AI Assessment Strategies

Practical Strategies for Life Sciences Instructors in the Age of Generative AI
As AI tools become increasingly common among students, many instructors are rethinking how they design assessments and verify learning. While there is no single solution, emerging best practices suggest that assessments should place greater emphasis on process, application, and authentic demonstration of understanding.

Guiding Principles

1. Assess the Process, Not Just the Product
Rather than evaluating only a final submission, consider assessing stages of student work, such as:
• Research questions or hypotheses
• Annotated outlines
• Concept maps
• Draft submissions
• Lab planning documents
• Reflective statements describing the evolution of the student's thinking

These intermediate steps provide insight into student learning and make it more difficult to rely exclusively on AI-generated work.

2. Use Course-Specific and Contextualized Assignments
Assignments are more resistant to inappropriate AI use when they require students to:
• Analyze class-generated data
• Interpret laboratory results
• Apply concepts to local ecosystems, case studies, or current research discussed in class
• Connect multiple course concepts to novel situations
• Reflect on their own observations, experiments, or field experiences

3. Incorporate Verification Opportunities
For major assignments, consider:
• Oral presentations
• Brief viva-style discussions
• Question-and-answer sessions
• Reflective "creator statements"
• Version-history review for written assignments

These approaches can help confirm authorship and deepen student engagement with their work.

4. Increase Transparency Expectations
Some instructors have found it helpful to:
• Require assignments to be written in platforms with version-history functionality
• Require access to document revision histories
• Include clear syllabus statements about acceptable evidence of the writing process
• Clearly articulate permitted and prohibited AI use

5. Use Scaffolded Assessment Design
Breaking large assignments into smaller components can:
• Encourage continuous engagement
• Reduce last-minute AI dependence
• Support stronger learning outcomes
• Provide multiple opportunities for feedback

Examples include:
• Topic proposal
• Literature search summary
• Annotated bibliography
• Draft methods section
• Final report or paper


Assessment Ideas for Life Sciences Courses

Laboratory Courses
• Lab notebooks reviewed periodically
• In-class data interpretation exercises
• Oral defense of experimental methods
• Reflection on unexpected results and troubleshooting

Lecture-Based Courses
• In-class case analyses
• Short written responses to recent class discussions
• Concept mapping activities
• Application-based examinations using novel scenarios

Resources

• The AI Pedagogy Project (metaLAB at Harvard): case studies, instructor guides, and amazing assignment ideas designed to integrate, or not, genAI in a constructive way. Harvard's Academic Integrity and Teaching Resources.
• International Center for Academic Integrity (ICAI): Asks us to rethink the punitive nature of many academic assessments. Publishes strategic toolkits.

 

AI in Teaching
1. ‎Course and Content Creation: Large Language Models/Chat Bots.  Can help with tasks like answering questions, drafting documents, summarizing text, creating images, or automating processes.  Note. It is highly recommended to use, and cite, the LLM supported by your institution that conforms to their privacy and data protection policies.
  •  Cogniti – specifically designed for education, integrates with several learning management systems
  • ChatGPT Edu – specifically designed for education
  • Microsoft Copilot Chat
 Instructor Use Examples
  1. Generate course content – drafting framework/outline, summarizing topic, breaking topic down into manageable sections, creating examples
  2. Improving accessibility to material – generating alt text, creating summaries
  3. Assess student work - generating practice and/or discussion questions, facilitating Q & A, creating rubrics, creating personalized feedback
 Student Use Examples
  1. Using AI in assessment creation – students can develop a document that is a record of their interactions with AI as an appendix to an assessment
  2. Critically evaluate genAI responses –. students use genAI to generate a response to a question, then critical evaluate, elaborate on, and correct the AI-generated response.
  3. Ethics and Responsibility in AI - students engage with generative AI tools to critically assess their outputs, identifying and challenging inaccuracies or distortions that may contribute to biased or discriminatory outcomes.
2. Course and Content Organization AI.   Organizes course content, answer FAQs, automates queries, prompts students to check their understanding of material.  
  • Google Notebook
Uploaded materials may include:
  • Course-specific materials (lecture notes, slide text, course assignments, lab manuals, etc.) 
  • Transcripts from lecture recordings 
  • Supplementary reading materials (might include articles, reports, case studies, online resources) 
  • Textbooks 
  • Research publications (avoid multi-column pdf, find single column sources)
Instructor and Student Use Examples. Either instructors or students can use to organize course documents and summarize, create glossaries, quiz students, answer FAQs. 
  
Reading List