Give

Guiding Principles

While Generative AI tools like ChatGPT, Claude, and others can enhance research efficiency and creativity, they also come with important risks and limitations. Understanding these challenges and adopting thoughtful best practices can help researchers use generative AI effectively while safeguarding research integrity and accuracy. When using Generative AI for research, it is important to recognize that these models are trained on large datasets that may reflect underlying biases, inaccuracies, and inconsistencies. AI-generated content is not inherently reliable, it reflects patterns in the training data rather than verified knowledge. As such, researchers should approach AI-generated outputs with a critical mindset, carefully evaluating the quality, accuracy, and appropriateness of the information produced.

Below are general best practices to follow when incorporating generative AI into research:

  1. Verify and Fact-Check AI Outputs
    • AI-generated content can be highly convincing but may contain inaccuracies, incomplete information, or even fabricated details (“hallucinations”).
    • Cross-reference AI-generated data, summaries, and analyses with peer-reviewed sources and primary literature.
    • When generating citations or references, verify them manually, as AI models may produce fictitious or misformatted citations.
  2. Address Bias and Ethical Considerations
    • AI models reflect the biases present in their training data. This can lead to skewed or misleading outputs, particularly when dealing with sensitive topics like race, gender, and cultural issues.
    • Be aware of potential biases in the data and methodology behind the AI tool you are using.
    • Strive for balanced representation and ensure that AI-generated outputs are not reinforcing harmful stereotypes or inaccuracies.
  3. Maintain Confidentiality and Data Security
    • Avoid inputting sensitive, unpublished, or proprietary information into AI tools, especially those that store user interactions or are cloud-based.
    • Use locally hosted or secure, privacy-protected AI models for research involving confidential or sensitive data.
    • If working with human-subject data, ensure that AI usage complies with institutional review board (IRB) requirements and data privacy regulations (e.g., GDPR, HIPAA).
  4. Understand AI’s Limitations with Language and Context
    • AI-generated translations may not accurately reflect technical terms, cultural nuance, or the intended tone of the original text.
    • For research in specialized fields, consult human experts to ensure the accuracy and contextual relevance of AI-generated translations.
    • Be cautious when generating content in languages other than the AI’s primary training language.
  5. Clearly Attribute and Document AI Use
    • Acknowledge AI-generated contributions in research papers, presentations, and other outputs where applicable.
    • Follow institutional or publisher guidelines for citing AI-generated content.
    • Maintain transparency by describing how AI was used in the research process (e.g., data analysis, writing assistance, hypothesis generation).
  6. Use AI as a Tool—Not a Replacement for Critical Thinking
    • AI can automate certain research tasks (e.g., data coding, text summarization), but it should complement rather than replace human judgment.
    • Critically assess AI-generated content, especially when it comes to forming research conclusions and interpretations.
    • Combine AI-generated insights with traditional research methods and expert analysis for a more rigorous and balanced outcome.
  7. Be Mindful of Overreliance on AI
    • AI can generate useful drafts and ideas but should not substitute for deep domain knowledge and critical evaluation.
    • Develop a balanced workflow that leverages AI for efficiency while maintaining human oversight and intellectual rigor.
    • Monitor how AI is affecting your research methodology and adjust as needed to maintain research integrity.
  8. Keep Up with AI Developments and Institutional Guidelines
    • AI technology is rapidly evolving, and new capabilities and limitations emerge frequently.
    • Stay informed about changes to the AI tools you use, including updates to their training data, algorithms, and privacy policies.
    • Consult your institution’s policies on AI use in research to ensure alignment with academic integrity standards.