News

How can AI be used within learning and development to enhance output?

How can AI be used within learning and development to enhance output?


Daryl Hedley, a Project Engineer at Kwantic, explores the transformative impact of artificial intelligence on Learning and Development (L&D) and its potential to elevate the learning experience.

 

How has AI changed the L&D sector?

AI seems to have created either excitement or anxiety amongst people I know – both for good reasons. While it may not have changed too much right now, the potential of having AI in the L&D sector could be a game changer; I anticipate a significant increase in the integration of tools and the presence of generative content. For us at Kwantic, our main goal is to use AI as an assistant. What we mean by this is to allow the AI to enhance the HI (Human Intelligence). AI is great at doing tasks – especially ones that can take much longer when done by humans. By taking these tasks away from humans, we can channel our energy towards the more creative aspects.

How can AI be used to personalise learning and make it more accessible?

When we consider the tedious or time-consuming tasks, AI is so good at getting it done. While it’s not consistently accurate, it generally does a good job. This is why we’ve gone with an assistant approach, with the human always checking the AI output.

At Kwantic, we’re always looking at ways to personalise by envisioning beyond the scope of a traditional course. We believe the power comes from multiple sources that are all connected – and it turns out that AI is incredible at doing this. AI possesses the capability to match content, link it, apply skills tags and then, with a tool like Kwantic, you can have it displayed to the end user in an engaging way.

We’re also exploring diverse methods of content access.  Considering an AI that can link to content, comprehend it, and provide an engaging search and filtering experience, could the future lie in an assistant guiding you to discover the next relevant content and seamlessly transitioning you to related pieces?

How can AI support data analysis?

I believe we’re merely scratching the surface of AI’s potential in data analysis. This appears to be the most significant realm for enhancement in the L&D field. Right now, we are stuck in the era (or past multiple era’s) of SCORM. We need to move away from it – and maybe AI is the key to unlocking this. If we’re able to push more analytics data out from our learners’ experiences, we can use this to better our own content creation, find missing gaps in knowledge, and identify to the end learner where they can potentially improve. I’m truly excited about this aspect of AI and L&D!

 

Can AI be used to automate content creation?

Right now, no. But there is potential for it to happen in the future – especially for more basic or generalised content. We’ve seen tools that are already able to do this, but the results are mixed. However, when aiming for distinctive experiences that captivate individuals through imaginative methods of presenting media and narratives, the ideal approach involves human involvement alongside an AI assistant. The numerous specialised distinctions among companies or products means a human will likely always be involved.

What AI could bring to the table, is a way for learners to engage with content creators. It’s always felt one-way when creating content, even though we spend some time upfront looking at initiatives, finding the problems and working out how internal businesses work before creating our content. Once the content is created and consumed, it’s easy to see this as a job done. But we should be using AI here to ask questions about how to improve the learning from the learner’s perspective. Maybe we missed something. Could AI help identify this via a conversation with the end learner?

 

Are there any limitations or challenges with AI in learning and development?

There’s a lot of talk about hallucinations  – which is probably the easiest to overcome by having a human look over the content. Hallucinations in AI refer to instances where the AI system generates inaccurate, unrealistic, or nonsensical outputs that do not accurately represent the input data or real-world context.

I’d be more worried about how the AI models have been taught and what that content contains. Only the last few decades have seen real change in the world around racism, sexism and equality. Yet, these AI models have been trained on data that predates this. There have been studies to show that AI can act in a racist or sexist way and with many people battling for these causes, AI has the potential to take us backwards. Once again, my inclination would be towards a more cautious approach, ensuring rigorous fact-checking and verifying that any output from AI is checked for equality.

 

Can AI help small L&D teams?

Yes. The way we see AI as an assistant can really help smaller teams by taking the load off them. However, this should not be interpreted as a mere acceleration of processes. We’re working smarter and this opens up the door for more creative work. Our view at Kwantic is to remove the 95% of content input down to 5% so we can focus on more engaging and story-based content. It’s important to remember that L&D has always had a feeling of always moving fast.

 

With AI, we have the potential to really dig into the strategy around learning initiatives, build more engaging content that can be personalised and use creative marketing campaigns to engage and enhance learning experiences. I believe that we’ll start to see the next generation of tools that become assistive to smaller L&D teams that enable them to create truly wonderful experiences that feel like something only a larger company would have had the budget for. This levels the playing field, and it comes down to human creativity.

 

Is it too early to integrate AI into some learning modules?

This is a hard question to answer. The context here really matters. For quick question and answer style types of integrations, AI is great. But the context must be just right, and the prompts need to be tested thoroughly. As we engage with this new technology, we should really be tracking the inputs and outputs of both the learner and AI to see if the conversations are still on track (and not experiencing hallucinations). I struggle to see this happen so quickly with SCORM as the defacto tracking standard. But those who are open to trying something more innovative could see something come together faster and more beneficial to the end learner.

Explore how Kwantic can help you unlock ways to elevate your content creation here.

Share this entry

Straight to your inbox

Be the first to know the latest news and insights about the world of L&D by signing up to our newsletter.