Prompts // 10 (Conversational Agents)

My last AI-related post, about prompt engineering and curriculum design, took me down a little sideroad as I worked on it, and now I know where it’s going.

Note: As with some of my recent AI-related posts, I have beenusing ChatGPT 4 to get me going and to help me ‘research’ the topic. What you see below is something that is me and ChatGPT, prompts have been modified and tweaked to form the article here.

Before we start I want to declare that, until recently, I have avoided ‘chatbots’. My experience with them has been the kind of ‘bot you find in a banking app, website customer support, or any other variation on this theme. Like many I’ve spoken to about this, the experience has been pretty negative – responses that bear no relevance to your original query, responses that ignore you and respond with what it thinks you need (not what you asked for) or, even worse, just not responding at all and telling you to phone them … despite your original phonecall directing you to the ‘bot.

The chatbot has, until recently, not been a success. But this has changed in the second half of 2023, with the advent of OpenAI and other companies providing the ability to create your custom chatbot.

While single, well-engineered prompts are crucial in guiding AI to provide a specific insight or analysis, the development of a custom chatbot designed for curriculum design encompasses a broader and more interactive approach.

Single Prompt Engineering:

• Focused and Specific: A single prompt is typically a one-time query designed to address a specific question or need. It should be highly targeted, aiming to elicit detailed and relevant information or analysis on a particular aspect of the curriculum.
• Limited Scope: The scope of a single prompt is confined to the question at hand. It does not adapt to follow-up queries or provide interactive guidance based on the user’s responses.
• Ease of Use: Single prompts are straightforward to create and can be used without needing extensive AI programming knowledge. They are ideal for quick, targeted insights.
• Example Application: A curriculum designer might use a single prompt like “Provide an analysis of the current trends in eLearning adoption in higher education” to gather specific insights for a potential course update.

Custom Chatbot for Curriculum Design:

• Interactive and Dynamic: A custom chatbot is designed to handle a series of interrelated queries, offering a more interactive experience. It can adapt to the user’s inputs, providing more conversational and comprehensive assistance.
• Broader Capabilities: Such a chatbot can cover a wide range of curriculum design aspects, from providing data-driven insights and trend analyses to suggesting best practices and innovative course materials.
• Complex Development: Building a custom chatbot requires more sophisticated AI programming and an understanding of the curriculum design process. It’s a more resource-intensive endeavour but offers greater versatility.
• Example Application: A chatbot could guide a learning designer through the process of integrating new technologies into a curriculum, offering suggestions, resources, and examples, and responding to specific concerns or queries raised by the designer.

The developing nature of AI-driven ‘conversational agents’ offers a range of functionalities that can potentially transform how educators, curriculum designers, learning developers, and university administrators approach their work. However, like any technology, chatbots come with their own set of advantages and disadvantages.

Advantages of Chatbots in Curriculum Design

Enhanced Interactivity and Engagement:

Chatbots offer a dynamic, interactive platform for curriculum designers, facilitating a more engaging design process. They can provide immediate responses and feedback, making the process more conversational and less static.

A chatbot can instantly suggest relevant educational theories or design principles when queried, enhancing the designer’s workflow.

Scalability and Accessibility:

Chatbots can handle multiple queries simultaneously, offering scalable support to a large team of curriculum designers. They are accessible at any time, providing flexibility in work schedules.

Multiple designers can interact with the chatbot simultaneously to get insights on different aspects of course development, regardless of their location or time zone.

Consistent and Up-to-Date Information:

Chatbots, when regularly updated, can provide the latest information and data, ensuring that curriculum design is aligned with current educational trends and research.

A chatbot can offer the latest findings in digital pedagogy or the newest educational technology tools that can be integrated into the curriculum.

Disadvantages of Chatbots in Curriculum Design:

Limited Understanding and Complexity:

Chatbots may lack the ability to fully understand complex, nuanced queries related to curriculum design, which can sometimes lead to irrelevant or superficial responses.

When asked about integrating complex subjects like interdisciplinary studies, a chatbot might struggle to provide comprehensive guidance.

Dependence on Quality of Input:

The effectiveness of a chatbot is largely dependent on the quality and specificity of the queries it receives. Vague or poorly framed questions can lead to less useful responses.

A generic query like “improve course” might not yield actionable insights unless further specified.

Ethical and Privacy Concerns:

The use of chatbots raises concerns regarding data privacy and security, especially when dealing with sensitive educational material and student information.

If a chatbot is used to analyse student performance data, there must be stringent measures in place to protect student privacy and comply with regulations.

Balancing the Pros and Cons:

Implementing with Caution:

While embracing the advantages of chatbots, it’s crucial to be cautious of their limitations. A balanced approach that combines the efficiency of chatbots with human oversight can be effective.

Use chatbots for initial data gathering and analysis but rely on human expertise for final decision-making in curriculum design.

Continuous Improvement and Feedback:

Regularly updating chatbot systems and incorporating user feedback can help mitigate some of the disadvantages over time.

Periodically refining the chatbot’s database and algorithms based on user experiences and emerging educational trends can enhance its effectiveness.

Chatbots present a development in the field of curriculum and learning design, offering significant advantages in terms of efficiency, scalability, and access to information. However, their limitations in understanding complex and ethical considerations necessitate a balanced and cautious approach. By acknowledging and addressing these challenges, curriculum designers can effectively leverage chatbots as a valuable tool in their toolkit, augmenting their expertise and creativity in designing educational programs.

Photo by Emiliano Vittoriosi on Unsplash