Embracing AI: transforming teaching with responsible prompts

During the AI ChangeMakers project in the summer of 2023, we explored the potential roles of Generative AI (GenAI) tools in higher education and the importance of their responsible use by students.

MediaCentral Widget Placeholderhttps://mediacentral.ucl.ac.uk/Player/81Ij411CWatch the video on MediaCentral. Download the transcript [Word].  

What could happen 

We mapped out various scenarios to understand how this integration might unfold. Our conclusion is that, given their pervasiveness, it would be impractical to ban GenAI tools.  

If we consider two variables (‘technology adoption’ and ‘ethical consideration’) as ‘high’ or ‘low’, we get four scenarios: 

Low adoption and low ethical consideration

Nothing changes, AI is avoided – and students are completely unprepared for the world of work. 

High adoption and low ethical consideration

 rampant material plagiarism and near-identical student output and results, through AI. Is Higher Education obsolete? 

Low adoption and high ethical consideration

AI is not made available because ethical considerations outweigh the benefits. As a result, most students miss the chance to learn about AI and those who do use it gain an unfair advantage. 

High adoption and high ethical consideration

Where learning and results improve, teachers use AI to encourage critical thinking. The student journey involves fostering knowledge responsibly and ethically, as well as benefitting from the benefits AI has to offer. 

Emerging themes

We concluded that these tools can significantly enhance students’ learning journeys if used responsibly. The university plays a pivotal role in ensuring the realisation of high adoption and high ethical consideration – where students use the technology responsibly as part of their learning journey. 

We identified two emerging themes which are crucial in shaping responsible AI guidelines:

• prompt engineering directly influences the accuracy and relevance of AI responses, making it a key skill for students to master.
• fostering trust between teachers and students is essential for creating a safe, transparent, and effective learning environment where AI is used responsibly. 

Responsible prompting for students to ground understanding 

GenAI tools such as ChatGPT can be used as an always-available teacher to help build an understanding of a topic or reading. This can help students flesh out thought processes and ideas to develop a deeper understanding of the topic. The student can then validate this understanding with a teacher through online forums, classes or other direct teacher-student contact.

The student can tell ChatGPT the topic they wish to explore, along with any specific learning objectives (defined by the teacher), and then the student can delve into the topic. This can be enhanced by pasting an excerpt from one or more required readings, giving ChatGPT a good foundation for its positions and judgements.   

An example of this approach from our project

...
Anonymous: so this might complement well with kotter’s 8-step change model? which does provide a step-by-step guide

ChatGPT: Yes, the McKinsey 7-S model and Kotter’s 8-Step Change Model can complement each other well when managing organizational change. While the 7-S model focuses on assessing and analyzing the current state of an organization, identifying areas of alignment or misalignment, and highlighting where change may be needed, Kotter’s 8-Step Change Model provides a more structured and step-by-step approach to planning and implementing those changes.
Here’s how they can work together

…
Anonymous: is it safe to say that the 7-S model, although provides a comprehensive view of change, it doesn’t provide a practical path to implementing it?

ChatGPT: The McKinsey 7-S model is primarily a diagnostic tool designed to help organizations assess and analyze their current state in terms of alignment among various internal elements. While it provides a comprehensive view of an organization’s current situation and identifies areas of misalignment, it does not inherently provide a detailed, step-by-step path for implementing change. Instead, it highlights where changes may be needed.

…
Anonymous: Can you tell me about McKinsey’s 7-S model?

ChatGPT: Certainly! The McKinsey 7-S model is a management framework developed by consultants at McKinsey & Company in the late 1970s. It is designed to help organizations analyze and align various aspects of their internal environment to achieve strategic goals and improve overall performance. The model is based on the idea that there are seven key elements within an organization that need to be properly aligned for it to be successful. These seven elements are often referred to as “S’s,” and they are as follows...

Alternatively a student could send their notes to ChatGPT and prompt it to expand the notes to help improve one’s understanding. ChatGPT can then give both expanded notes or suggestions on what else the student should consider about a topic, and the student could follow this up with a conversation.  

An example from our project where a student has used the tool to expand notes on a topic

Anonymous: I’ve written up some notes on Porter’s five forces, can you review them and see how they might be expanded to better help my understanding and comprehension of the five forces? The competitive five forces that shape strategy (Porter, 2008) is a HBR article that goes over five forces that impact the competitiveness of a business. These are:

• Threat of new entrants: new entrants bring a new desire to gain market share and put pressure on prices, costs, and the rate of investment.

ChatGPT: Your notes provide a good summary of Porter’s Five Forces framework, but they can be expanded to provide a deeper understanding and comprehension of each force. Here’s an expanded version of your notes:

1.    Threat of New Entrants:

• Barriers to Entry: Consider what barriers may exist that deter new entrants from entering the market. These barriers could be financial (high capital requirements), regulatory (government restrictions), or technological (proprietary technology or patents).
• Economies of Scale: Discuss how established companies may have economies of scale, which allow them to produce goods or services at a lower cost per unit. This can be a deterrent to new entrants.
• Brand Loyalty and Switching Costs: Explain how existing customer loyalty to established brands or high switching costs for customers can make it difficult for new entrants to gain a foothold in the market.

The students must remain critical of the information ChatGPT is giving them to ensure accuracy, relevance and understanding.   

Integrating prompt engineering in teaching delivery 

Prompt engineering can be incredibly useful, making the classroom more interactive and engaging. For instance, we can incorporate prompt engineering exercises into our classroom delivery for formative activities. This might include generating concepts related to a specific topic or learning outcomes. Such activities can sharpen critical thinking skills as students write prompts and debate the results.

Meanwhile, teachers can critique the findings and guide students on what constitutes acceptable and non-acceptable prompts. This critique could become an activity in itself, where students learn how to verify data and findings via a range of methods. This process fosters trust between students and teachers, as teachers observe how students develop prompts while students learn to distinguish responsible prompts from others. 

Nurturing trust through collaborative AI integration 

As students engage with AI tools to augment their learning, establishing a foundation of trust is paramount. Teachers, by validating and guiding students in navigating the use of GenAI, foster an environment where students feel supported and empowered in their educational pursuits. This collaborative approach strengthens the teacher-student bond, instilling confidence in students to utilise GenAI responsibly.

Ultimately, this trust empowers students to embrace AI tools as valuable resources to develop skills such as critical thinking, problem-solving, and effective communication, while ensuring responsible and ethical utilisation within the educational landscape. 

Embrace and encourage 

It’s a fact that GenAI tools have disrupted the Higher Education landscape. However, as tutors, we can embrace this technology and encourage students to use it responsibly. This requires a fundamental paradigm shift in our teaching delivery.

We can embrace AI to significantly enhance the learning experience for students, broadening their horizons and offering personalised learning opportunities.

We trust our students to use GenAI tools by integrating them into our teaching and encouraging their responsible use. As teachers, our responsibility extends beyond imparting knowledge. It involves nurturing a generation that thinks critically and acts responsibly.

Echoing Jeremy Bentham, we aim to use GenAI ‘to create the greatest happiness for the greatest number.’ 

Disclaimer: the video scripts are inspired by ChatGPT. 

Find out more

To learn more about this project, watch the AI Lunch & Learn session (minute 32:00 onwards) 

Other UCL AI ChangeMaker projects are on the UCL Arena blog.

On prompt engineering as an academic skill: A model for effective ChatGPT interactions, Times Higher Education

Bronwyn Eager and Ryan Brunton, 'Prompting Higher Education Towards AI-Augmented Teaching and Learning Practice', Journal of University Teaching & Learning Practice, Volume 20 Issue 5 Quarterly Issue 2.