Integrating Responsible Uses of GenAI
Welcome to our webpage on the integration of GenAI into learning activities in higher education. Here, we provide practice examples and guidance on how to effectively incorporate GenAI into teaching and assessment design while promoting critical engagement and learning.
Incorporating GenAI tools into higher education requires a thoughtful and transpartent approach, with ongoing monitoring and evaluation, and must be guided by a commitment to enhancing the learning experience and upholding ethical standards. Through the (AI)2ed project we encountered numerous ways that GenAI could be integrated responsibly in higher education settings, alongside appropriate training (critical engagement) and acknowledgement of use.
Please explore the links below to find specific information you are looking for on this page.
Using GenAI Produced Answers as Templates
As was noted in several of the project case studies, GenAI software is capable of producing essay-style answers of a reasonable standard.
However, a common thread across the case studies, was that these answers lack any real depth or critical engagement with course content.
While these AI generated answers lacked critical engagement, their overall structure was deemed acceptable, and therefore could serve as a template or exemplar for students (on structure, flow of writing, clarity of language, grammar, etc.) as they begin their work on a particular assignment.
GenAI as Integral Part of the Assessment Design
In one of our case studies, Philosophy from CACSSS, the lecturer decided to fully integrate the use of GenAI into their assessment design, making its use a requirement for the assignment task.
Positioning GenAI front and centre in the assessment design acted to prevent or discourage the misuse or undisclosed use of GenAI. This also provided students with a guided introduction to the technology, enabling them to learn best practice in using GenAI technology.
Students were specifically told to include screenshots and transcripts of their interaction with the GenAI tool, as well as providing a rationale for the prompts that they used with respect to the issues and texts discussed throughout the module.
Identifying Gaps in GenAI Responses
Several of our staff-student pairings noted that using content produced by GenAI in a compare-and-contrast type exercise proved useful. Comparing GenAI outputs on a given topic with their own ideas/research on that topic generally lead to either:
- New points raised that they hadn’t considered and could now explore further
- Discovery of the limitations of the GenAI tool when compared with their own work
With this in mind, students can be asked to compare their own work with outputs from GenAI and reflect on and critically analyse these. This can help develop students’ understanding of how these tools work and encourage them to engage with them critically.
Make Students Aware of the Downsides of GenAI
It is widely acknowledged that GenAI is highly proficient at producing code. However, the Computer Science students working on this research project found that while GenAI is indeed capable of producing code to a decent standard, it has its limitations.
They found that code generated, through ChatGPT in this case, did not adhere to best practice in a number of ways. The most pressing of these is the code’s robustness from a security perspective. The GenAI output tends to be the simplest, and most straightforward answer to prompts, which potentially could leave the code produced open to hacking and other forms of cyber-attacks.
Also, code produced by these tools can lack maintainability. Maintainability refers to the practice of organisation-wide coordination on coding, that allows multiple software engineers to work on a single project because of consistency in formatting, functioning, and coding methods. Code produced by GenAI tends not to be consistent in these terms, even when appropriately prompted.
Students who rely on code produced by GenAI, that is sometimes suboptimal and ignores best practice, will not learn the necessary foundations to program effectively as they continue through their studies and move into the professional world. For this reason, it is vital that students are aware of the downsides of GenAI from an early stage and are able to develop their own skills.
This principle can be applied across the disciplines. As noted throughout this guide, while GenAI can produce essays to a certain standard, they lack depth and proper engagement with subject matter. Highlighting this to students, making them fully aware of the superficiality of GenAI answers and the need for rigorous critical engagement, is vital.
GenAI as a Study Aid
GenAI can be used as a study aid in a number of ways:
- GenAI can be utilised in a similar way to how a student might use a search engine or other such online resource. A student can engage a GenAI tool in a conversation, helping them to clarify concepts. As is the case with any research or study aid, a student’s capacity to critically engage with and analyse these outputs is vital.
- GenAI can also be used as a starting point to brainstorm ideas for research projects or essays, but again, students must ensure the content generated is reputable and legitimate.
- Literature Review: GenAI can assist researchers by summarising and highlighting key information in academic papers. For example, a student can input a block of text from a journal, from which GenAI can pull the key points and present them in simple bullet points. Students should be aware of copyright and how/where the GenAI tool stores and uses data when inputting content to it.
- Data Analysis: GenAI can help process and analyse research data, generating visualisations and insights that aid in research projects. Sensitive data should not be input into these tools unless it is protected and not used for training the tool.
- Grammar and Proofreading: Depending on the requirements of the course or module, GenAI can be used to help identify grammatical and punctuation errors in a student’s writing.
- Time management: GenAI can be used to break down tasks into smaller segments and create a study timetable.
Accessibility
More research is required into the impact GenAI can have in terms of accessibility and access to higher education for students with learning difficulties.
Some preliminary examples:
- Text-to-Speech and Speech-to-Text: GenAI can be integrated with accessibility tools to assist students with visual or hearing impairments.
- Language Translation: GenAI can help break language barriers by offering translation services for international students. It should not be used for this purpose to bypass learning in any course where the students' ability to speak/write the language is being evaluated.
Challenge ChatGPT
A student could be tasked with challenging ChatGPT to deliver answers of a certain standard.
In the case study from Neuroanatomy, the lecturer tasked the students with developing prompts for ChatGPT that would produce a detailed and comprehensive summary of a given topic.
The students were required to conduct traditional research, which would then inform their detailed prompts to ChatGPT, which in turn produced a detailed response from the software, matching the level and standard of their traditional research.
This act of challenging ChatGPT does not undermine intended learning outcomes, as the student has already completed the necessary research.
Also, this serves as a good example of developing a student’s AI literacy, woven into traditional teaching and learning.