AI-Generated Tools for Academic Research

AI-generated research tools can support or hinder your intellectual, creative, and professional growth. Accordingly, we expect that you will use AI-based tools responsibly, with care, and with an eye toward understanding their systemic harms and benefits. Use the TAAP framework to guide your ethical AI practices:

Transparency & Disclosure
Authority Evaluation
Accuracy Validation
Privacy Considerations

1. Transparency & Disclosure

• Fully document AI's role in research:

  •    Cite AI tools used and clearly explain their specific contributions rather than presenting AI-generated content as your own work
  •    Highlight known limitations of the tool and provide context for AI-generated content
  •    Document your verification process for AI-generated citations
  •    Understand the tool's stated methodology and capabilities, while recognizing:
    • Many AI systems function as "black boxes" where the decision-making process is not fully transparent or reproducible
    • AI-powered search results are not perfectly reproducible because:
      • The same search may yield different results when run multiple times (inherent randomness)

      • Tools and databases change over time (updates and evolution)

      • Action: Document the date and version of AI tools used

2. Authority Evaluation

• Critically assess AI as a source:  
  • Investigate whether developers provide adequate information about how the tool was created, including: 
    • Training data sources and methodologies (what information was included or excluded)

    • Fine-tuning processes (how the model was adjusted)

    • Guardrails and content filters (what the system is designed to avoid)

• Check for potential conflicts of interest and consider the reputation of the AI tool's creators
• Do not automatically treat AI as an equivalent to human experts
• Recognize that general-purpose AI tools (ChatGPT, Gemini, Perplexity) are more prone to citation fabrication than specialized academic research tools (e.g., Undermind, Elicit, Consensus)

3. Accuracy Validation

•    Rigorously validate AI-generated content:
  •    Understand that AI systems can produce two types of errors:
    • Factual hallucinations: generating information that is simply untrue
    • Source faithfulness errors: citing sources that don't actually support the claim made, or generating "ghost references" (fabricated citations that don't exist)
  • Verify both types of errors: Check that citations exist and support their claims, AND independently verify factual information from multiple sources
  • Remember that even specialized academic tools are not 100% accurate—approach all AI-generated content with critical thinking and a skeptical stance

4. Privacy Considerations

•    Be mindful of potential privacy implications:
  •    Avoid uploading sensitive or personal information  
  •    Review the basic privacy policy of AI platforms 
  •    Be cautious about data sharing and storage