Utilizing AI to revolutionize how we approach traditional L&D challenges
Generative AI is changing the learning and development landscape. Age-old problems such as scaling individualized learning and engaging our people in continuous learning suddenly have new solutions. But the pace of change is fast and furious and there is much snake oil being presented along the way. In this blog, we’ll delve into what AI is, and perhaps more importantly what it isn’t, as well as explore utilizing AI as a learning and development professional.
When discussing Artificial Intelligence (AI), it’s best to start with what underpins it – and that’s data. AI needs huge swaths of information to train it. Without data, AI wouldn’t exist. We talk a lot at Learning Pool about intelligent data and what we mean by that is actionable data – data that impacts what happens.
The first level and key to everything is good data management. We need to capture, store and organize the data. We also need easy ways to visualize the data so we can see patterns, insights and even future trends. Once we have good data management in place, we can start to use the data in a variety of ways.
The next level is scripts. Scripts are a great example of actionable data. They are often built into learning platforms. But what are scripts? They’re really just very basic coding. For example, IF a learner completes a learning experience, THEN enroll them into this group. Or IF they are a member of this group or audience, THEN remind them they have access to this or might want to look at that. Pretty simple but hugely effective and a great time saver for learning administrators.
Slightly more complex scripts are often used for recommendations – for instance, user A liked this and so did user B, User A then completed this, User B might like it as well – recommend it to them!
On a much more complex scale, this is what the likes of Spotify and Netflix are doing when they give you recommendations. They are constantly capturing data both from users and their expert content curators. They then use a combination of machine learning* and deep learning** (which are forms of AI) to process and apply the data.
*Machine learning means that humans are training the AI model
**Deep learning means the computer is training itself by finding the patterns in the data provided
Key to the power of using AI in this way is that machines can find patterns that humans probably can’t. This is what makes them so effective and their recommendations uncannily accurate. More importantly, they can also do perform this in a fraction of the time that a human could meaning that number crunching is taken to a whole other level.
But let’s take a couple of steps back…
What is AI?
Artificial Intelligence (AI) is an umbrella term for computer software that mimics human cognition in order to perform complex tasks and learn from them. There are three key types: Traditional, Generative and General.
I’ll start with Artificial General Intelligence, but you don’t need to worry about this one as it doesn’t exist! This is where a machine would have an intelligence equal to humans in all aspects or even surpass the intelligence and ability of the human brain. Much science fiction which predicts the end of humanity is based on this but we’re a long way off and none of the latest developments gets us much closer despite the media frenzy that often tells us otherwise!
Traditional Artificial Intelligence, also known as narrow AI or Weak AI, is trained and focused to perform specific tasks. Weak AI drives most of the AI that surrounds us today, for example, Google Maps and the Spotify recommendations mentioned above.
Generative Artificial Intelligence goes a step further… by creating new data similar to its training data.
“Generative artificial intelligence (AI) describes algorithms that can be used to create new content, including audio, code, images, text, simulations, and videos. Recent breakthroughs in the field have the potential to drastically change the way we approach content creation.” McKinsey, January 2023
Generative AI has been around for a while, but its power only really became apparent to the majority with the release of ChatGPT. At the time, it was the fastest-adopted new platform, reaching one million users in the first five days and 100 million in two months.
ChatGPT is a Large Language Model (LLM). LLMs are a type of machine learning model that is designed to understand and generate human language text at a high level of complexity and fluency. They can perform a wide range of language-related tasks, such as text generation, language translation, sentiment analysis, text summarization and role play.
It is models such as ChatGPT that power most generative AI or at least the way we interact with them. For the first time, we can talk to a computer and be almost always guaranteed a ‘good’ response. This capability has allowed advances in other content generation, such as voice-to-text, text-to-voice, image and video generation and even language translation.
But generative AI is not only about text, there are also lots of media creation possibilities including images, videos, video editing, speech transcription and text-to-voice creation. And even writing code from natural language instructions.
Why is generative AI important in learning and development?
There are two key benefits generative AI brings to learning and development, efficiency and effectiveness.
As learning and development professionals we can use generative AI to help us speed up the learning design process but it shouldn’t be an excuse to create more content. We’re all drowning in a vast sea of content. Rather than mindlessly adding to that, we should consider how generative AI can not only save us time but make us better learning designers. It’s worthwhile starting by asking yourself…
What do we need people to do and how can we best support them to do that?
Generative AI can help you get off the starting blocks by working with you to draft objectives, summarize key points from lengthy documents and help you focus on the performance and behaviors that are needed and find the right learning experiences for those.
It can also act as a writing co-pilot and help you write questions, scenarios and activities. But don’t just take what it says as correct or the best way. You are the learning expert! To get the best from your Learning Design co-pilot you still need expertise, good source content, a subject matter expert or thorough research.
Finally, you can use it for media creation. You can create highly relevant graphics, talking heads or social media videos with little expertise, but again focus on the learning expertise and make every media asset have a purpose that helps your learners meet their goals.
This is where we can really improve what we do. Things are now possible at a scale that haven’t been possible before.
Personalization of content
Understanding a learner’s goals, behaviors and needs allows us to create personalized learning pathways or even create content specifically tailored to them and their situation.
People pay a fortune for the best personal tutors. Why? They can tailor their responses to yours and fine tune the learning you are given, even to the point of breaking it down and talking you through step by step. Generative AI can do this too! DuoLingo has used GPT-4 to create their Explain my Answer feature. It doesn’t just tell you you’re wrong – it talks you through the answer you gave and clears up misconceptions so that you’re far less likely to make the same mistake next time.
Coaching is a structured conversation. Using generative AI to ask learners what their goal is, where they are currently and then helping them to find ways forward is something a well trained AI bot can now do really well. The Khan Academy has done something similar with Khanmigo, their AI powered coach which can guide teachers through creating more effective lesson plans or even guide students in writing tasks. It’s not giving them the answers, it’s just nudging them along the right path.
Review of text or voice inputs
It’s been hard until now to truly assess open inputs or assignments at scale. Open input questions in online learning courses are often nothing more than an opportunity to give a model answer. Learners are often disengaged and simply hit enough keys on their keyboard to be allowed to progress. Now imagine how that might change if you know you’ll get proper feedback that takes your answer and compares it against an expertly crafted framework. The same technology could be applied to transcripts from call center calls or long-form assignments.
Enabling learners to practice the skills they are learning, whether it’s coding, writing or communication, has been difficult to perform online until now. But all of this is possible with Generative AI. Learning Pool has developed a whole new product that allows new managers to practice complex conversations by roleplaying with an AI team member.
Using Generative AI for skills practice at scale
The capabilities of AI Conversations
AI Conversations uses generative AI to allow managers to build their communication and management skills by practicing the types of difficult conversations they may need to have with people who report to them.
Users not only get feedback from the AI in the form of realistic, conversational responses, but they also receive detailed personalized feedback on how they performed – down to the specifics of the exact words they used and how they met certain pre-set criteria.
Does an A.I. Conversation work?
Over the last few months, we’ve been testing a prototype AI Conversation tool and we’ve found that even from just using it a couple of times:
- On average it increased confidence in having difficult conversations by 9.4%.
- 70% of the test cohort rated the tool as being high quality.
- 63% of the test cohort said they learned something new about handling difficult conversations at work.
To see an example of AI Conversations in action, watch this short video.
Got a learning problem to solve?
Get in touch to discover how we can help