AI in Learning & Development
Ideas on how to think about AI in the context of Learning & Development
In my current role as product consultant, supporting the British Council in an exciting journey of innovation in the language teaching, learning and assessment space, I recently chatted to British Council Corporate English Solution’s Head of Services & Talent, Kate Sullivan, about AI in the context of Learning & Development.
Kate and her team support businesses, higher education institutions, government, and NGOs with solutions and insights in all things learning and development.
Common questions, that get asked these days:
Should we adopt AI? What can AI bring to the table? How do we adopt AI? What are the risks and challenges? What will the future bring?
Here are some of my answers to Kate’s questions:
What one trend do you think shaped learning in 2024?
With our clients we have seen a major interest in task-focused skills. Skills that allow solving complex problems and challenges, that involve creativity, human interaction and collaboration. Soft(er) skills, if you like.
I have also observed a shift away from traditional classroom-based learning towards personalised coaching and mentoring. At least a wish for that; I’m not convinced that the market has solved that problem sufficiently well quite yet.
Recruiting for “my” (product) team, I have seen a shift away from knowledge to creative and strategic thinking, and communication. In a nut-shell: not how you code, but what the right code for the right problem is.
What key changes have you seen in learning and assessment this year?
A clear focus on cost efficiency and scale, and with that an AI sticker on many products. I’m not sure though that all these products actually deliver on the AI-promise the advertise…
What advice would you give to L&D professionals implementing programmes during change?
Because of my last point: Don’t jump on the hype.
Many client organisations I speak with ask me what they should do with AI. My response always is: what problem do you need solving?
Once we can answer that, then we can talk about a) whether AI helps address this problem and b) whether the organisation has the stamina and capabilities to adopt AI. Because — certainly for now — AI, beyond the most basic use cases, requires quite some investment for organisations in terms of up-skilling, policy and governance, and implementation. Also, dependent on use case, there is one should expect considerable! effort for sourcing and curation of data, model training or tuning, validation, etc.
AI will only work well if it is ‘fed’ with high quality data during training, tuning and operation. Making high quality data available is not an easy undertaking.
Right now, the answer in many cases to first question is “No, AI doesn’t solve the problem”. If the answer is “Yes” then, frequently the answer to the second question is “No”.
And that’s not a disaster, AI is in it’s early stages, most organisations are well-advised to tread carefully.
In fact, I want to clarify that the term “AI” is a bit of a red herring. We shouldn’t think “AI”. We should think in terms of simple and complex algorithms that might be useful.
I am only negative “AI” in the sense of it being an over-hyped solution category, positioned uncomfortably close to snake oil by many of the thought leaders in that the. I am, however a big fan of using algorithms to achieve better outcomes…
A minute, but very important distinction.
How have AI and technology impacted learning and assessment?
As I mentioned earlier, as with all hypes, ‘AI’ has been added to far too many product names, frequently without substance or need. (Remember blockchain? Not quite as bad, but similar).
On the other hand, AI has opened a number of doors for efficiencies and scale: speech recognition (for assessment), text to speech and LLMs (for interactions), genAI for item generation, ML and NLP in general for assessment.
But we mustn’t forget that all of these are still quite emergent. So my team treats all of these as additional ingredients, not as replacements. Certainly not as replacements for humans. We also don’t deploy into high stakes use cases, for ethical reasons (and also for regulatory reasons).
What tips do you have for integrating AI into L&D practices?
- Find the sweet spot where AI can really solve your problems better and or more efficiently than you currently can.
- Off-the-shelf systems are less likely to make a real difference. To gain real value from AI you’ll have to ‘train’ or ‘fine tune’ the system, to make it contextual to your ‘problem’. Don’t be tempted by what this technology might be able to do for your organisation in an ideal scenario. Instead focus on clearly articulating desirable outcomes in actionable terms and understanding what conditions and resources (e.g.) you need to put in place. For instance, can you sufficiently well articulate desirable learning goals? Or, if you wanted to leave that to a model, do you have the data to train a model?
- Don’t underestimate the effort to implement an AI based system.
What do you see as the role of learning and development in championing diversity, equity and inclusion (DEI)?
We frequently hear how problematic AI can be in regards to bad goals or bias. And while that is true, AI can also help democratise learning and broaden access.
So if we, as an industry, get DEI right and make these tools available widely, then we can contribute to a global increase in skills and knowledge.
I often speak at conferences about ethics — I always emphasise that our knee jerk reaction is to keep knowledge to ourselves or within our organisations. However, history has shown that a highly educated global workforce means we all all better off….
What practical steps could L&D take to advance DEI in their initiatives?
I also speak about compliance a lot at conferences. I am a strong believer of continuous compliance. I.e. compliance that is part of the process from day one, where ethical and other compliance concerns are clearly articulated, regularly revisited and enacted by cross functional teams at every step of the process.
For this to work, the first step is to have a clearly articulated, reasonable, actionable stance on DEI.
What are the main challenges L&D has faced this year?
My wife is a meditation and yoga teacher, and from her I hear that stress and anxiety are a key factor. So I wonder whether including coping skills in L&D isn’t obligatory.
What strategies do you believe are most effective to support well-being and resilience?
I have found yoga and Buddhist meditation techniques helpful — be this breath work, mindfulness practices or gaining insight into behavioural patterns….
Looking ahead, what do you see as a crucial trend that will influence L&D in 2025?
On the (technology) solution side I will be watching what happens with personalised AI coaches, auto content generation, and automated assessment. Chatbots, while not quite there yet, are a solved problem and we’ll (only) see (incremental) improvements.
As a result of the interest in and need for tech, we should expect highly skilled (and more expensive) talent coming from locations that are currently — generally — less well-educated and cheaper.
Thanks to British Council’s Kate Sullivan for being interested in my take on AI and L&D. If you have any questions on how the British Council can support your organisation, contact Kate Sullivan or myself, or visit their website. Also, an excellent free overview by the British Council about AI adoption in this space can be found here:
