In this module, we explored generative AI and ways to assess content and tools applicable to learning. In this blog, I will detail my experience using two different AI websites, ChatGTP and Perplexity, supplemented with some thoughts on AI’s role in education, then, I will conduct an SECTIONS analysis of a peice of media used in this EDCI 337 course.
Testing the Limitations of Generative AI in Education
I decided to try ChatGPT because of its growing popularity in the mainstream population. After learning about the limitations of generative AI in this module, I decided to put ChatGTP to the test and see how well it would preform a complex task. The task I chose was designing a full lesson plan, specifically around creating a balanced diet.
To my surprise, ChatGTP performed quite well and spat out a decent first draft for a lesson plan, even if it was only a skeleton structure with no actual content. However, there were some bugs: my first prompt generated ‘half-way assessments’ and a ‘final cumulative assessment’ after each module, rather than one half-way and one at the end; and although my second attempt (editing the first prompt) somewhat corrected this, it only provided a course with two modules, labeled ‘module 2’ and ‘module 4’; while my third prompt (second edit of first prompt) spat out the most usable lesson plan, but was still lacking in terms of actual applicability, i.e., it would take some rejiggering and filling in blanks to make it a workable lesson plan (see Figure 1 and Appendix 1). As you can see, ChatGTP is a tool – not a paid assistant. Thus, what I learnt in Module 5 was proven to be true, ChatGTP’s output is only as good as it’s inputs, and it certainly requires an experienced and literate eye to check, refine and apply its work.
I then decided to compare the work of ChatGTP to another AI generative tool, Perplexity. I chose this as comparison because it is an open-source tool that does not require an account – so I assumed it would be less efficient and poorer quality. I began by plugging in the final prompt (3rd edit) used from ChatGTP. To my surprise, it was just as good (or bad?) as ChatGTP, with this complex task at least. It too had flaws, like reiterating my criteria for course overview rather than expanding on it with details, but for the most part, it was half-decent and likely took out ¼ of the time it takes to generate a lesson plan (see Figure 2 and Appendix 1).
Generative AI’s Role in Education
All in all, I think that AI generative technology certainly isn’t replacing a teaching degree anytime soon. That said, I think if it is used wisely, ethically and skillfully, generative AI can be provide tremendous support in terms of saving energy and time on tedious tasks or provoking inspiration.
SECTIONS Evaluation of Media
To understand how to effectively evaluate media for learning, I undertook an evaluation of our professor’s introductory “Welcome to EDCI337” video using the SECTIONS framework, posted on the EDCI website (link to video).
- Students’ Access: The video is readily accessible to students enrolled in an online asynchronous course. It is expected that these students are equipped with the necessary tools, such as a computer and internet access, to engage with the course content.
- Ease of Use: The video is user-friendly, leveraging the ubiquity of video media. This familiarity ensures that most students will have no difficulty engaging with the video content.
- Costs: The creation of the video is presumed to be time-efficient and cost-effective. The estimated time investment includes the video’s duration and an additional fourfold for planning. The cost is minimal, considering the availability of free and accessible video editing platforms.
- Teaching Functions: The video excels in supporting effective teaching by offering students a personal connection to their instructor. This humanizes the learning experience, fostering a positive attitude and confidence as students embark on the course, providing reassurance of the instructor’s commitment to their education.
- Interaction: The video facilitates a moderate level of interaction. It serves as an excellent medium for student-content engagement, employing dual coding theory for enhanced learning. Interaction between students and the instructor, though limited by its asynchronous nature, is still meaningful. However, it does not facilitate direct student-to-student interaction.
- Organizational Support: Adequate institutional support is anticipated for the development, maintenance, and access to this video. While direct assistance in scriptwriting may not be provided, technical support for video creation and website hosting is likely available through the university.
- Networking: The video is conducive to fostering learning networks. Its ability to be shared and accessed across various platforms enhances its potential to integrate and connect new learners within and beyond the course.
- Security and Privacy: The security and privacy aspect is somewhat addressed. While the video does not compromise student security, it offers limited protection of the instructor’s personal identity.
This assessment underscores the video’s effectiveness as an introductory tool, highlighting its strengths in accessibility, ease of use, and teaching efficacy, while also noting areas for potential improvement in interaction and privacy protection. To conclude, through my evaluation, I found this specific media to be shinning example of multimedia learning content. The only thing I might add is the ability for students to comment directly on the webpage to facilitate more student-student interaction.
References
“Please create a lesson plan for me with the following…” ChatGPT, 13 Feb. version, OpenAI, 10 Mar. 2024, chat.openai.com/chat.
“Please create a lesson plan for me with the following…” Perplexity, Aug. version, OpenAI, 10 Mar. 2024, perplexity.ai.