AI Research Ethics: IRB Adaptation

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ethics-safety-AI-and-research-ethics-adults

1. A researcher is designing a study that uses an AI chatbot to conduct initial participant screening. Before submitting to the IRB, which action would best prevent revision delays during the review process?

   1. Incorporate AI-specific ethics considerations into the protocol design from the start
   2. Wait for the IRB to request additional AI-related information
   3. Use the same protocol template from a non-AI study with minor word changes
   4. Submit the protocol without mentioning the AI chatbot to gauge initial reactions

2. Which of the following risks is considered an AI-specific concern that researchers must address when using AI in their studies?

   1. Standard participant time commitments
   2. Standard informed consent procedures
   3. Algorithmic bias perpetuating existing inequities
   4. Standard data collection costs

3. A research team plans to use machine learning to analyze participant survey responses. What element must be explicitly disclosed in their IRB protocol?

   1. The funding source for the study
   2. The team member who will operate the computer
   3. The comprehensive description of how AI will be used in analysis
   4. Only the final analysis results they expect to obtain

4. Why is updating consent language particularly important when participants will interact directly with an AI system during a study?

   1. The IRB requires consent language updates for all studies regardless of AI involvement
   2. Consent language changes are optional and can be addressed after the study concludes
   3. Participants must understand they are interacting with AI, not a human, to make informed decisions
   4. Updated consent language is only required for studies involving medical interventions

5. A longitudinal study begins using AI for data analysis but later wants to add AI-powered participant recruitment. What must the research team do?

   1. Wait until the study completes to address the new AI use
   2. Submit a completely new protocol for the recruitment component
   3. Submit an IRB amendment detailing the new AI recruitment method
   4. Continue the study without informing the IRB since analysis already uses AI

6. Which of the following statements accurately reflects a limitation of using AI in research ethics planning?

   1. AI can substitute for human judgment in determining whether research meets ethical standards
   2. AI eliminates the need for IRB review when used in research protocols
   3. AI tools cannot predict every possible ethics issue that may arise during a research project
   4. AI can automatically generate compliant consent language for any research scenario

7. An IRB protocol mentions that collected data will be used to train a future AI model. Which element should this protocol address?

   1. The academic credentials of the research team
   2. The cost of cloud computing services
   3. Only the immediate research analysis plan
   4. AI training and re-use considerations for the collected data

8. What distinguishes AI-specific ethics considerations from generic research ethics in IRB protocols?

   1. There is no distinction; all research ethics are essentially the same
   2. AI-specific considerations are only relevant for technology-focused institutions
   3. Generic research ethics already cover all AI-related concerns
   4. AI-specific considerations address risks like algorithmic bias and hallucination-generated misinformation

9. A researcher discovers their AI recruitment tool is inadvertently excluding participants from certain demographic groups due to biased training data. What should be done?

   1. Submit findings to the IRB only after the study concludes
   2. Continue using the tool since recruitment targets are being met
   3. Replace the AI with a human recruiter without documenting the issue
   4. Address this bias as an AI-specific risk requiring immediate protocol revision

10. Which action would violate the principle of comprehensive AI disclosure in IRB submissions?

    1. Listing AI use in both recruitment and analysis sections
    2. Providing detailed technical specifications of the AI systems used
    3. Describing all AI tools and their purposes in the protocol
    4. Mentioning only the primary AI application while omitting auxiliary AI uses in data processing

11. Why must researchers stay current on institutional and federal AI research guidance?

    1. Institutional guidance is optional and rarely affects protocol approval
    2. Federal guidance only applies to studies with government funding
    3. AI ethics frameworks and regulations are evolving rapidly and may change during a study
    4. IRBs never update their requirements

12. A study uses an AI system that occasionally generates factually incorrect but plausible-sounding responses. Which risk does this represent?

    1. Hallucination harm, where AI generates misleading or false information
    2. Standard measurement error inherent in all research methods
    3. Statistical software configuration issues
    4. Data loss during participant interviews

13. When an AI system processes participant data, what specific data handling consideration should be documented in the IRB protocol?

    1. The meeting schedule for research team briefings
    2. How the AI system stores, accesses, and potentially shares participant data
    3. The brand of computer used for data entry
    4. The ambient temperature of the server room

14. A researcher argues that their AI-powered analysis is just like traditional statistical software and therefore doesn't need special ethics consideration. What is the flaw in this reasoning?

    1. AI analysis is faster, so ethics review can be shorter
    2. AI systems introduce unique risks like bias amplification and opacity that traditional software lacks
    3. Statistical software has been proven to be completely unbiased
    4. Traditional software has more stringent ethics requirements

15. What happens if researchers skip AI disclosure in their IRB submission?

    1. The protocol may be rejected or require substantial revisions upon discovery
    2. The study will proceed faster without unnecessary oversight
    3. The IRB will automatically approve it without review
    4. Funding agencies will not notice the omission