Creativity and ethics: Human centric AI for learning

I don’t know about you, but I’m spending a lot of time learning about and playing with Generative AI.  I’ve been impressed by the technology’s ability to create content, such as text, images, audio, and video, from simple prompts.  Everywhere I turn it seems there are new ideas and applications being thought of.  I trained as an engineer and computer scientist originally, and it’s impressive to see how far the technology has come since I qualified.  I’ll go so far as to say I’m an advocate for AI, and I’m interested in seeing how it can help us reach our potential in society and business. 

I’m particularly interested in how GenAI can help in the field of learning and development, and I’ve been spending quite a lot of time researching this and experimenting to see how far we can push the frontier.  Every day there are more examples from our creative colleagues, using it to accelerate content creation, helping design and deliver better learning experiences, and even embedding it in the learning itself as a way to help reinforce learning through simulations and digital coaches.  

Personally, I think we need to make sure we use the technology in a ‘human centric’ way, as David explores in his post “How do we make AI more human centric?”.  My dad was one of those who would have said “never invent a lifeform that is smarter than you are”, and while I think we are still a while away from Applied General Intelligence, the point about making sure the technology was there to help, not replace or undermine is well made.   My colleague Robert Bolton made this point equally eloquently in our 2016 paper “Rise of the Humans – the integration of digital and human labour”

But how do you use it safely, ethically, and effectively in your job? And what does it mean for your role and your skills? 

I thought it would be helpful to share some of what I’ve discovered, as I continue my own journey in this area.  My big takeaways so far are:

1. Building a culture of curiosity: This field is moving fast, and we’re still very much exploring the art of the possible.  By the time you read this paper, it will probably be out of date.  So taking time to get people aware, curious, and conscious of safe usage is essential.  As L&D professionals, we all want our organisations to have a Learning Culture, and this applies to AI too.
2. Skills for AI: It’s important to develop clear views on what skills people will need in a future where GenAI tools are widely adopted.  This isn’t just about training more data scientists.  It’s going to change the skills you need to do lots of different jobs, and that means the learning needs to change too.
3. Avoiding AI Amnesia:  We need to teach the teams who are deploying these amazing tools how people learn.  If we don’t we’re missing a trick.  Lots of people are writing prompts that inadvertently introduce barriers to learning, and encourage ‘forgetting’ rather than ‘remembering’. We need to help unlock the skills potential of our people.

Here’s some highlights from what I’ve learned so far:

GenAI is already having an impact on learning and development

There’s a lot going on, and here are my ‘top four’ areas where GenAI is already having an impact on learning and development:

• Levelling the skills gap. It’s probably the most obvious point, but it’s worth saying that Generative AI is already helping many learners close their skills gaps, without meaning to do so.  I quite like Ethan Mollick’s perspective which shows how GenAI is a “Skills Leveller”.  Having a tool that knows pretty much everything on the internet at your fingertips means there’s much less call to ‘ask an expert’.  Out of the box, it can help learners access skills such as research, writing, analysis, coding, or design skills. It can also help learners access and understand information that may be complex or inaccessible otherwise, such as scientific, legal, or technical information.
• Boosting your productivity and creativity. Generative AI can help you create high-quality content faster and easier, saving you time and resource. I often use it to speed up my preliminary research instead traditional manual research techniques, and I see tools like Perplexity.ai starting to get popular in education too.  Personally, I also use GenAI tools like Microsoft Copilot daily, to challenge my own thinking. I find it can inspire me to think differently by offering new perspectives, acting as a bit of a digital buddy to give me new ideas.  I even use it to create graphics for me: I fed this paper into an AI image generator and asked it to generate the hero graphic at the top, which is a trick I learnt from Ethan. Thinking to the future, with tools like Synthesia, Pika, and HeyGen then AI generated video is closer than we think.  It’s also important to be mindful of potential copyright infringement though, a hot topic as I write this paper…! 
• Helping learners feel more engaged and motivated.  We all know that learner experience has a direct impact on engagement, and consequently on the impact of learning.  Generative AI is being used to provide personalized and adaptive learning experiences, tailored to the needs and preferences of each learner. It can also create immersive and interactive scenarios, such as simulations and games, that can make learning more fun and engaging. 
• Improving your learners’ outcomes and performance. Generative AI can support learners throughout their learning journey, providing feedback, guidance, and reinforcement.   There are also some really interesting applications on enhancing tutor / facilitator practice (as summarised in this article from Stanford).  And of course, there is a lot of debate on the role of AI in measuring and evaluating learner progress and performance.  I do feel that’s a particularly risky area, and it’s good to see that emerging regulation such as the EU AI Act is setting some parameters on what is acceptable use.  

But there are some tricky bits for us to get right too

So while we get all excited about the future potential of the areas above, and more – what do we need to keep an eye out for to ensure we are doing things the right way? 

• Managing risks and ethical dilemmas.  You’ve probably heard the term ‘hallucination’ in relation to GenAI by now.  If you haven’t, this is describing the instances where GenAI creates content that is inaccurate, misleading, or even harmful. It can also infringe on the rights and privacy of individuals and groups, such as by using their data without consent or using intellectual property without permission (and some early lawsuits are already being filed on the subject). You need to make sure that you use generative AI in a responsible and transparent way, having established safe usage guidelines, and following the rules and regulations, such as the recent EU AI Act.
• Still requiring human oversight and accountability. We’re already seeing examples of where people have just thrown tasks at a Large Language Model (LLM) and used the output without checking.  That’s like a chef failing to taste a dish before serving it – not ok!   You need to make sure that you keep an eye and a hand on the generative AI systems you use, checking and fixing their outputs, and stepping in and correcting their mistakes. It’s fine to delegate tasks, but you need to make sure you don’t abdicate your accountability.  Keep the human in the loop, as the popular phrase goes.
• Changing the skills landscape. I believe Generative AI is going to fundamentally change the skills landscape, creating new demands and opportunities, but also significantly reducing the need for some legacy skills to do some tasks.  Hence my first call to action: as learning professionals we need to be considering what skills our organisations will need in a future where GenAI is prevalent.  Not just skills on how to use GenAI tools, but how to work in a world with GenAI.  Critical thinking is high up on my list, as are data skills (including ethics and data protection), digital skills, and the more human skills of creativity and emotional intelligence.  And let’s not forget important areas such as AI ethics, safe usage, legal implications, and suchlike. And maybe the time is right to consider what learning should be retired from the curriculum, as it becomes less needed.  I suspect there are swathes of ‘awareness level’ learning products that may soon fit into this category. 
• AI induced amnesia? We all know that ‘doing’ is one of the best ways to learn.  But given the human tendency to follow the path of least resistance, I think there’s all too much risk that people use the technology to ‘do’, not learn.  If you surveyed a group of your learners and asked how they’ve been using GenAI so far, I suspect more will say “It wrote my report” than “I got it to review my report and offer suggestion to improve my writing skills”. The Learning and Performance Institute also wrote about this in their recent article on When AI meets Human, which comments on the risk of ‘Digital Amnesia’, and I think we need to be proactive about this.  So here’s my second call to action: as learning professionals, we need to be helping organisations identify the critical skills, and make sure the people building tools and writing prompts are doing so using proven pedagogical and andragogical approaches to develop skills.

So where should we start?

Generative AI feels like it is going to have a significant impact on the working world, and I believe it will have just as big an impact on learning and development.  But it also requires careful and thoughtful use, balancing the opportunities with the tricky bits, and making sure of the safety, ethics, and effectiveness of use. As learning and development professionals, we have a unique opportunity and responsibility to help shape how generative AI is used in our field and our profession.

One question I’ve been asked a lot recently is ‘where should I start?’.  The advice I would give is that we should be:

• Learning about generative AI. We need to educate ourselves and stay updated on the latest developments and trends in generative AI, understanding how it works, what it can and cannot do, and what are the best practices and standards for using it.
• Experimenting with generative AI. We need to explore and try out different generative AI tools and platforms, such as Copilot, ChatGPT, DALL-E, Gemini, and HeyGen, and see how they can help us create and deliver better learning experiences and outcomes.
• Collaborating with generative AI. In the main, I believe that GenAI is best deployed as a local productivity tool.  Each of us need to work out how we can best work with generative AI as a digital colleague, leveraging its strengths and compensating for its weaknesses, and integrating it into our working practices.
• Evaluating generative AI. Like everything we do in learning, it’s important that we measure and assess the impact and value of generative AI for our learners and organizations, using data and evidence to inform our decisions and actions, and to demonstrate our results and achievements.
• Influencing generative AI. We need to shape and guide the development and use of generative AI in our organisations, helping our businesses focus on the right skills based approach, and ensuring that it reflects our values and principles, and that it serves our purposes and goals.

Generative AI is widely in the public domain now, and will continue to see widespread adoption and evolution.  It is up to us to join the conversation on how we want to use it, and what we want to achieve with it.  So to mis-quote John F. Kennedy: “Ask not what Generative AI can do for Learning and Development, ask what Learning and Development can do for Generative AI”.