Blogging With An AI-Powered Content Generator
Back in the early days of flying, pilots had to be brave and competent at the same time. Pilots provided engineers with feedback to improve the airplane for the next flight and in the end, this contributed to the current safety of the airline industry. This metaphor illustrates the interaction and collaboration between humans and machines. Right now, we find ourselves in the same early days of collaboration between humans and Artificial Intelligence (AI)-driven chatbots. This article is an experiment about the interaction between an AI chatbot and myself. I chose to collaborate with OpenAI GPT and their AI-driven chatbot for content generation: ChatGPT.
The questions and comments are mine. The responses generated by ChatGPT (in block quotes!) are unedited. This is an experiment with ChatGPT (test version) to experience the opportunities and shortcomings of learning content generation. The experiment is not about questions like “do you love me?”, but fun questions that will be enlightening as well. The topic is relevant for Learning and Development (L&D) practice, with a discussion around formal and informal learning. Red alert! Reading this article might cause ChatGPT-related FOMO.
ChatGPT
The first question is, “what is ChatGPT3?”
ChatGPT:
is a chatbot that uses the GPT-3 language model developed by OpenAI. GPT-3 is a large, powerful language model that has been trained on a massive amount of text data and can generate high-quality content on a wide range of topics. ChatGPT allows users to interact with GPT-3 through natural language conversation, allowing them to ask questions, provide input, and generate text on demand. ChatGPT is designed to be easy to use and to provide users with quick and accurate responses to their queries.
Let’s Have Some Fun With The AI Chatbot
We start with a simple question to warm up the chatbot. “Write a song about formal and informal learning.”
ChatGPT:
Verse 1: In the classroom, we learn the facts In a structured way, that’s a given But outside the walls, there’s a whole world Waiting to be explored and driven
Chorus: Formal and informal, they go hand in hand Each has its place, in this learning land Formal can teach us, what we need to know But informal shows us, how to make it grow
Verse 2: Informal learning, it’s all around In the experiences, that we have found It’s in the conversations, and the things we do It’s how we learn, to think and be true
Chorus: Formal and informal, they go hand in hand Each has its place, in this learning land Formal can teach us, what we need to know But informal shows us, how to make it grow
Bridge: Together they work, to make us whole Formal and informal, they both play a role In our journey of learning, they guide the way Towards a brighter future, with each passing day
Chorus: Formal and informal, they go hand in hand Each has its place, in this learning land Formal can teach us, what we need to know But informal shows us, how to make it grow.
This is cool! Creating lyrics for a song in a couple of seconds! This is the power of AI. While these lyrics are perhaps a bit awkward, they can surely compete with the current top 40. And just as quickly, comparable text can be generated with the opposing view! “Write a song against informal learning.”
ChatGPT:
Verse 1: Informal learning, what a shame It’s unstructured and without a name No curriculum, no plan Just a bunch of random, haphazard scans
Chorus: Formal learning, that’s the way Structured and organized, here to stay Teachers and instructors to guide the way Tests and exams to measure what you say
Verse 2: Informal learning, it’s just a game No recognition, no one knows your name No sense of accomplishment, no motivation Just aimless wandering, no destination
Chorus: Formal learning, that’s the way Structured and organized, here to stay Teachers and instructors to guide the way Tests and exams to measure what you say
Bridge: Formal learning, it’s the key To success and achievement, you’ll see It’s the foundation of all we know Informal learning, let it go
Chorus: Formal learning, that’s the way Structured and organized, here to stay Teachers and instructors to guide the way Tests and exams to measure what you say.
Garbage In, Garbage Out
It’s the same old story. If you don’t provide effective input, you don’t get the required output. A similar pattern applies to the interaction between humans and AI chatbots (by Donald Clark). That’s something to reflect on when using these AI-powered tools. The proof of the pudding is in the eating. How will ChatGPT respond to two different questions symbolic of the debate of “formal versus informal learning”?
The first question for ChapGPT is: “Write a short article about formal and informal learning—assuming they are not two separate entities.”
ChatGPT:
Formal and informal learning are two different types of learning that are often distinguished by their structure and delivery. Formal learning refers to learning that takes place within a structured environment, such as a school or training program, and is typically organized and planned. Informal learning, on the other hand, refers to learning that occurs outside of a formal learning environment and is often unstructured and spontaneous.
While these two forms of learning may seem quite distinct, they are not two separate entities. In fact, they are closely intertwined and can often be found together in a single learning experience. For example, a person may attend a formal training program that includes a mix of lectures, discussions, and practical activities, and then apply what they have learned in their daily work. This work experience may then provide opportunities for informal learning, such as through observations, feedback, and reflection.
One reason why formal and informal learning are not two separate entities is that they both involve the acquisition of knowledge and skills. Whether learning takes place in a structured environment or not, it is still learning. And whether learning is planned or spontaneous, it still involves the acquisition of new knowledge and skills. In this sense, formal and informal learning are two sides of the same coin.
Another reason why formal and informal learning are not two separate entities is that they both involve the application of learning in real-world settings. Formal learning often includes practical activities and simulations that allow learners to apply what they have learned in a realistic environment. Informal learning also often involves the application of learning in real-world settings, such as through on-the-job training or self-directed study. In both cases, learning is not just about acquiring knowledge, but also about applying it in a way that is useful and relevant.
Furthermore, formal and informal learning are not two separate entities because they both involve the participation of learners. Learning is not something that is done to learners, but something that learners do themselves. In both formal and informal learning, learners are actively involved in the process of acquiring and applying knowledge and skills. This means that learning is not just a one-way process, but a two-way exchange between learners and their environment.
In conclusion, formal and informal learning are not two separate entities, but two different perspectives on a single learning process. Both forms of learning involve the acquisition and application of knowledge and skills, and both involve the participation of learners. By understanding the connections between formal and informal learning, we can better support learners and help them to achieve their learning goals.
In a matter of seconds, this article provides a rather generic view of formal and informal learning. But if you assume informal learning is not relevant for acquiring and applying knowledge and skills, you might prefer to ask ChatGPT to “write a blog against informal learning in organizations.”
ChatGPT:
Informal learning, also known as “learning on the job” or “self-directed learning,” is a type of learning that takes place outside of structured educational settings. While informal learning can certainly have its benefits, it is not without its drawbacks, especially when it comes to learning in organizations.
One of the main problems with informal learning is that it is often unstructured and lacks a clear learning objective. Without a specific goal in mind, employees may end up learning things that are not relevant to their job or the organization’s needs. This can lead to a lack of focus and direction, and can ultimately result in wasted time and resources.
Another issue with informal learning is that it is often dependent on the individual’s willingness to learn. Some employees may be more motivated to learn on their own, while others may be less inclined to do so. This can lead to a lack of consistency in learning outcomes, and can create unequal opportunities for learning and development within the organization.
Additionally, informal learning is often not supported by the organization. Without proper support, employees may not have access to the resources and materials they need to learn effectively. This can lead to a lack of knowledge and skills, which can ultimately hinder the organization’s ability to achieve its goals.
In conclusion, while informal learning can certainly have its benefits, it is not without its drawbacks. In organizations, informal learning can lead to a lack of focus, inconsistency in learning outcomes, and a lack of support. For these reasons, it is important for organizations to prioritize structured learning opportunities that are aligned with the organization’s goals and provide employees with the resources and support they need to learn effectively.
This experiment demonstrates the power and the weakness of using ChatGPT for creating learning content. Like pilots at the beginning of the airline industry, L&D professionals have to be skilled in creating the right input to get the required output (according to the standards of the L&D profession). “Skilled” in the previous sentence isn’t a reference to twenty-first-century skills, but to L&D professionals who are able to validate learning design and content creation based on scientific research. Evidence-informed learning design should be the professional standard within L&D and will play an imperative role in improving AI chatbot-driven content generators. More information about evidence-informed learning is available from thought leaders like Will Thalheimer, Clark Quinn, Mirjam Neelen and Paul Kirschner, Donald Clark, Julie Dirksen, Jane Bozarth, and many others.
Opportunities And Shortcomings For Generating Learning Content
It’s clear that content generation with ChatGPT is incredibly quick and easy! But that’s not enough to create learning content according to the standards of the L&D profession (e.g., evidence-informed learning). That’s why I am going to ask ChatGPT about the advantages and limitations of creating learning content in different ways. Will we get honest answers?
Question 1: “How can ChatGPT be supportive of generating learning content?”
ChatGPT:
Can be useful for generating learning content because it can generate high-quality, accurate text on a wide range of topics. GPT-3, the language model that powers ChatGPT, has been trained on a massive amount of text data, which allows it to generate content that is coherent, relevant, and informative. This can be useful for educators and trainers who need to produce learning materials quickly and efficiently, or who want to generate ideas for new learning content.
Another way that ChatGPT can be supportive for generating learning content is by providing learners with a personalized learning experience. Because ChatGPT allows users to interact with it through natural language conversation, learners can use it to ask questions, provide input, and receive personalized responses. This can make learning more interactive and engaging, and it can help learners to focus on the topics and concepts that are most relevant to them.
In addition, ChatGPT can be useful for generating learning content that is accessible and inclusive. Because ChatGPT can generate text in response to user input, it can be used to create learning materials that are tailored to the needs and abilities of different learners. For example, ChatGPT can be used to generate learning content that is written in plain language, that uses alternative formats, or that is available in multiple languages.
Overall, ChatGPT has the potential to be a valuable tool for generating learning content. It can help educators and trainers to produce high-quality, accurate learning materials quickly and efficiently, and it can provide learners with a personalized and accessible learning experience.
This is a promising response from ChatGPT, but it also requires plenty of feedback from L&D experts to improve. L&D experts interacting with ChatGPT, or another AI-driven chatbot, will be valuable in the creation of high-quality learning content in a very short period. This would not only be cost-effective but also support L&D to stay at the speed of the business. Let’s continue the experiment and read the response to next question.
“What are the limitations to generating learning content?”
ChatGPT:
There are several limitations to using ChatGPT for generating learning content. First and foremost, ChatGPT is a language model, which means that it is trained to generate text based on the input it receives. However, it is not specifically designed to generate learning content, so its outputs may not always be relevant or accurate.
Another limitation is that ChatGPT is not able to browse the internet or access external sources of information. This means that it is limited to the knowledge and information that it has been trained on, which may not be comprehensive or up to date.
Additionally, ChatGPT is a “black box” model, which means that it is not transparent in how it generates its outputs. This can make it difficult to verify the accuracy and relevance of the content it generates and can also make it difficult to customize the model to specific learning needs.
In conclusion, while ChatGPT may be able to generate text that is relevant to learning content, it has several limitations that make it less than ideal for generating high-quality learning materials. These limitations include its lack of specialized training, its lack of access to external information sources, and its lack of transparency.
A logical next question for bloggers in the L&D sphere who want to use ChatGPT is: “What are the limitations of using ChatGPT for creating blogs?”
ChatGPT:
While ChatGPT and the GPT-3 language model it uses can be useful for writing blogs, there are some limitations to be aware of. One limitation is that ChatGPT is not a human, so it may not be able to write in the same style or voice as a human writer. This can make its writing feel less personal or engaging, and it may not be able to capture the same depth of emotion or nuance that a human writer can.
Another limitation of ChatGPT is that it is not perfect. GPT-3 is a powerful language model, but it is not perfect, and it may generate content that is inaccurate, irrelevant, or nonsensical. While the quality of GPT-3’s output is generally high, it is not guaranteed, and users should be prepared to edit and revise the content it generates.
In addition, ChatGPT is not a replacement for human writers. It is a tool that can be used to assist with writing, but it is not a substitute for human creativity and insight. Bloggers who use ChatGPT should still use their own judgement and expertise when writing and editing their content and should not rely solely on ChatGPT to generate their blog posts.
Overall, while ChatGPT and GPT-3 can be useful for writing blogs, they are not perfect and should be used with caution. Bloggers should be aware of their limitations.
Let’s end this experiment with a killer question for ChatGPT: “Are learning styles based on scientific evidence?”
ChatGPT:
There is no scientific consensus on the concept of learning styles. Some researchers argue that learning styles, as typically conceptualized, do not accurately reflect the way people learn, while others maintain that they can be useful in certain contexts.
The idea of learning styles is based on the notion that individuals have unique preferences for how they learn new information. For example, some people may prefer to learn through visual means, such as by seeing diagrams or pictures, while others may prefer to learn through auditory means, such as by listening to lectures or discussions.
However, many researchers have criticized the notion of learning styles, arguing that there is little evidence to support the idea that people have unique learning preferences. They point out that most studies on learning styles have been flawed, and that the results are not consistent or reliable.
Moreover, some researchers argue that the focus on learning styles can actually be detrimental to education, because it can lead teachers to treat students as if they all have the same learning preferences. This can result in a “one-size-fits-all” approach to teaching, which can be ineffective for many students.
In conclusion, while the concept of learning styles is widely accepted and commonly used, it is not supported by scientific evidence.
A helpful informal response if you are familiar with the academic background of this continuing discussion in our field, such as this overview by Jane Bozarth, this short movie from Guy Wallace, and so on. If not, it requires deep knowledge and effort to validate the response in an evidence-informed way.
Reflections On Using The AI Chatbot
ChatGTP is not a magic bullet for L&D in generating learning content or solving problems. However, the potential of these algorithms is too promising to be ignored by the L&D community.
My experience with ChatGPT was cool. It is an extremely fast content generator and can support the L&D community with the creation of learning content. Creating content is the first step. Validation of the content is the second and more significant step:
- Check the content for errors, nonsense, mistakes, and illogical reasoning.
- Validate the content with Subject Matter Experts.
- Validate the (scientific) resources, since there is no transparency about the sources in ChatGPT (at this moment).
- Check the content for our professional standards (including evidence-informed designs of learning experiences).
L&D professionals working with ChatGPT need to check and double-check the input and output. L&D professionals aren’t displaced by ChatGPT, but this tool can alleviate some of your work if used responsibly, with the right intentions, and in accordance with L&D professional standards. The L&D community is facing the challenge of supporting the improvement of these algorithms.
Now is the time for L&D to learn how to effectively use ChatGPT to get assistance in creating learning content within seconds, solving (complex) problems, and so on. The rule is simple: the better the input, the “probably better” the output. “Probably better” should be replaced by “better” with the support of the L&D community. The challenge: is the learning profession able to adapt fast enough to use ChatGTP, as organizations will leverage ChatCPT to drive their business?
It’s like the old days in the airline industry. We must work hard and smart to improve ChatGPT (or similar chatbots). As with pilots, L&D professionals should also possess a very strong level of expertise to provide the right input and feedback to improve ChatGPT and add professional, evidence-based knowledge and practices. But, unlike the beginning of aviation, we will not end up in the hospital if anything goes wrong…
