This document is a return-to list for grounding articles, talks, and product education in credible, citable evidence about how artificial intelligence intersects with civil-society organizations (including churches), donors, staff, volunteers, and the people they serve, especially youth and families in the United States.
It is scoped for the Movemental organizational site context: mission-driven institutions that care about trust, stewardship, formation, and long-horizon outcomes, not only efficiency.
How to use this list
- Start with the tier that matches your claim: opinion → start with Pew/Gallup/Ipsos; macro labor/costs → BLS/OECD/Stanford AI Index; governance → NIST + your jurisdiction’s regulator; nonprofit operations → NTEN/TechSoup/Candid-class sources; scholarship → Semantic Scholar/OpenAlex/PMC.
- Prefer primary documents (the report PDF, the agency page, the dataset) over secondary blog summaries.
- Date-stamp everything: AI moves quickly; cite the report year, field dates, and revision where applicable.
- Distinguish evidence from advocacy: trade associations, think tanks, and campaigns can be useful for arguments and framing but should not be presented as neutral statistics without corroboration.
- Use the router first for drafts: Movemental content angles (writer’s router) maps common questions to sections and lists typical mis-use patterns.
Source-quality rubric (quick)
| Signal | Strong | Weak |
|---|
| Method | Published methodology, sample frame, limitations | “Survey says” with no methods |
| Independence | Nonprofit research institute, government statistical agency, peer review | Vendor whitepaper as sole source for market size |
| Fit | Sector-specific (nonprofit) + general AI trends triangulated | Single viral headline |
| Humility | Confidence intervals, replication, dissenting studies | Certainty language on contested social effects |
<span id="movemental-content-router"></span>
Movemental content angles (writer’s router)
This section does not add new authorities. It maps recurring questions from nonprofits, churches, and movement leaders to the numbered source sections below, with misuse warnings so drafts stay honest under Movemental’s mix of trust, stewardship, formation, and outcomes.
| Layer | What it can prove | Where it usually lives here |
|---|
| Empirical | Prevalence, attitudes, reported behavior, institutional context, economics (with methods limits) | Sections 1–8, 11–12 |
| Ethical / pastoral / theological | Obligations, virtues, wise response, congregational discernment | Section 9 (plus your tradition’s primary texts and credentialed interpreters, not covered in this doc) |
| Debate / advocacy | Competing values and policy stakes | Section 10 (always label incentives) |
In public pieces, do not present the second layer as if it were representative survey data, or vice versa.
Question → start here → avoid this
| Audience question or writing job | Start in these sections | Primary entry points (repeat) | Common mis-use to avoid |
|---|
| “What are US teens actually doing with AI?” | Sections 1, 2 | Pew: Teens & technology, Pew: Artificial intelligence, Common Sense research | Turning cross-sectional survey shares into causal claims about wellbeing or character |
| “How should parents/pastors think about mental health and screens?” | Sections 1, 9 | Surgeon General / HHS (search advisories on social media and youth), WWC when making learning claims | Collapsing social media evidence into “AI” without saying what changed in the claim |
| “What do Americans (or our people) believe about AI?” | Sections 2, 12 | Pew: Artificial intelligence, PRRI | Cherry-picking one demographic bar without trend/context |
| “Do we need AI governance (especially with volunteer boards)?” | Sections 3, 4, 8 | NIST AI RMF, NIST AI RMF Playbook, NTEN AI hub | Importing enterprise maturity theater (badges, vanity scores) as if it were fiduciary duty |
| “What must we protect: donor data, HR, beneficiaries?” | Section 3 | NIST Cybersecurity Framework, FTC Business Guidance Blog | Assuming “we’re small, so attackers don’t care” (nonprofits are targets of opportunity) |
| “Is using AI in fundraising ethical / legal / trust-safe?” | Sections 3, 5 | AFP ethics, Fundraising Regulator AI guidance, FTC Business Guidance Blog | “AI wrote it, not us” defenses (accountability still sits with the institution) |
| “What’s changing in philanthropy and grant-seeking?” | Sections 4, 5 | Candid, Lilly Family School of Philanthropy, GivingTuesday | Projecting platform hype onto aggregate giving without checking the actual report’s scope |
| “Will staff lose jobs? What do we owe people?” | Sections 6, 7, 10 | BLS, OECD.AI, EPI for labor-critical framing | Single vendor forecast as destiny; enterprise workforce surveys as “nonprofit reality” |
| “What does AI cost (tokens, vendors, consulting)?” | Section 7 | AI Index | Quoting headline ROI from a consultant study without sample and sector |
| “How mature are we compared to peers?” | Sections 4, 8 | NTEN AI hub, NIST AI RMF | Confusing tool adoption with risk governance or mission alignment |
| “What does AI change about learning and formation?” | Sections 1, 9 | ERIC, WWC, Semantic Scholar | “Science says” from preprints or EdTech marketing |
| “We serve EU partners / donors: what applies?” | Section 3 | EDPB, EU AI Act overview | US-only blog posts presented as global compliance |
| “Map the debate without pretending neutrality” | Section 10 | Pick two opposing frames + one rights/governance anchor (e.g., EFF, AI Now, ITIF) | “Both sides” symmetry when power and harms are asymmetric. Name what is at stake. |
These sources help answer questions about what young people are actually doing with AI, how adults perceive risks, and how digital environments relate to wellbeing (often adjacent to “AI” specifically).
| Source | Why it is authoritative | Entry points |
|---|
| Pew Research Center | Gold-standard public opinion and behavioral surveys; frequent releases on teens, schools, and technology. | Topic hub: Teens & technology. AI topic hub: Artificial intelligence. Internet & technology: Internet & technology. |
| Common Sense Media | Large-scale US surveys on media/tech use with youth and parent angles; useful for household norms and concerns. | Research landing: Research. |
| US Surgeon General / HHS | Official federal health guidance; especially strong for adolescent mental health and social media (often the upstream frame for “screens, identity, isolation”). | Advisory on social media and youth mental health: HHS press release index (search “Surgeon General advisory social media”). Surgeon General site: SurgeonGeneral.gov. |
| National Center for Education Statistics (NCES) | Official US K–12 and postsecondary statistics; helpful for institutional context (device access, connectivity) that shapes AI use in schools. | NCES. |
| CRITICAL CONSUMPTION NOTE | Longitudinal causal claims (“AI causes X”) are often not established in pop coverage. Prefer sources that separate association, mechanism hypotheses, and intervention evidence. | When writing for churches/nonprofits, pair survey facts (what % do what) with pastoral ethics (how should we respond), and label each layer clearly. |
2. Public opinion, trust, and civic attitudes (US and comparative)
Use this tier when you need “what people believe / fear / hope” statistics.
| Source | Why it is authoritative | Entry points |
|---|
| Pew Research Center | Deep crosstabs; trend questions; religion and technology intersections. | AI hub: Artificial intelligence. Religion: Religion. Datasets: Pew Research datasets. |
| Gallup | Long-running tracking on workplace, wellbeing, leadership, and change management themes that show up in nonprofit HR conversations. | Gallup topics. |
| Ipsos / Eurobarometer | Strong for global and EU comparative opinion on AI regulation and trust. | Ipsos: Ipsos.com. Eurobarometer: European Commission Eurobarometer. |
| YouGov | Useful for fast-moving topical polling; treat as one pollster among many unless corroborated. | YouGov. |
3. Governance, ethics, safety, and compliance (what boards and exec teams should know)
| Source | Why it is authoritative | Entry points |
|---|
| NIST (AI Risk Management Framework) | Widely referenced voluntary baseline for AI risk management in the US ecosystem; includes generative AI profile materials. | AI RMF home: NIST AI Risk Management Framework. Playbook: NIST AI RMF Playbook. |
| NIST Cybersecurity Framework / privacy engineering | Nonprofits still must protect PII, donor data, HR records; AI expands attack surface and data flows. | NIST Cybersecurity. |
| FTC | US enforcement perspective on consumer protection, deceptive claims, and emerging AI marketing issues. | FTC Business Guidance Blog. |
| EDPB / EU AI Act (for global NGOs) | If you serve EU residents, partners, or donors, EU materials matter even for US-based orgs. | European Data Protection Board: EDPB. EU AI Act portal (Commission): Digital strategy / AI. |
| Markkula Center for Applied Ethics (Santa Clara University) | Durable applied ethics framing and cases; helpful for board-ready ethics language beyond compliance checklists. | Markkula Center for Applied Ethics. |
| Partnership on AI | Multi-stakeholder AI organization publishing on responsible practices and ecosystem topics (evaluate each paper on its own methods). | Partnership on AI. |
4. Nonprofit-specific: adoption, capacity, digital equity, and field infrastructure
| Source | Why it is authoritative | Entry points |
|---|
| NTEN | Nonprofit technology field builder; practical governance and equitable adoption framing. | AI resource hub: AI for nonprofits (NTEN). |
| TechSoup / field partners | Ecosystem surveys and benchmarks on nonprofit tool adoption (including AI) appear intermittently via partners; treat each report as a snapshot, not eternal truth. | TechSoup: TechSoup. Example of a widely circulated 2025 benchmark report (verify the latest edition before citing): Climate Advocacy Lab resource page referencing nonprofit AI benchmarking: Climate Advocacy Lab resources. |
| Candid (Foundation Center / GuideStar) | Best for philanthropic market structure, grant data, and nonprofit financial transparency. Not usually focused on AI mechanics, but essential for fundraising claims grounded in sector economics. | Candid. |
| Independent Sector / state nonprofit associations | US nonprofit policy voice; helpful for workforce, policy, and operating environment context. | Independent Sector. |
| Bridgespan Group | Influential nonprofit management research and guidance; strong for strategy, operating model, and leadership topics when they publish on technology change. | Bridgespan. |
5. Fundraising, philanthropy ethics, and donor trust
6. Workforce, training, reskilling, and volunteer ecosystems
| Source | Why it is authoritative | Entry points |
|---|
| Bureau of Labor Statistics (BLS) | Official US occupational and wage statistics; anchor claims about roles, employment, and pay carefully. | BLS. |
| OECD AI Policy Observatory | International policy papers and indicators on AI diffusion, skills, and labor market themes. | OECD.AI. |
| World Economic Forum (Future of Jobs report) | Widely cited survey-of-employers view of skills demand; good for directional trends, not gospel. | WEF publications. |
7. Economics: inference costs, corporate adoption, consulting markets, and ROI narratives
| Source | Why it is authoritative | Entry points |
|---|
| Stanford Institute for Human-Centered Artificial Intelligence (HAI): AI Index | The default macro annual compendium for AI R&D, investment, models, and selected responsible-AI indicators. | AI Index report hub: AI Index. Example year page: AI Index Report 2025. |
| McKinsey / Deloitte / Gartner (use carefully) | Useful for enterprise adoption patterns and executive surveys; often non-representative of small nonprofits. Prefer labeling as “survey of X respondents / firms.” | McKinsey: McKinsey featured insights. Deloitte: Deloitte insights. Gartner: Gartner. |
| US BEA / Census (as needed) | If you make claims about economic output or business dynamics, prefer official statistics over vendor estimates. | BEA. US Census Bureau. |
8. AI maturity models, adoption frameworks, and “how mature are we?”
| Source | Why it is authoritative | Entry points |
|---|
| NIST AI RMF | Not a “maturity badge,” but a serious risk governance framework nonprofits can adapt. | NIST AI Risk Management Framework. |
| ISO/IEC standards family (via ISO portal) | Some orgs require ISO-aligned governance language for procurement; access is often paywalled. Cite the exact standard number and edition. | ISO standards catalogue. |
| Industry maturity rubrics (Gartner/Forrester/etc.) | Helpful for enterprise buyers; translate carefully for nonprofits (capacity, IT staffing, vendor dependence). | Treat as diagnostic language, not moral truth. |
This is where theology/ministry sources and learning science should meet empirical research, not replace it.
| Source | Why it is authoritative | Entry points |
|---|
| ERIC / What Works Clearinghouse (WWC) | Evidence standards for education interventions; helps you avoid mirage “learning gains” claims. | ERIC. WWC. |
| APA / major psychology publishers (open summaries) | For psychologically sensitive claims (addiction analogies, anxiety, cognition), prefer peer-reviewed routes via Semantic Scholar/OpenAlex (below). | American Psychological Association: APA. |
| Faith-adjacent academic centers (selective) | Use for ethical reasoning and pastoral practice, not for market statistics. Examples: university centers publishing on technology and moral formation. | Search via OpenAlex/Semantic Scholar for authors/institutions your audience trusts; cite as theology/ethics, not as empirical prevalence. |
10. Strong critics, strong proponents, and “who says what?” (use with explicit labeling)
This section is for mapping debates, not for laundering ideology as neutrality.
| Voice / org type | Typical emphasis | Entry points / examples |
|---|
| AI safety / catastrophic risk organizations | Long-horizon risk, governance, model evaluation debates. | Center for AI Safety. Future of Life Institute. |
| Digital rights / civil liberties | Surveillance, discrimination, due process, consent. | Electronic Frontier Foundation. EPIC. ACLU. |
| Labor-focused critique | Worker power, deskilling, contingent labor, data extraction. | Economic Policy Institute. |
| Industry-friendly innovation policy | Competitiveness, standards, deployment, growth framing. | Information Technology and Innovation Foundation. |
| “Humane tech” reformers | Attention economy, design ethics, organizational habit change. | Center for Humane Technology. |
| AI Now Institute | Critical policy voice on power, accountability, and institutions. | AI Now Institute. |
When writing for churches and nonprofits, name stakeholder incentives explicitly: who funded the study, who benefits from the recommended policy, what the study cannot measure.
Use these when you need mechanisms, effect sizes, replication, and debate rather than headlines.
| Source | Why it is authoritative | Entry points |
|---|
| Semantic Scholar | Fast paper discovery, citation signals, OA PDF pointers. | Semantic Scholar. |
| OpenAlex | Open catalog of scholarly works/institutions; good for mapping a field. | OpenAlex. |
| PubMed Central | Biomedical and adjacent psych research with OA full text when available. | PMC. |
| arXiv | Preprints (not peer-reviewed); great for engineering, risky for social “facts” without corroboration. | arXiv. |
12. Religion surveys and “what US congregations believe/do” (statistics, not homiletics)
| Source | Why it is authoritative | Entry points |
|---|
| Pew Research Center (Religion) | Representative religion surveys in the US and globally. | Religion. |
| PRRI | Public religion and values research with policy and cultural salience. | PRRI. |
Maintenance
- Review cadence: quarterly for AI governance + quarterly for Pew/AI Index releases; annually for standards/regulator pages.
- Versioning: when you cite a source in content, copy the exact URL and retrieval date into your draft’s reference list.