Skip to main content
CSP LabBy Educator Today

Pilot · L11 · 60 minutes

Who benefits from an innovation

Use two fictional evidence packs to identify benefits and unintended harms for specific stakeholders, distinguish device access from effective participation, trace bias to training data and design rules, evaluate crowdsourced contributions, and propose a mitigation while stating what it leaves unsolved.

Topics: 5.1 Beneficial and Harmful Effects · 5.2 Digital Divide · 5.3 Computing Bias · 5.4 Crowdsourcing

Goals and vocabulary

By the end you can:

  • Identify a benefit and an unintended harm of one innovation for named stakeholders.
  • Distinguish owning a device from being able to participate effectively.
  • Explain how training data and design rules can introduce bias.
  • Evaluate how crowdsourcing provides local knowledge and how its reports can be validated.
  • Propose a mitigation and state a limitation that remains.
Stakeholder
A person or group affected by an innovation.
Unintended harm
A negative effect the designers did not plan.
Digital divide
Unequal access to computing devices, connections, skills and their benefits.
Computing bias
Systematic unfairness that a program’s data or design introduces or repeats.
Training data
Past examples a system learns its patterns from.
Crowdsourcing
Gathering contributions, such as reports or data, from many people.
Mitigation
A change that reduces a harm, usually without removing it completely.

Benefits and harms depend on who you ask

A computing innovation can help one group while harming another, often in ways its designers did not intend. To judge an innovation, name a specific stakeholder, say how they are affected, and give evidence.

One innovation, several stakeholders
InnovationStakeholderEffect
Online-only club sign-upStudents with home internetBenefit: sign up any time from home.
Online-only club sign-upStudents without home internetUnintended harm: miss sign-up and lose places.
Online-only club sign-upClub leadersBenefit: instant lists; harm: smaller, less varied membership.

Access is more than owning a device

The digital divide is unequal access to computing and its benefits. Owning a device is only the start. Effective participation also needs a reliable connection, time, skills, content in a language you read well, and an interface that works with your needs.

Barriers to participation
BarrierExample
ConnectionA family phone with limited data shared by four people.
Skills and confidenceNever having filled in an online form before.
LanguageInstructions only in English for families who mainly read Spanish.
TimingA form open only in the evening when a student works.

A targeted intervention removes a named barrier for a named group. Giving everyone a new laptop does not help a family whose barrier is language.

Evidence pack 1: ClubConnect and bias

Worked example

Where bias enters
SourceEvidenceEffect
Training dataEarly-bus students rarely appear in past sign-ups.Their interests are underrepresented, so recommendations fit them poorly.
Design ruleRank by members who log in three times a week.Students who share a device or have limited data log in less, so their clubs rank lower.
FeedbackOnly the top five are shown.Popular clubs get more sign-ups, which makes them rank even higher.

Collecting three more years of data the same way would not fix the problem: the new data would underrepresent the same students. The fix has to change how data is collected or how clubs are ranked.

Evidence pack 2: AccessMap and crowdsourcing

Worked example

Crowdsourcing gathers contributions from many people. Residents know things no central office can see quickly, such as a lift that broke this morning. But participation is uneven, and without checks, wrong reports spread as fast as right ones.

Strengths and weaknesses
StrengthWeakness
Fast, local knowledge from people who use the routes.Areas with fewer contributors have fewer, staler reports.
Many observers can catch problems an office would miss.False or outdated reports can mislead until someone checks them.

Mitigation and what remains

A strong response names the harm, proposes a change that targets its cause, and admits what the change does not fix.

Mitigations and remaining limitations
InnovationMitigationRemaining limitation
ClubConnectAdd a paper and lunchtime sign-up, and rank clubs by sign-ups rather than login frequency.Past data still underrepresents early-bus students until new data builds up.
AccessMapRequire independent, current corroboration and moderator review before marking a report “confirmed”; recruit volunteers outside the centre.Confirmation slows urgent reports, and some areas may still have few contributors.

Explore each concept

Choose a concept, change its inputs and follow the explanation. Each example models the stated rule or scenario; completing it is practice rather than evidence of mastery.

  • Physical, software and conceptual innovations
  • Benefits, harms and scale
  • Creative expression and cross-field effects
Open this example on its own

Change it · step through · explain

Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.

Ready · Make a prediction

Ready to explore: Intended benefits and unintended reuse
  1. Innovationmedical-image-toolWaiting
  2. UseintendedWaiting

Ready to explore: Intended benefits and unintended reuse

Why: Change an input, predict the first effect, then advance one step.

Try this: Change one input or choose another listed case. Predict the effect, reset and compare the states and governing rule.

Step through and experiment

Change it · step through · explain

Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.

Ready · Make a prediction

Ready to experiment
  1. Input 1online-onlyWaiting

Ready to experiment

Why: Predict the first change, then advance one step.

Try this: Compare all three policies. Which barrier does each change, and which remains?

Step through and experiment

Change it · step through · explain

Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.

Ready · Make a prediction

Ready to experiment
  1. Input 115Waiting
  2. Input 2deviceWaiting

Ready to experiment

Why: Predict the first change, then advance one step.

Try this: Try a budget one below the cost, then exactly equal to the cost.

Step through and experiment

Change it · step through · explain

Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.

Ready · Make a prediction

Ready to experiment
  1. Input 120,80Waiting
  2. Input 28,4Waiting

Ready to experiment

Why: Predict the first change, then advance one step.

Try this: Change group sizes while keeping very different error rates. Does a high overall accuracy hide a gap?

Step through and experiment

Change it · step through · explain

Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.

Ready · Make a prediction

Ready to experiment
  1. Input 18Waiting
  2. Input 22Waiting
  3. Input 3same-sourceWaiting

Ready to experiment

Why: Predict the first change, then advance one step.

Try this: Keep the counts fixed and change provenance. Then set both counts to zero.

Common misconceptions

  • A device solves access

    The belief: Once everyone has a device, the digital divide is closed.

    What is actually true: Connection, skills, language and timing can still block participation.

    Check yourself: What else does a student need to use the device for sign-up?

  • Computers are neutral

    The belief: A program cannot be biased because it just follows rules.

    What is actually true: Its rules and training data come from people and past records, which can be unfair.

    Check yourself: Who is missing from the data it learned from?

  • More data fixes bias

    The belief: Collecting more data will always remove bias.

    What is actually true: More data collected the same way repeats the same gaps.

    Check yourself: Would the new data include the early-bus students?

Guided practice

Guided: trace the effects of ClubConnect

Read evidence pack 1 and use the “where bias enters” table.

  1. Name one stakeholder who benefits from ClubConnect and one who is harmed, with evidence from the pack.
  2. Explain how the login-frequency rule disadvantages students who share a device.
  3. Explain why three more years of the same evening online data would not fix the problem.
  4. Propose one mitigation and state one limitation it leaves.

Independent practice

Independent: evaluate AccessMap

Read evidence pack 2.

  1. Give one benefit of AccessMap for wheelchair users, using the station-lift evidence.
  2. Explain why reports are uneven across the town and who is affected.
  3. Propose a validation rule for reports and explain one cost of that rule.
  4. Choose one barrier to participation and design a targeted intervention for a named group.

Exit check

  1. Name one stakeholder harmed by an online-only sign-up and explain why owning a phone may not be enough.
  2. Give one way bias can enter a recommendation system that does not come from the training data.

Check yourself

Practice only: answers are checked in your browser and nothing is saved or sent to your teacher. Each explanation links back to the section it practises.

Lesson practice

  1. Question 1 of 6An online-only club sign-up opens at 7 pm and automatically gives club leaders an instant, sorted membership list. Which statement describes a benefit for a specific stakeholder?

    Select one answer.

  2. Question 2 of 6An online-only club sign-up has limited places allocated first come, first served. Places fill between 7 and 8 pm, when some students must work and cannot access the form. Which is an unintended harm?

    Select one answer.

  3. Question 3 of 6Which pair correctly describes a benefit and a harm of the same innovation, AccessMap?

    Select one answer.

  4. Question 4 of 6Every student in a fictional class now has a school tablet. Which situation shows that access to a device is not the same as effective participation?

    Select one answer.

  5. Question 5 of 6A fictional health service publishes appointment booking online only in English. Which barrier does this mainly create?

    Select one answer.

  6. Question 6 of 6Which change most directly targets the barriers described?

    Select one answer.

    Read this first

    Fictional situation. In a district, 30% of families report no home internet, and 20% mainly read Spanish. ClubConnect sign-up is online only, in English, from 7 pm. The school is choosing one change for next term.

Your practice so far

Check an answer to see its rationale.

Apply it somewhere new

  1. Question 1 of 6Which evidence from the pack shows underrepresentation in the training data?

    Select one answer.

    Read this first

    Fictional evidence pack — ClubConnect. A school launches ClubConnect, an app that recommends after-school clubs. It learns from three years of past sign-ups, which were collected only through an online form open in the evenings. Students who ride the early rural bus rarely appear in those records. The app ranks clubs by how many past members logged in at least three times a week, and shows only the top five.
  2. Question 2 of 6Which part of ClubConnect introduces bias through a design rule rather than through the data?

    Select one answer.

    Read this first

    Fictional evidence pack — ClubConnect. A school launches ClubConnect, an app that recommends after-school clubs. It learns from three years of past sign-ups, which were collected only through an online form open in the evenings. Students who ride the early rural bus rarely appear in those records. The app ranks clubs by how many past members logged in at least three times a week, and shows only the top five.
  3. Question 3 of 6The school proposes training ClubConnect on three more years of sign-ups, collected the same way. Why would this not fix the bias?

    Select one answer.

    Read this first

    Fictional evidence pack — ClubConnect. A school launches ClubConnect, an app that recommends after-school clubs. It learns from three years of past sign-ups, which were collected only through an online form open in the evenings. Students who ride the early rural bus rarely appear in those records. The app ranks clubs by how many past members logged in at least three times a week, and shows only the top five.
  4. Question 4 of 6How does the station-lift report show the value of crowdsourced contributions?

    Select one answer.

    Read this first

    Fictional evidence pack — AccessMap. A town launches AccessMap, a shared map where residents report step-free routes, broken lifts and blocked pavements. Anyone with a smartphone can add a report. After one month, 80% of reports come from the town centre, where most volunteers live. A wheelchair user’s report of a broken lift at the station was shown within minutes; a false report of a closed road stayed up for a week.
  5. Question 5 of 6Which validation process best addresses the false road report while keeping the map useful?

    Select one answer.

    Read this first

    Fictional evidence pack — AccessMap. A town launches AccessMap, a shared map where residents report step-free routes, broken lifts and blocked pavements. Anyone with a smartphone can add a report. After one month, 80% of reports come from the town centre, where most volunteers live. A wheelchair user’s report of a broken lift at the station was shown within minutes; a false report of a closed road stayed up for a week.
  6. Question 6 of 6Why does AccessMap give a less complete picture of areas outside the town centre?

    Select one answer.

    Read this first

    Fictional evidence pack — AccessMap. A town launches AccessMap, a shared map where residents report step-free routes, broken lifts and blocked pavements. Anyone with a smartphone can add a report. After one month, 80% of reports come from the town centre, where most volunteers live. A wheelchair user’s report of a broken lift at the station was shown within minutes; a false report of a closed road stayed up for a week.

Your practice so far

Check an answer to see its rationale.