David James is CLO at 360Learning, host of The Learning & Development Podcast and former Director of L&D for The Walt Disney Company.
getty
There is a lot of discussion right now in learning and development circles about the potential—and deficiencies—of ChatGPT. But despite criticisms of its inaccuracies and "lack of humanity," these will not prevent it—or its imminent successors—from challenging the field of instructional design.
This may seem alarmist and like a knee-jerk reaction to a product that only launched at the very end of November 2022—but hear me out. Near-future iterations of ChatGPT—and its like—could offer a more comprehensive solution to the needs of L&D leaders than decades of instructional design.
What Is ChatGPT?
ChatGPT and GPT-3 are different in that the former is the application and the latter is the machine learning model that the application uses. It can be difficult to describe without a technical explanation and is easily misinterpreted in comparison to technologies we’re more familiar with—which is a note of caution. As explained in TechTarget:
“GPT-3, or the third generation Generative Pre-trained Transformer, is a neural network machine learning model trained using internet data to generate any type of text. Developed by OpenAI, it requires a small amount of input text to generate large volumes of relevant and sophisticated machine-generated text.”
The consequence of this is: “GPT-3 is better than any prior model for producing text that is convincing enough to seem like a human could have written it.”
Some examples of what this technology has been used to create include articles, poetry, stories and news reporting.
ChatGPT responds to the text that a user enters into it, which may usefully include summaries of large amounts of text, and can even respond in code for software programmers.
Why Instructional Designers Are At Risk
What’s always frustrated me about instructional design is the low barrier to entry. Anybody with access to an authoring tool can create a course, and L&D leaders have been filling courses with mediocre content and facile tests for decades. What does it take to be "knowledgeable" in a field? Reading one book and a handful of articles? I’ve witnessed e-learning modules based on wafer-thin "expertise" and learning tech platforms in organizations around the world crammed full of substandard content. Arguably, ChatGPT could already challenge run-of-the-mill learning content production by responding to specific questions from users about how they might respond in specific situations. Learning content providers beware!
So AI consolidating ideas and models in order to outline and feed learning resources is disruptive enough for instructional designers everywhere. But what I’m really excited about is ChatGPT’s ability to be trained.
So much learning content may pass a quality test for up-to-date thinking but remains remiss of any organizational context, let alone the context of departments, teams and roles. For many in L&D reading this right now, the natural response could be to ask: "It’s okay saying it, but how exactly does a small team with limited resources actually do this?" And this is where ChatGPT (or its successors and competitors) could be truly transformative.
ChatGPT does not trawl the internet and consolidate relevant sources; it pulls from data that it has been trained on in order to generate its responses. This data includes an enormous array of texts from different sources, including articles, books and websites. But imagine training it to understand the nuances of your culture, your customers, interpersonal and team dynamics, your products, services, processes, technologies—and your implicit expectations. Then imagine how quickly and usefully it could answer questions, advise and guide people to do more of the right things to achieve their desired outcomes.
All of a sudden, anything below exceptional and bespoke instructional design seems superfluous. Any arguments about it not being human enough are inconsequential compared to its ability to respond to the unique context of your organization for the benefit of those who need to interpret it.
‘Anything’ Was Never Enough And It Certainly Isn’t Now
ChatGPT doesn’t have to be better than the best instructional designer to disrupt L&D; it just needs to be good enough to compete with the oceans of mediocre courses and content creators. This is what makes this exciting technology such a big wake-up call.
So many conversations in organizations go like this:
• Employee/line manager/stakeholder to L&D: “Do you have anything on time management?”
• L&D: “Sure, let me send you a link to the content we have on time management.”
What’s so important about this short interaction, you may ask? It’s as common as the sun rising every morning. Well, the most important word in all of this is "anything." When somebody asks us if we have anything on a topic, that’s a very low bar for relevance. People aren’t used to finding exactly what they need from L&D because we’ve never believed it’s actually been possible. So in the absence of true relevance, "anything" will do.
But what if an AI chatbot learns how to interpret our organizational context and learns more and more from feedback over time? Will "anything" do six months from now? No. That’s because we have higher expectations of what we, in L&D, are capable of and the value that brings to employees, line managers and other stakeholders.
Answering questions and creating useful resources from its wealth of sources is just the beginning. AI chatbots being trained on what makes our organizations unique, and how that applies to specific roles and tasks, is a whole new ball game. As a field, L&D could realize its potential and provide organizations with skilled employees ready to perform in new roles, now and in the future. This could offer a true answer to the upskilling and reskilling priority that links technical capability with cultural nuance, guiding and supporting employees to perform the tasks and roles the organization requires of them with confidence and competence.
So should all instructional designers be worried? No, not immediately. But the huge number of mediocre and subpar designers should be.
Forbes Human Resources Council is an invitation-only organization for HR executives across all industries. Do I qualify?
