I mentioned in yesterday’s post that one thing I do in getting objectives is focus on decisions. And, simple ones will get automated; we can train AI to handle these. What will make the difference between ordinary and extraordinary organizations is the ability to make decisions in this new VUCA environment (volatile, uncertain, complex, and ambiguous). And it made me wonder how you develop the ability to make better choices.
AI can be trained in a couple of ways to answer questions and make these decisions. We can use machine learning to train a system on a historical database (watching out for bias). We can use semantic analysis to read documents and make a system that can answer questions about them. But such systems are very limited; they can’t handle questions at the periphery of the knowledge well, and fall apart at related areas. Which people are better at, if their expertise has been developed.
Now, developing this expertise isn’t straightforward. If there were simple decision trees, we could automate them as above. Instead, what works best is expert models that have been abstracted across dialog and practice. This needs to be augmented with an awareness of adjacent fields. So, for instance, for instructional design, we should have an awareness of interface design, graphic design, media production, etc. So how do we develop this?
We certainly need to develop the expert models we know play a role. But this gets circular with the above unless we find a way to break out of the predictable. I suggested one approach to this with my ‘shades of grey’ post, having groups work together to make categorization choices: is this or is this not legal. This was, however, more focused on compliance and there’s a much wider situation.
We first need to identify the situations, the relevant models, and the scope of likely variation. We can’t provide specific data (or we’d train the system on it), so we need to anticipate a spread. And we could just train that, but I want to go further.
I’d want to use such a process to choose situations, and then design group work, for the reasons I identified here. (Resourced with models and examples, of course.) We want to get learners working together to address complex problems. We want them to use their various understandings to illuminate the underlying models. If you can get productive discussion (and this needs to be designed in and facilitated), the learners’ thinking will be enriched. (And they may have folks to call on when the situations do arise ;).
Collaboration in learning is second best to collaboration in problem-solving. We should do the latter when we can, but we should do the former anyway. For better learning, and for those times when there isn’t the luxury of working with others.
I reckon this would lead to better decision-making ability. What do you think?