Schedule

Thirteen sessions, September 9 to December 9, 2026, in a weekly three-hour block. The tables list the sequence and the deadlines; the reading for each session follows below.

Enrolled students receive the rest in Canvas, where each session has a page for what to do beforehand and a page for what happens in the room. The syllabus holds the policies, the grading, and the course description, and the assignments are published in full with their rubrics.

Module 1, Foundations (Weeks 1–2)

Working definitions, the history of technology in social services, and the sociotechnical protocol used all semester.

Module 1, Foundations (Weeks 1–2), session by session with what is due
#DateSessionDue
1Sep 9Definitions of technology and social good
2Sep 16History of technology in social services

Module 2, AI (Weeks 3–6)

Algorithmic bias and fairness, how generative AI works, critical and safe use, and organizational adoption.

Module 2, AI (Weeks 3–6), session by session with what is due
#DateSessionDue
3Sep 23Algorithmic bias and fairnessEthics Case launched
4Sep 30Language model mechanics
5Oct 7AI literacyEthics Case: case and research plan (Oct 7, 9am)
6Oct 14AI adoption in organizationsDigital Story: track and problem chosen (Oct 14, 9am)

Module 3, Data, industry, and capacity (Weeks 7–10)

Data rights and privacy, digital storytelling, the digital maturity of an organization, and how the technology industry is structured.

Module 3, Data, industry, and capacity (Weeks 7–10), session by session with what is due
#DateSessionDue
7Oct 21Data sovereignty and privacyStrategy Pitch: organization named (Oct 20, 5pm); Ethics Case due (25%)
8Oct 28Digital storytellingDigital Story: ABT sentence; brief or story of self sheet drafted in class
9Nov 4Digital maturity, with Lauri Goldkind (Fordham) and Jason Cummings (Empowerment Strategies)Strategy Pitch: capacity assessment and interview protocol in lab
10Nov 11Structure of the technology industry, with Jes Kane (former Google.org)

Module 4, Equity and implementation (Weeks 11–13)

Production and accessibility, design justice and ICT4D, digital equity, and the strategy pitch presented in class.

Module 4, Equity and implementation (Weeks 11–13), session by session with what is due
#DateSessionDue
11Nov 18Production and accessibilityVideo notes (Nov 17, 5pm); rough cut, ungraded; Capacity brief
12Dec 2Design justice, digital equity, and environmental costsDigital Story due (25%), memo draft for the practice review
13Dec 9Strategy pitchesPitch (of 25%); memo due Dec 16

Readings by session

Everything assigned, in order. Course readings are open access or available through the University of Michigan Library. Where a reading is a video, a report, or a primary document, the link goes to it directly.

Week 1. Definitions of technology and social good

The internet was announced as a great equalizer, and for a long stretch the evidence agreed. Speaking, organizing, and publishing became cheap enough that Benkler credited the network with a new, commons-based mode of production, with free software and Wikipedia as his leading exhibits.

Topics

  • Affordances and scripts
  • Design and power
  • Kranzberg's first law
  • Person-in-environment
  • Engineering move and solution-focused move
  • Technosolutionism
  • Amplification thesis

Readings

Week 2. History of technology in social services

American poor relief was housed in buildings before it was filed in records. Boston opened its first workhouse in 1660, and by the nineteenth century the poorhouse was the main form of public aid, its minimal food, forced labor, and crowding kept deliberately punitive to discourage anyone with another option. An 1883 investigation exposed conditions at the almshouse in Tewksbury, Massachusetts, where Anne Sullivan, later Helen Keller's teacher, spent part of her childhood; she called her years there "a crime against childhood."

Topics

  • Scientific charity and the casework registry
  • Reform maps
  • Technology as public infrastructure
  • Early computerization debates

Readings

Week 3. Algorithmic bias and fairness

The MiDAS case below documents the harm side of algorithmic bias. On overrides, at Allegheny County's Family Screening Tool, a CHI study found that caseworkers who overrode the algorithm's risk score cut the Black-white screen-in disparity from 20 to 9 percent, drawing on context the algorithm lacked, such as a family's engagement with private services that never enter administrative records. A follow-up analysis on updated county data confirmed the finding.

Topics

  • Four sources of algorithmic bias
  • The New Jim Code
  • Allegheny Family Screening Tool
  • Competing definitions of fairness
  • Audit outcomes

Readings

Week 4. Language model mechanics

The mechanism drafts a plain-language version of a dense notice, produces a first-pass translation for a person to check, and role-plays a hard conversation on demand. The Trevor Project trains crisis counselors against a simulated teenager built this way, and Be My Eyes describes images for blind users with it. Capability has also moved: METR times how long a task a model can finish with 50 percent reliability and finds that horizon doubling on the order of months, and the systems arriving in agencies now call other software and take many steps without checking back, which the agentic systems case below treats.

Topics

  • Next-token prediction
  • Transformer
  • Hallucination and the training objective
  • Strengths of the mechanism

Readings

Week 5. AI literacy

When a chatbot receives the same health question in English and in Arabic, the Arabic answer is measurably worse. In that study, bilingual experts scored four chatbots on 15 questions about five infectious diseases; English answers averaged 4.6 on a 5-point scale, in the excellent range, while Arabic averaged 4.1. The same drop has been measured for Spanish, on real patient questions about epidurals, where bilingual obstetric anesthesiologists rated Spanish answers significantly lower than English ones, and for Chinese and Hindi across three expert-annotated health question sets. The tool performs worst for many of the people social-impact organizations serve.

Topics

  • Calibrated trust
  • Ask AI or ask a person
  • Language gap
  • Error, accountability, and participation
  • Misinformation and teach-back
  • AI in social work practice

Readings

Week 6. AI adoption in organizations

Randomized trials continue to find that generative AI helps the least-experienced workers most, on the kind of writing work intake involves, and a 2026 RCT found a similar pattern along education lines: with AI access, an education-based productivity gap of 0.548 standard deviations fell to 0.139, closing about three-quarters of it, and participants kept part of the gain in a follow-up round without AI. Automated systems in this same administrative space, including one run by this state, have hurt people at scale.

Topics

  • Adopt, defer, or refuse
  • Vendors and procurement
  • Gap between task and deployment
  • Adoption as an implementation problem
  • Organizational AI policy
  • Evidence checks

Readings

Week 7. Data sovereignty and privacy

Module 2 asked whether the classification is biased. This week asks who governs the data at all. Ask of any harm case whether the affected community could have refused the data collection, and almost every one (MiDAS, Allegheny, SafeRent) had no way to refuse. The conceptual move is from privacy, an individual defensive right, to data sovereignty, a collective governing right. An individual can consent to a survey, but a community cannot un-consent to what is built from the aggregate.

Topics

  • Privacy and data sovereignty
  • Networked privacy
  • Four power questions
  • CARE and OCAP in production
  • Refusal as a justice claim

Readings

Week 8. Digital storytelling

Madden, Harrison, and Vafeiadis argue that nonprofit fundraising works better, and more ethically, when it rests on relational care: treating the storyteller as a partner in an ongoing relationship instead of a means to a donation. Their ethics-of-care model treats fundraising as relationship-building and extends the same dialogue-based obligations to donors, employees, beneficiaries, and volunteers. On that argument, final edit belongs with the person on screen, because the exposure is entirely theirs and the organization's incentives, a stronger ask, a longer video, a better shot, pull against theirs. An experiment on poverty portrayals found that negative, deficit-framed depictions measurably lowered viewers' ratings of the subject's own agency, an editorial choice an organization under fundraising pressure is tempted to make.

Topics

  • Story circle
  • Public narrative
  • Storytelling as organizing
  • Poverty depictions
  • Consent and labor
  • Voice and listening

Readings

Week 9. Digital maturity

A technology recommendation needs an assessment of the organization's own capacity. This week covers that assessment, which asks who owns technology inside the organization, what the data is held in, which of the minimum safeguards are in place, and what the system costs to operate after the grant that bought it is spent. The Strategy Pitch launches on it, and the memo rests on a single interview that gathers this information in one conversation.

Topics

  • Self-assessment instruments
  • Accidental techie
  • Data infrastructure
  • Minimum safeguards
  • Total cost of ownership
  • Realist evaluation
  • Semi-structured interview

Readings

Week 10. Structure of the technology industry

The people who build a product decide what it does, and some of those roles are open to clinicians and practitioners. Grow Therapy's clinical quality and documentation integrity manager posting accepts a licensed clinical social worker among its eligible credentials and puts that person on the clinical review step for product initiatives, checking that clinical requirements are represented correctly before anything ships. AI governance work has become its own profession, and a 2025 survey by the International Association of Privacy Professionals (IAPP) found 77 percent of organizations building AI governance programs, rising to about 90 percent among organizations already using AI. The function has no established organizational home, splitting mainly across privacy, legal, and IT teams, and only 1.5 percent of the surveyed organizations reported no additional AI-governance staffing needs for the coming year. Supporters argue that changing who holds these roles changes the product before it ships.

Topics

  • Product cycles and economics
  • Role map
  • Product roles and governance roles
  • Labor behind the product
  • Worker organizing
  • Grantor-grantee relationship

Readings

Week 11. Production and accessibility

The Digital Story is due next week, and today is the last supervised build time. Holding a video back until it is fully accessible has a real cost: auto-generated captions are roughly 90 percent accurate, cost nothing, and are available right away, and a systematic review of assistive technology in education finds real, measurable inclusion gains once tools are actually used, imperfect or not. The same evidence base includes a warning about what is never measured: a systematic review of 11 randomized trials of digital health interventions for children in underserved and rural settings found improved outcomes alongside a near-total absence of cost-effectiveness or equity-disaggregated analysis.

Topics

  • Accessibility law
  • Automatic captions
  • Production lanes
  • Accessibility overlays
  • Imagery ethics

Readings

Week 12. Design justice, digital equity, and environmental costs

Design justice, digital equity, and the environmental cost of computing are usually taught as separate subjects, and this session takes them together because a plan written for a real organization has to answer all three. The federal money behind digital inclusion has been withdrawn: the Digital Equity Act, the largest federal investment in digital inclusion ever made, was terminated in May 2025 with the money unspent in most states, and the litigation to restore it is still running. The Affordable Connectivity Program had already lapsed in June 2024 after serving more than 23 million households with $30 to $75 monthly internet subsidies, and the Congressional Research Service weighed the remaining alternatives, FCC Lifeline at $9.25 a month, VA telehealth-connectivity programs, and ISP low-cost plans, and concluded that no single one would fully replace it.

Topics

  • Design justice
  • Toyama's amplification thesis
  • Three levels of the divide
  • Federal retreat
  • Environmental numbers
  • Infrastructure ownership

Readings

Week 13. Strategy pitches

Before class

  • Your slides