Skip to main content

Trinity College Dublin, The University of Dublin

Trinity Menu Trinity Search

Centre for Academic Practice, Trinity Teaching & Learning

You are here Resources > Generative AI Hub > GenAI and Assessment

GenAI and Assessment

Many of the key concerns about GenAI are related to assessment and academic integrity. Staff have questions about how GenAI can be used to support assessment processes without compromising academic integrity. In a world where GenAI is widely used, many students are unsure what constitutes acceptable use of GenAI in an assessment context.

This section aims to support you to:

  • recognise three  academic integrity principles to help face key GenAI challenges in relation to assessment;
  • reflect on what plagiarism means in the context of GenAI;
  • consider how to design GenAI in or out of the assessment process;
  • recognise and evaluate using the traffic light model for Assessment and GenAI;
  • define key questions when planning using GenAI in Assessment.

Click on the titles below to expand each section:

Academic Integrity in an age of GenAI

Academic integrity can be described as a commitment to, and compliance with:

ethical and professional principles, standards, practices and consistent system of values, that serve as guidance for making decisions and taking actions in education, research and scholarship
NAIN, 2021 Lexicon, p.3

But what does academic integrity mean in an age of GenAI, where it is increasingly difficult to ascertain what content has been generated by GenAI tools and what has been developed by a student? How do we maintain a focus on authentic assessment without engaging in an unproductive arms race to try to outwit or detect GenAI use? Traditional plagiarism detection tools are rendered ineffective in this context, as AI becomes more capable of human-like text (Dwivedi et al., 2023).

In response, universities worldwide are shifting from merely focusing on detecting GenAI uses to strengthening proactively academic integrity. This shift involves (re) designing assessment strategies emphasising critical skills and creativity, and fostering a culture of integrity in all academic staff and students do.

In this new context, Crawford (2023) argues that GenAI is challenging our understandings of what “integrity” in “academic integrity” functionally means in the face of AI.  He argues that we need to define and then apply a reframed academic integrity lens to artificial intelligence whereby “integrity” in assessment is taken to mean:

Authenticity

This principle highlights the importance of academic staff ensuring that the student’s submission has been developed independently. In the era of GenAI, applying this principle requires academic staff to think about strategies for obtaining evidence of student authorship of submitted work. For example, adding reflective elements to the assignment, asking for progressive draft submissions with tracked changes to evidence students’ progress, and adding creativity or critical thinking to the assessment questions.

Responsibility

This principle refers to the crucial role of academics guiding students to understand how to utilise GenAI tools ethically and effectively. Simultaneously, there is a need for academic staff to responsibly (re)design assessment strategies with greater robustness to make them resistant to misuse. These strategies should also focus on gathering evidence about students' ability to think critically and to apply their knowledge in specific and professional contexts.

Accountability

This principle promotes and recognises the shared responsibility between academic staff and students in maintaining clear rules, understanding, and transparency when using GenAI in any assessment. It calls for a cohesive alignment with the institutional and faculty academic policies, which need to be updated to remain clear and relevant for all.

Click here for more information on Academic Integrity.

What does “Plagiarism” mean in the context of GenAI?

If a student generates content from a GenAI tool and submits it as their own work, it is considered plagiarism, which is defined as academic misconduct in accordance with the College Academic Integrity Policy. If a sentence or quotation from a GenAI output is  used by a student in their work, then it must be referenced. Cases of plagiarism are considered under College Academic Misconduct Procedures.

To maintain  academic integrity when using GenAI in an assessment, the specific uses of GenAI  must be disclosed, properly acknowledged and referenced. Students should explain in detail what, when, and how they used GenAI in their assignment. For detailed guidelines on how to reference the use of GenAI, please see page: Using GenAI in Teaching and Learning - FAQs.

Designing GenAI in or out of the assessment process?

The decision to “design in” or “design out” GenAI may depend on several factors but should start with a review of current assessment design, and specifically a review of the alignment of current assessment(s) and learning outcomes. It may also be an opportunity to consider the volume of assessments, and the cohesion across and between assessments in a programme.

In reviewing learning outcomes and their associated assessment(s), you might consider the following questions:

  • Consider your module learning outcomes. What do you want your students to achieve with this assessment? What core skills do you intend for them to develop?  For more information on how to design assessments to align with learning outcomes see: Linking Assessment Methods with Learning Outcomes using Blooms Taxonomy - PDF 220KB.
  • Consider how exactly students may or may not use GenAI for your assessment in order to meet learning outcomes.
  • How vulnerable are these learning outcomes/assessments to Gen AI?  If the task focuses on knowledge, comprehension, and written accuracy, it will be vulnerable. However, learning outcomes/assessment focused on originality, specific contexts, and critical thinking, will be less vulnerable.
  • Is this an opportunity to review learning outcomes and consider the incorporation of higher order skills such as analytical thinking or critical analysis?
  •  If learning outcomes and assessments are already focused on higher order skills, could GenAI be used to further develop these skills?

For more information see Designing & Delivering Assessments.

For an example of how to apply this re-design process in practice, please see:

Assessment and GenAI: the Traffic Light Model

One of the main challenges when embedding GenAI within assessment practices is ensuring that the work and/or learnings presented are the students own. While initial responses to GenAI focused on strategies and tools for detecting the use of GenAI, it is now generally acknowledged that focusing on detection is unrealistic and unachievable. The challenge has now shifted to incorporating the use of GenAI into the assessment process while safeguarding the integrity of the assessment process.

One common means of achieving this is through the widely-adopted traffic light model (Perkins, et al., 2024) which offers a hierarchical structure for clearly communicating demarcations of acceptable AI usage within an assignment or task.  While helping students avoid inadvertent academic misconduct or accidental breaches of academic integrity, this hierarchical model performs a dual role, allowing the educator to guide their thinking in relation to incorporating GenAI into assessment redesign.

Building on Perkins et al.’s framework, the University of Limerick have adapted this model as outlined below (click on the left and right arrows to navigate the table):

Table 1: The Traffic Light model adopted in the University of Limerick (guidance developed with reference to Perkinset al., 2024). A downloadable PDF version of this table is available here.

Frequently Asked Questions:

To download these questions, please click here.

Can I or my students take ideas generated by GenAI and use them within an assignment?

If GenAI is used at any stage of an assignment process, it must be acknowledged in the same way as any other resource. Remember that long pieces of text from GenAI are not allowed, and students should disclose what tool they used and how they used it. Students should show evidence of their own work based on the GenAI output. For example, Students can get initial ideas for a project. Then, they must show how those ideas were transformed into original pieces of work.  

Can my students get GenAI to solve an equation?

Providing the use of GenAI is properly acknowledged, asking it to do mathematical equations or other calculations will not infringe academic integrity. You, as a lecturer, should define and clearly communicate to your students what role the equation or calculation played in the assessment. If the point of the exercise is to test your student's ability to do math, and this is done by GenAI rather than your student, then the assignment submission does not accomplish the principle of authenticity in academic integrity.

(Adapted from the School of Languages, Literatures & Cultural Studies, Student Handbook 2023-24)

Can my students use GenAI to write a computer programme?

GenAI is highly impacting coding. Companies such as Amazon Web Services and IBM have been promoting its use worldwide. However, in the context of higher education, students should demonstrate that they have achieved the knowledge and skills required for their programme learning outcomes. If, during any process of the coding assignment, students use GenAI (guided by clear academic staff instructions), they must transparently disclose how and where GenAI was used.

Can my students use GenAI to paraphrase something they have written and submit this content as part of an assessment?

As mentioned before, students must show evidence of the originality and authorship of their work. If they rely on GenAI's output to enhance and polish their original work, the result comes from the GenAI tool, and the student is no longer the author.  If, instead, your student provides evidence of asking GenAI something like "Find gaps or weaknesses ideas in my piece" or "Give me six ways this piece of writing could be improved", then they should provide the GenAI's output as an appendix and demonstrate how these recommendations were implemented in the original work.

Key Takeaways

  •  Universities worldwide are shifting their focus from detecting GenAI uses, to proactively strengthening academic integrity through (re)designing assessment strategies and fostering a culture of integrity in alignment with the three academic integrity principles:  Authenticity, Responsibility, and Accountability. These must be achieved for each student at all stages of the assessment process.
  • Plagiarism is deemed to have occurred when students submit a GenAI output as their work. To avoid plagiarism, students must disclose what, when, and how they used GenAI for their assignment submission.  GenAI in an assessment must follow the school guidelines and the relevant referencing format.
  • “Designing in” or “designing out” GenAI starts with reviewing the current assessment design and its alignment with the learning outcomes. If this alignment is cohesive, the next step is to analyse how vulnerable the assessment strategy may be. The more original, specific, contextualised, and critical thinking tasks involved in the assignment, the less vulnerable it will be.  
  • The traffic light model offers a hierarchical structure detailing the extent of legitimate GenAI usage within an assignment or task guiding students to avoid involuntarily engaging in academic misconduct, and guiding the thinking of academic staff  in relation to incorporating GenAI into assessment redesign.
  • When using GenAI in assessment: Students’ use of GenAI should follow the School guidelines, and should be disclosed in a transparent and accurate manner.

Resources

Previous Next