Research involving human participants requires ethical attention beyond methodological rigor. How do you ensure participants understand what they're consenting to? How do you protect privacy and confidentiality? How do you minimize harm while pursuing knowledge? When is research with vulnerable populations ethically justified? These questions are core to ethical research practice. Understanding research ethics frameworks and institutional review board processes ensures your research respects human dignity while advancing knowledge.
This lesson explores ethical principles guiding research, the institutional review process (IRB), and specific considerations for AI ethics research in the nonprofit and grants sectors. Whether you're conducting research as an independent nonprofit evaluating your own systems or as a researcher publishing in academic journals, ethical practice matters.
Modern research ethics rest on three foundational principles established by the Belmont Report (1979), a landmark U.S. document defining ethical research standards. These principles—respect for persons, beneficence, and justice—guide ethical decision-making across research types.
Respect for persons means recognizing individuals' autonomy and treating them as agents capable of making informed decisions about whether to participate in research. Operationally, this requires informed consent: participants must understand what research involves, what risks and benefits it carries, how their information will be used, and that they can decline or withdraw without penalty.
Respect for persons also demands protection of privacy and confidentiality. Researchers must safeguard personal information, limit use to specified purposes, and ensure data security preventing unauthorized access. For research on AI in grants, respecting persons means if interviewing nonprofit leaders about their AI adoption, they understand how findings will be used, that quotes may be attributed or anonymous as they choose, and that their organization's data is protected.
Beneficence means maximizing benefits and minimizing harms. Benefits of research on AI ethics in grants include advancing knowledge that could improve decision-making and reduce bias. Potential harms include time burden on participants, possible reputational risk if findings are critical, and risk that findings could be misused. Ethical research design asks: How do benefits outweigh harms? What precautions minimize harm?
Sometimes research offers direct benefits to participants—improved organizational processes, increased self-awareness, valuable insights. Other research benefits primarily the broader field. Both can be ethical, but transparency about who benefits matters.
Justice requires fair distribution of research burdens and benefits. Historically, research burdens (participation, risks) fell disproportionately on vulnerable populations—incarcerated people, people with mental illness, marginalized racial groups. Benefits accrued to privileged populations and researchers. Justice demands we don't exploit vulnerable people for research benefiting others.
For nonprofits and AI research, justice questions include: Are we asking overburdened nonprofit staff to participate in research without compensation? Are we studying AI's effects on marginalized communities without meaningful participation from those communities in defining research questions? Are we benefiting researchers and funders while nonprofit participants gain little?
IRBs are committees at research institutions charged with reviewing research involving human participants to ensure ethical standards. Most universities, many hospitals, and some nonprofit and research organizations have IRBs. If your research involves human subjects and is affiliated with an institution with an IRB, you typically must obtain IRB approval before beginning research.
IRB requirements vary by institution and jurisdiction, but generally, research involving human subjects requires review. This includes interviews, surveys, observations, data collection from people, or analysis of existing data containing personal information. Some research may be exempt or eligible for expedited review if minimal risk and simple procedures.
Research on your own nonprofit's AI system might not require external IRB review if conducted as organizational evaluation. But if publishing findings or seeking to contribute to external knowledge, IRB review becomes important. When in doubt, consult your institution's IRB office.
IRBs use three review levels. Exempt review applies to minimal-risk research like anonymous surveys with no sensitive questions or observation of public behavior. Expedited review applies to low-risk research that can be reviewed by a single IRB member rather than the full board. Full review applies to higher-risk research, involving sensitive topics, vulnerable populations, or significant intervention.
Research on AI ethics in nonprofits might qualify for expedited review if involving anonymous surveys of nonprofit leaders about AI adoption concerns. Full review would be needed if interviewing vulnerable populations about AI systems affecting their services, or researching how algorithmic decisions affect funding to marginalized communities.
Informed consent is central to research ethics. A proper consent process ensures participants understand what research involves and willingly agree. A consent form should explain: the research purpose, what participation requires, risks and benefits, how information will be used and protected, who to contact with questions, and that participation is voluntary.
For nonprofit leaders interviewed about AI adoption, consent might explain: "This research examines how nonprofits evaluate and implement AI systems. You will be interviewed for 60 minutes about your organization's experience. Your responses are confidential; we will not identify your organization in any reports. You can decline to answer any question or withdraw from the study at any time. Findings may be published in academic journals or nonprofit publications."
Special care is needed with vulnerable populations. If studying how AI affects funding to organizations serving justice-impacted or immigrant communities, ensure consent materials are accessible, available in relevant languages, and presented in ways respecting recipients' autonomy and dignity.
Privacy refers to individuals' rights to control their personal information. Confidentiality means researchers will keep information private. Data security involves safeguards preventing unauthorized access. Together, these protect research participants.
For AI ethics research, this means: storing interview recordings securely, separating identifying information from research data, using de-identified data for analysis, limiting who accesses data, and securely destroying data after specified periods. If promising confidentiality, keeping that promise is paramount. If anything compromises confidentiality, you're ethically obligated to inform participants.
Research on AI bias in funding touches sensitive topics. Organizations may fear findings revealing algorithmic discrimination could damage reputation. Researchers may face pressure from funding sources to find particular results. How do you maintain ethical integrity? By being transparent about methods and limitations, not selectively reporting findings, and protecting participants' identities when promised.
Vulnerable populations—those with limited power, resources, or agency—require heightened ethical protections. This includes nonprofit staff (whose jobs depend on cooperation with funders), people experiencing homelessness or other challenges, immigrants, people with intellectual disabilities, incarcerated people, and others facing systemic marginalization.
Extra protections include: ensuring truly voluntary participation without implicit coercion, providing compensation recognizing their time and expertise, involving community members in research design, and ensuring findings benefit those studied. If examining how AI affects funding to organizations serving vulnerable communities, engaging those communities in research design is ethically important.
Researchers are not neutral observers. Your position—race, gender, class, professional role, relationships to institutions—shapes what you notice, who trusts you, and what interpretations you construct. Ethical research acknowledges positionality and reflects on how it influences research.
If studying AI in grant allocation, your position matters: Are you a grantmaker studying grantees? A nonprofit leader studying other nonprofits? An external researcher? A person of color or someone from a historically marginalized group? An insider or outsider to nonprofit communities? Each position brings insights and blind spots. Acknowledging your position honestly strengthens research credibility.
Ethical research requires careful data management. Keep detailed records of what data you have, who it involves, what informed consents you received, how it's stored, who has access, and when it will be destroyed. Use secure storage—encrypted drives, password-protected servers, institutional data repositories. Limit access to research staff. Follow any data sharing agreements with participants or institutions.
For qualitative research like interviews, consider: Will interview transcripts be shared with research participants for review? Will identifying information be removed? How long will data be kept? When and how will it be destroyed? Transparency about these practices builds trust.
Publication ethics requires honesty about methods and findings. Don't selectively report results, exclude contradictory findings, or overstate conclusions. If funders or institutions pressure you to find particular results, resist. Research credibility depends on integrity.
Acknowledge limitations honestly. Every research has constraints: limited sample, specific context, methodological choices. Acknowledging these helps readers interpret findings appropriately. Fabricating data, plagiarizing, or misrepresenting findings violates research ethics fundamentally.
Research ethics frameworks vary internationally. The Belmont Report guides U.S. research. The Declaration of Helsinki guides international medical research. Different countries have different IRB structures and requirements. If conducting international research or publishing in multiple contexts, understand these variations.
Never begin research involving human participants without understanding and following applicable ethical requirements. Violations can harm participants, compromise your research credibility, and potentially result in institutional sanctions. When uncertain about ethical requirements, consult your institution's IRB office or ethics committee.
Research ethics protect human dignity while advancing knowledge. Core principles—respect for persons, beneficence, justice—guide ethical research design. IRB review, informed consent, privacy protection, and careful data management operationalize these principles. Ethical research requires honesty, transparency, and genuine commitment to minimizing harm while advancing knowledge that serves the field.
This lesson's seminar explores ethical decisions in AI ethics research through scenario analysis. Participants examine realistic research scenarios—interviewing nonprofit leaders about AI vendor evaluation, studying whether algorithms disadvantage organizations serving marginalized communities, analyzing grant allocation data for bias—and discuss ethical considerations. Through dialogue, you'll develop ethical reasoning skills applicable to your own research.
Research ethics aren't bureaucratic requirements to satisfy—they're core to meaningful research. Ethical research respects those who participate by protecting their dignity, privacy, and autonomy. It advances knowledge credibly by demanding honesty and integrity. It contributes to justice by ensuring benefits aren't concentrated while burdens fall on vulnerable populations. As you conduct research on AI ethics in nonprofits and grants, let ethical principles guide your work.
Learn to balance rigorous inquiry with ethical responsibility to participants and communities.
Explore Full Course