1 of 30

Ethics of �Human-centered �Product�Management

Wanheng Hu, Ph.D.

Embedded Ethics Fellow, Stanford

wanhenghu@stanford.edu

CS177, Fall 2025

2 of 30

2

Current Ethics Readings & Assignments

  • Working for Ethically Complicated Organizations
  • Fair Use or Piracy
  • The Internet‘s Original Sin
  • Disruptive Innovation

3 of 30

3

Managing Ethical Tensions

  • Personal values vs. organizational goals
  • Copyright/fair use vs. innovation and cost
  • Ad-based business models vs. privacy and autonomy
  • Tech progress vs. disruption/social consequences

4 of 30

4

Ethics of �Human-centered �Product�Management

5 of 30

5

Ethics of �Human-centered �Product�Management

6 of 30

Today’s Learning Goals for Ethics

  • Understand how ethics and values are built into products
  • Recognize design choices and their potential consequences
  • Exert your own agency in shaping technology/products for good

6

7 of 30

Today’s Learning Goals for Ethics

  • Two frameworks for thinking about product ethics
    • The Politics of Artifacts
    • The Social Construction of Technology (SCOT)

7

8 of 30

Do Products Have Politics or Ethics?

  • “Politics” → power relations & struggles for power
  • “Ethics” → what ought to be done, what’s right and moral
  • Can these live inside artifacts and products?

8

9 of 30

Is it ethically neutral?

9

10 of 30

So, What Makes an Artifact Political or Ethical?

  • Some artifacts just work differently depending on context
  • Others require certain social arrangements or enforce them
  • Design can express power (who can use, who’s excluded)
  • Design can encode values (what counts as right, safe, efficient)

10

11 of 30

Langdon Winner: “Do Artifacts Have Politics?” (Daedalus,1980)

  • Core claim: “The machines, structures, and systems of modern material culture can embody specific forms of power and authority.”
  • Key idea: Technologies aren’t neutral tools. They can be political in themselves — not just because of how people use them, but because of how they are designed and arranged.

11

Langdon Winner (1944-)

Rensselaer Polytechnic Institute

12 of 30

Two Ways Artifacts Have Politics

12

Type

Description

Classic Example

Modern/Product Example

1. Intentional or design-based politics

A device or system is designed to settle a social issue or enforce a power relation.

Robert Moses’s low bridges that kept buses (and thus low-income riders) from reaching Long Island beaches.

AI hiring filters designed to “streamline” recruitment but effectively exclude certain groups.

2. Inherently political technologies

Some technologies require or strongly favor specific forms of social organization or authority to function.

Production line or nuclear power requires centralized, hierarchical control for safety.

Autonomous weapons or AI surveillance infrastructures that require state or corporate centralization.

13 of 30

13

 Bridges on the way to Jones Beach, Long Island, NY

Production line of a 19-cent. cotton spinning mill

14 of 30

14

“Defensive designs” & “Hostile designs”

15 of 30

Why This Matters for Product Management

  • Winner’s takeaway: Technological choices are like legislative acts — they establish frameworks of order that endure.
  • PM implications:
    • Design decisions stabilize certain values and behaviors.
    • Once embedded in code, interfaces, and metrics, those values become hard to reverse.
    • “Neutral” design choices can quietly structure who has agency, who benefits, and who gets excluded.

15

16 of 30

Spotting Product Ethics

16

If technology can embody power and ethics, what kinds of political or ethical assumptions are built into the products you work on — even before users touch them?

Hint: Intentionally designed so? Unintentionally reinforce them? Require them? More compatible with them? Or in any other sense?

17 of 30

How do politics/ethics get into products?

17

One Answer: through social processes of problem definition, solution design, and debate closure, or the Social Construction of Technology (SCOT)

Trevor Pinch & Wiebe Bijker

“The Social Construction of Facts and Artifacts” (1984)

18 of 30

SCOT in a Nutshell

  1. Interpretive flexibility: the same technology means different things to different social groups.
  2. Relevant social groups (RSGs): identify who shapes and interprets the artifact (users, engineers, managers, regulators, investors, etc.).
  3. Problem definition: each group defines what the technology is “for” — what problem it solves.
  4. Closure: when one definition and design becomes dominant, the artifact stabilizes — and so do its embedded values and power relations.

18

19 of 30

The Case of the Bicycle

  • A number of competing designs in bike history, esp. in the1880s

19

“Betty Bloomers”-1894

“Penny-farthings”-1870s

“Boneshakers”-1820s

20 of 30

Social Groups

    • Producers
    • Sport cyclists
    • Tourist cyclists
    • Women cyclists

20

Designs

    • Brakes
    • Air tires
    • Spring frames
    • Lower wheel

Problems

    • Safety
    • Speed
    • Vibration
    • Dresses

21 of 30

  • Tall penny-farthing (high-wheel)
  • Speed
  • Elite
  • Safety bicycle (chain and equal wheels)
  • Safety
  • Inclusive

21

22 of 30

Social Groups

    • Producers
    • Sport cyclists
    • Tourist cyclists
    • Women cyclists

22

Designs

    • Brakes
    • Air tires
    • Spring frames
    • Lower wheel

Problems

    • Safety
    • Speed
    • Vibration
    • Dresses

Closure

  • Safety campaigns, mass adoption, and technical improvements.
  • The safety bicycle became the “real” bicycle.
  • Competing meanings faded; design stabilized.

23 of 30

TikTok and the Politics of the Feed

  • The “For You” page uses AI to autonomously curate each user’s stream.
  • Optimization goal: maximize engagement.
  • The algorithm learns, acts, and adjusts continuously — not just reacting, but pursuing its own objectives.
  • Raises questions of visibility, influence, and value.

23

24 of 30

A SCOT Lens at TikTok’s “For You” Algorithm

24

Relevant social groups

Their “problem to solve”

Values embedded

Platform engineers

Deliver personalized, sticky content

Technical efficiency, optimization

Executives / investors

Maximize user time, ad revenue

Profit, scale

Creators

Grow audiences, earn income

Visibility, fairness

Users / parents

Avoid harm, support well-being

Safety, mental health

Regulators

Prevent manipulation / disinformation

Accountability, transparency

Closure: Engagement as the dominant value metric — other meanings (well-being, diversity, civic health) marginalized.

Stabilization: Infrastructure + culture make alternatives hard to imagine — “That’s just how algorithms work”

25 of 30

Your Agency in the Social Construction of Technology

  • You’re part of the “relevant social groups.” The PM voice often defines “the problem.”
  • Problem framing is political. It decides which users, harms, and benefits count.
  • Closure isn’t neutral. When you ship, you stabilize one version of the world.
  • Agency: You can broaden the frame — include marginalized users, redefine “success,” and reopen closure points.

25

26 of 30

No more boring things?

26

27 of 30

But who decides what’s boring and what’s meaningful?

    • What perspectives or kinds of work are being left out or undervalued?
    • How do such design assumptions reflect particular social, cultural, or economic values?
    • What kind of human–AI division of labor is being imagined here — and for whose benefit?

27

28 of 30

From Managing Ethics to Building Ethics Into Products

  • Technologies aren’t neutral: they embody power and values (Winner).
  • Politics gets in through design: problem definitions, social groups, and closure (SCOT).
  • Products make value choices material: optimization goals become infrastructures of power.
  • PMs have agency: every metric, feature, and default stabilizes one version of the world.

28

29 of 30

29

Gather with your group project members, and reflect on the following questions:

  1. Which “relevant social groups” are involved in your project?
  2. How have your groups framed the design problem? How might it be framed differently?
  3. How did you reach a consensus on the problem and its solutions?
  4. Who does the product benefit and who might be left out?

Unpacking Product Designs

30 of 30

30

References

  • Winner, L. (1980). Do Artifacts Have Politics? Daedalus, 109(1), 121–136.
  • Pinch, T. J., & Bijker, W. E. (1984). The Social Construction of Facts and Artefacts: Or How the Sociology of Science and the Sociology of Technology might Benefit Each Other. Social Studies of Science, 14(3), 399–441.

Thank you!

Acknowledgements: Content created with assistance from Louie Ortiz and Emily Lindsay Tianshi.