Grades 6–8 how-do-ride-share-apps-worksurge-pricing-lesson

How Do Ride-Share Apps Work? A Grades 6-8 AI Lesson

Paper-collage illustration of a city street map with a glowing ride route, a car, and a price tag showing how ride-share apps work.

A student flags you down after class. She just noticed her ride home cost $14 — but her friend, who requested a Lyft from the same corner three minutes earlier, paid $8. Same destination, same evening, same block. She wants to know why. You want to give her a real answer. You don’t have a lesson for it.

That moment — a concrete, observable injustice produced by an invisible algorithm — is exactly the kind of opening that AI literacy is supposed to prepare students to navigate. The problem isn’t that the algorithm is a mystery. The problem is that nobody handed you the curriculum to explain it. Not your fault. This post exists because that gap is real, and this question is a genuinely teachable one.

TL;DR: Ride-share apps use three AI systems. An ETA predictor uses maps, live traffic, and millions of past trips to estimate your driver’s arrival — Uber’s is called DeepETA. A surge-pricing algorithm watches riders-versus-drivers in your area and raises the fare to draw more drivers in and thin out demand. A matching algorithm waits a few seconds, evaluates several nearby drivers at once, and picks the pairing that lowers everyone’s wait — not just yours. All three learn from real trips every day.

How do ride-share apps work? The three systems behind every trip

Paper-collage overview of the three AI systems behind a ride-share trip: a route line, a price tag, and paired car-and-rider pins.

Every trip involves three distinct AI systems running in the background before a driver even accepts your request.

  • ETA prediction — estimates how long until the driver arrives, then keeps recalculating as the car moves.
  • Surge pricing — watches the ratio of riders to available drivers in small map zones and raises the fare when demand exceeds supply.
  • Driver-rider matching — evaluates multiple driver-rider pairs in a brief batch window and picks the pairings that lower total wait across everyone in the area, not just the first person who tapped the app.

Each system is a separate machine-learning model. They run together, they share inputs, and their outputs interact — a high-surge zone also changes which drivers the matching algorithm has to work with. That’s the foundational concept to put on the board on day one: one app, three algorithms.

How does the app predict your arrival time?

The ETA estimate appears within a second of opening the app, but the computation behind it is more layered than a simple map lookup.

A routing engine first calculates a baseline: it finds the fastest path using current map data and live traffic feeds. That baseline is then adjusted by a machine-learning model — Uber calls theirs DeepETA — which corrects the estimate using signals like time of day, weather conditions, and how that specific route actually performed for thousands of past trips. If 6 p.m. traffic on that road historically runs 40 percent slower than the map data suggests, DeepETA already knows.

The app also runs “map matching” — a process that snaps the driver’s GPS coordinates to the actual road network rather than treating them as a raw dot on a grid. Without it, a driver sitting at a red light one lane over from the route would appear off-course. As the driver moves, the estimate updates continuously, incorporating new traffic data in real time.

For a companion look at how the same ETA and routing logic appears in navigation apps, the post how AI navigation apps predict traffic covers the underlying prediction architecture in more depth.

Why does the price keep changing? Surge pricing explained

Paper-collage of a hexagon city grid where some zones glow brighter, with upward price arrows illustrating surge pricing.

Surge pricing exists to solve a real logistics problem: when too many riders request trips at the same time, some of them would wait a very long time or get no match at all. The algorithm tries to rebalance by raising the price.

Here is the fare structure in plain terms: the final charge is (base fare + per-mile rate × distance + per-minute rate × time + booking fee) × the surge multiplier. A 2x surge on a $10 base trip doesn’t cost $12 — it costs closer to $20, because the multiplier applies to the whole variable portion. Students who have worked with ratios and proportional reasoning will recognize the structure immediately; the surge multiplier is the same operation as a scale factor applied to an entire expression, not just one term. If you want a resource that connects this math directly to data literacy, the MS AI Math Bundle ties proportional reasoning to real-world pricing models.

The geography of surge pricing is also worth teaching explicitly. Cities are divided into small map zones — Uber uses a hexagonal grid system — and surge is calculated per zone. Two riders standing half a block apart can see different prices because they fall into different hexagons. That’s not a glitch; it’s the system working as designed. Walking one block is sometimes a genuine financial decision.

How does the app choose your driver?

Paper-collage showing several available cars evaluated at once and matched to waiting riders by the ride-share matching algorithm.

The intuitive guess is that the app always assigns the nearest available driver. The actual algorithm is more deliberate than that.

Instead of immediately locking in the closest car, the matching system waits a brief window — a few seconds — and collects multiple driver-rider pairs in the same batch. It then solves the batch together, looking for the combination of assignments that minimizes total wait time across all riders in the area. A driver two minutes away from you but four minutes away from everyone else might be assigned to you. A driver three minutes away from you but thirty seconds away from someone with zero other options might go there first.

The algorithm also runs a safety filter: it blocks any pairing where the driver and rider have each given the other a 1-star rating in a prior trip. That one constraint is worth pausing on in class — it shows students that the algorithm reflects human choices about what the system should value, not just efficiency. Every algorithm embeds a decision about whose wait matters more.

The fairness question: when an algorithm sets your price

This is the section where the literacy gap becomes visible. The three systems described above are not neutral pipes — they produce outcomes that differ by person, location, and moment in ways that aren’t always transparent to the people paying.

A Consumer Reports investigation published in June 2026 tested 30 routes across multiple cities. Two people requesting the same ride within the same minute saw a median price difference of 42.4 percent. Every one of the 30 routes showed at least two distinct price tiers. (Consumer Reports, June 2026)

The fairness question isn’t limited to riders. An Oxford study published in June 2025, drawing on 1.5 million UK trips, found that after Uber introduced its 2023 dynamic-pricing change, driver median hourly pay fell from approximately £22 to approximately £19 in inflation-adjusted terms — while Uber’s effective commission rose from around 25 percent to around 29 percent, with some trips exceeding 50 percent. (TechXplore, June 2025)

Regulators have begun responding. Connecticut and Maryland moved to restrict certain personalized-pricing practices. The FTC opened a public comment period on personalized pricing in 2026. (Freshfields, 2026)

None of this means the algorithm is malicious. It means the algorithm was designed for specific goals — and that people affected by it have a right to understand the mechanism well enough to ask good questions. That’s the literacy gap. A student who understands surge pricing is not just better at timing her rides. She’s better equipped to read a news story, evaluate a regulation, or notice when she’s being treated differently. Nobody handed her teacher a lesson for it — but that’s what this post is for.

For the data-privacy dimension — what the app records about every trip, device, and location — the post how targeted ads work in middle school covers how platforms use behavioral data and what students should understand about who holds it.

Teach it: a no-device “Surge Auction” activity (about 15 minutes)

This activity needs no devices, no accounts, and no prep beyond printing two sets of index cards. A typical class works through all three rounds in under 20 minutes.

Setup (0–3 min)

Print or write 8 “rider” cards and 3 “driver” cards. Post a base fare of $8 on the board. Each rider card says: “You need a ride. How much will you pay?” Each driver card says: “You’ll take a trip if the fare is worth it.”

Round 1 (3–8 min)

Only 3 riders decide they need a trip. Three drivers are available. Supply meets demand — the price stays at $8. Record the multiplier: 1.0x.

Round 2 (8–12 min)

All 8 riders now want a trip. Three drivers are still available. Ask the class: if drivers won’t accept rides below a certain fare, what price will thin the rider pool down to three? Students negotiate aloud. Someone will suggest $12. Someone will hold out. Record the number the class lands on — that’s their surge multiplier. Write it as a decimal: $12 / $8 = 1.5x.

Reveal (12 min)

That rising number is exactly what the surge algorithm computes — automatically, per map hexagon, in under a second. The class just did manually what the algorithm does in milliseconds. The difference is that the algorithm runs for thousands of zones simultaneously.

Debrief (12–15 min)

One question, written before the discussion: Who does surge pricing help? Who does it hurt? Is a price a computer sets “fair”? Students write for 90 seconds, then share. The goal isn’t consensus — it’s articulation. A student who can name two stakeholders and one tradeoff has met the standard.

Standards crosswalk

Activity stepStandard
Explaining the 3 systems / “ride-share is really three algorithms”ISTE 1.5.d (algorithmic thinking)
The Surge Auction supply-vs-demand simulationAI4K12 Big Idea #2 (representation & reasoning)
“Learns from millions of trips” discussionAI4K12 Big Idea #3 (learning from data)
The fairness / who-does-it-help debateISTE 1.3.d (build knowledge by actively exploring real-world issues)
Discussing what data the app collects about ridersISTE 1.2.d (manage personal data / privacy)

For a written extension, CCSS.ELA-LITERACY.W.7.8 fits directly: students gather and assess information from multiple sources — the Consumer Reports investigation, the Oxford study, and a Lyft or Uber public blog post on how pricing works — and synthesize the differing perspectives into a short argumentative paragraph on whether algorithmic pricing should be regulated.

Standards source: ISTE Student Standards. ISTE is a registered trademark of the International Society for Technology in Education. These resources are not affiliated with or endorsed by ISTE.

The lesson in the question

When a student asks why her ride cost more, she’s doing something most adults don’t: noticing that an automated decision affected her and asking for an explanation. That instinct is exactly the one AI literacy is supposed to build — not fear of the algorithm, not uncritical acceptance, but the habit of asking: what is this system doing, and who designed it to do that?

Teachers weren’t given a curriculum for this. That’s the gap, not a personal failing. The good news is that ride-share pricing is one of the most teachable AI systems out there — every student has seen it, most have felt its effects, and the math sits right inside proportional-reasoning territory students already own.

If you want the full lesson sequence — from ETA prediction through data ethics and standards documentation — the How AI Works MEGA Bundle covers 16 machine learning and AI concept lessons for grades 6-12, ready to drop into any unit. For a shorter entry point scoped specifically to grades 6-8, the AI Everyday Life Lesson Bundle focuses on the three AI encounters students are most likely to name themselves: recommendations, pricing, and navigation.

This post was drafted with AI assistance and human-finalized.

Quick questions

Ride-share apps use AI algorithms that factor in your location, time of day, and local demand versus available drivers. A June 2026 Consumer Reports investigation found a median 42.4% price difference between the highest and lowest fare offered for the same route. Companies say differences reflect real-time marketplace conditions.

The app uses ETA prediction. A routing engine maps the fastest path, then a machine-learning model (Uber's is called DeepETA) corrects that estimate using time of day, current traffic, weather, and how that route performed on past trips. The estimate updates as the driver moves toward you.

Surge pricing multiplies the normal fare when riders outnumber available drivers in your area. The goal is to attract more drivers and thin out demand. Cities are split into small map zones, so surge is calculated per zone — walking a block can sometimes change your price.

Get the free AI-Proof Assignment Toolkit

10 ways to redesign any assignment so an AI chatbot structurally can't do it — plus a redesign worksheet, a 45-minute lesson, the “Spot AI Work” card, and parent templates. One email, all 5 pieces.

Straight to your inbox — plus a short, practical AI-teaching email most school days. No spam. Unsubscribe anytime.

Prefer the full breakdown? See everything inside the toolkit →