish it in minutes with no capital equipment. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for this physical checking task, so any AI-based approach would require additional sensors/hardware costing more than the negligible marginal time a human spends checking equipment. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs complete vehicle equipment testing autonomously. Some diagnostic tools exist for electronic systems, but they require human setup and interpretation, and physical testing of mechanical components remains entirely manual. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed consumer product performs pre-trip vehicle safety checks autonomously for taxi drivers; this remains a manual, physical task. |
Perform minor vehicle repairs, such as cleaning spark plugs, or take vehicles to mechanics for servicing.
10CI 5–15 · exposure 0 · augmentation 25 · click for rater detail
Perform minor vehicle repairs, such as cleaning spark plugs, or take vehicles to mechanics for servicing.
10| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi driving remains a labor sector with limited AI adoption for physical tasks. While some fleet telematics assist with diagnostics, actual repair automation is not occurring in production settings and the sector lags in digitization of mechanical work. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Vehicle maintenance and taxi driving are low-digitization, physical-world sectors with minimal AI/robotics adoption for hands-on repair tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could provide some assistance through diagnostic tools or maintenance scheduling reminders, but the physical repair work itself offers limited opportunity for meaningful human-AI collaboration on this task. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could provide diagnostic guidance or help locate/schedule mechanics via apps, but offers little assistance for the actual physical act of cleaning parts or performing repairs. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation of vehicle components and hands-on mechanical work that current AI systems cannot perform. While diagnosis could be partially automated, the actual repair execution and deciding when to take a vehicle to a mechanic depend on embodied mechanical skills and judgment that AI lacks. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manual task requiring hands-on manipulation of vehicle parts or driving to a mechanic; no current AI system can perform physical repairs or drive a vehicle to a shop autonomously in this context. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | There are significant barriers including safety and liability concerns with automated vehicle repairs, insurance and regulatory requirements around vehicle maintenance, and the fundamental need for a physically present human to manipulate the vehicle and assess conditions on-site. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for a driver to clean spark plugs, but the practical barrier is physical embodiment - AI has no physical presence to perform repairs, though this isn't a regulatory/liability barrier per se. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Even AI systems that could theoretically assist with diagnostics would require specialized hardware, integration, and oversight costs that far exceed the time saved on basic tasks like spark plug cleaning or deciding to visit a mechanic. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for physical repair work, so cost comparison favors the human doing it directly; deploying robotics for this would be far more expensive than a driver's own time or a mechanic's labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can physically perform minor vehicle repairs or make reliable field decisions about when vehicles need mechanic service. AI has no way to interact with physical vehicle components or assess mechanical condition through hands-on inspection. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical minor vehicle repairs or manages vehicle servicing logistics for taxi drivers; this remains purely a human physical task. |
Provide passengers with assistance entering and exiting vehicles, and help them with any luggage.
5CI 0–10 · exposure 0 · augmentation 0 · click for rater detail
Provide passengers with assistance entering and exiting vehicles, and help them with any luggage.
5| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Taxi services remain highly fragmented, labor-intensive, and physically dependent; no sector-wide shift toward automation of passenger assistance is evident, and physical robotics adoption in this context is minimal. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Physical, in-person service tasks in transportation are among the slowest sectors for AI/robotic adoption, with no meaningful displacement occurring for this specific task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI offers no meaningful assistance to a human driver providing passenger boarding help or luggage handling; the task is fundamentally physical and interpersonal, not information-mediated. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers no meaningful assistance to a human driver performing physical luggage handling or passenger assistance. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation in unstructured environments and direct human contact assistance, which current AI systems cannot perform. Robotics capable of reliably helping passengers enter/exit and handling luggage are not deployed at scale in taxi operations. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation of luggage and physical assistance to a person, which is beyond the capability of any current AI system without embodied robotics that don't exist in this context. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Passengers expect and legally require human assistance for accessibility; liability for passenger injury, luggage damage, and duty-of-care create hard barriers. The physical and personal-contact nature of the task creates strong organizational and regulatory friction against automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing barrier exists, but physical assistance to vulnerable passengers (elderly, disabled) creates liability and safety expectations that favor human presence. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Humanoid robots capable of passenger assistance are extremely expensive to acquire and maintain, far exceeding the loaded wage cost of a taxi driver providing this service. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI-based substitute for physical assistance, so any comparison favors the human doing the task at essentially the only available cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No commercial taxi service deploys robots or automated systems to assist passengers with boarding and luggage handling in production. This remains an entirely human-performed service in all deployed taxi operations. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product provides physical passenger and luggage assistance; autonomous vehicle pilots explicitly lack this capability and typically exclude assistance services. |
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.