Conversational AI has taken the world by storm in the last few months and is already a disruptive force across many industries. As an industry, we are just beginning to understand how AI will transform skills development and validation.
On the one hand, AI can act as a force multiplier, reducing the quantity of things learners must know and retain.
On the other, it provides a set of tools that enable us to better develop and more effectively, efficiently and accurately validate skills.
Corey Hynes
Executive Chairman of Skillable
Our foundational belief, what drives us every day, is found on our History page.
Skilling is the future of training and it’s hands-on. Experiences will be supplemented by videos and instructors, not vice versa. Challenge-centric learning is the next instructional design model. Scored labs and shared labs will innovate hands-on learning and we’re leading the way. Performance testing is the future of certification and badging.
The new generation of AI tools based on platforms such as ChatGPT, Bing, Bard, CoPilot and more only bring us closer to realizing that belief. As we collectively begin to adopt new AI tools, here are a few ways digital skilling, and labs in general, are evolving for the better.
It is easier to build labs that validate skill.
One of the greatest uses of AI is its ability to provide a first draft of anything. For Lab Developers, this is particularly useful. Developers can generate very usable first draft lab documents complete with instructional text, step-by-step procedures and even assessment questions.
In some cases, these drafts have been good enough to go directly to our test and QA teams. For cloud labs such as Azure, AWS or GCP, we have successfully generated useable deployment templates, security policies and configuration scripts. This has cut significant time and cost from the lab development process, as well as reduced the skill set required for a Lab Developer to get started. More people can build better labs – in less time.
Today, a simple “step by step” lab is not good enough.
Today we need labs that not only provide instructional context but ensure learners can take skills developed in the lab and apply them to real-world scenarios, while providing the organization evidence of that skill. Scored labs are the tool to accomplish this, however creating scored labs can be challenging.
This is where AI helps the most.
Used correctly, AI through creative prompt engineering can provide a very usable scoring script in seconds. Scoring is no longer reserved for the most complex lab written by the most qualified authors. Any author can use any lab to validate and generate evidence of skill.
Having used AI to create labs over the past several months, I have found the time to develop scoring scripts is reduced drastically – up to 90%. What took hours now takes a minute. Developers don’t need to write the script, they can simply ask for it.
The skill required to build a scored lab is reduced significantly.
If you can read script and modify a few lines, you can be successful. You do not need to write the logic or worry about structure, syntax or documentation. Let AI do that for you.
The amazing thing is you can do all of this today with AI. Microsoft and GitHub just announced CoPilot for business. A code-specific AI, positioned as the best AI code generator yet. You can use ChatGPT and Bing to create all the samples and starter templates mentioned above right now. We have Skillable team members doing this every day and are realizing how powerful these tools are.
What’s next for AI and Skillable?
If you are building labs or have a desire to evolve to scored labs that provide evidence of skill, keep an eye out for a few things.
- Join our Lab Developer Workshops. AI will be a topic in upcoming Skillable workshops. Get practical tips, advice and guidance on how to leverage AI as part of your hands-on content creation and training programs.
- Keep an eye out for platform enhancements. We have a number of platform enhancements in flight to bring AI to the forefront of lab development and delivery. We are building tools to introduce AI into the development of scoring scripts, lab instructions and more. We are also extending our learner experience, providing AI-based instructional assistance, mentoring and support.
“Proof of work” is more important than “proof of knowledge.”
Without getting overly academic, most validation/certification questions are based on frameworks like Bloom’s Taxonomy. These models are used heavily in learning applications to design programs and assessments that move a learner beyond knowledge and into analysis and comprehension.
For example, the higher your assessment question is on Bloom’s Taxonomy, the “better” it is. Even so, you are asking a candidate to read, comprehend, interpret and analyze information, and then decide on a course of action (proof of knowledge). You are not asking to execute the decision or perform the task (proof of work).
Some applications of AI are beginning to devalue proof of knowledge as a form of assessment. There are numerous examples of AI passing law, business and medical exams. If you are building a skills validation program and looking to collect verifiable evidence that a person possesses a skill, collecting proof of work by having them perform tasks that generate evidence becomes exponentially more important.
This evolution underscores the need to evolve beyond labs that are focused on completing a process and focus on labs that challenge a user to demonstrate skill and then provide you with evidence of that skill. It furthermore increases the need for hands-on assessments to be an integral part of the entire employee lifecycle. From pre-employment skills assessments to ongoing skills validation to ensure employers have verification that employees have the right skills before they take on larger responsibilities.
I’m fully invested in the notion that “doing” will replace “understanding and remembering” as evidence of skill. I’m convinced that advancements like conversational AI will break down the barriers to building what are inherently more complex items and that advances in technology will make them available to more people, in more formats. I think the next two years will see massive evolution in the way we produce evidence of skill and ultimately certify that evidence with proof of work trumping proof of knowledge.
Instructors can be elevated from educators to coaches.
This is where we will see the biggest direct impacts of conversational AI in the skills development and validation space. In the past, many platforms have implemented some form of simple chatbot. Something that could take keywords from questions and perform a search or provide an AI-driven answer.
This alone is a compelling path of evolution for an Instructor. As AI tools become more conversational and have access to more information, it presents an opportunity for Instructors to redefine the role they play in skills development.
Imagine a learner being able to interact with any instruction, goal or objective in a scored lab, or any task in a Skillable Challenge, and learn from AI in real time. Then imagine being able to interact with AI with phrases like “I still don’t understand” or “Tell me more” to further explore or summarize the topic.
Imagine if that experience was situationally aware, knew what kind of lab you were taking, on what topic, and which topics you have already trained on.
Now what if we could look at patterns over time, of the topics that people “asked about” and dynamically change the content of the lab to personalize it and focus on your weak areas, while skipping your strong ones.
What if the experience you have in the lab is akin to sitting with a tutor who watches you do things, gives you tips where you struggle and explains things you don’t understand, fully personalized based on your existing skills and experiences.
Skillable: Harnessing the power of AI for better, future-proof skilling.
At Skillable, we are working towards a future where AI is a fully integrated, transparent part of the learning experience.
With only a few enhancements we will bring conversational AI to the forefront of the learner experience in a lab, while simplifying and streamlining the work of the author.
This is where our excitement lies.
Conversational AI allows us to produce ten times the content in half the time, and our tools allow us to choose which content to display and when, in real-time, based on behavior of the user and characteristics of the lab.
The result is labs that start with a simple framework but dynamically evolve over time so every learner’s experience is unique.
100% dynamic, 100% personalized, 100% interactive, 100% validated skilling experiences.
