Harvard psychologist and education professor Howard Gardner recently shared his view on how AI could change education. He believes cognitive tasks will become optional for humans, freeing people to pursue their interests. The actual interview between Rick Hess and Howard Gardner is certainly more nuanced than just the statement and makes for an interesting read.
While I can't say I agree with his bold statement of cognitive tasks becoming optional for humans, what is true is that when all our lives, humans have defined their intelligence based on cognitive tasks, and then, AI takes over some of those cognitive tasks, we are suddenly left thinking more deeply about what we thought was exclusive to us... our intelligence!
But I think the real leap in this discussion is to stop making these lists of "things human can do that AI can't" (because, realistically, that list is shrinking by the hour). The more important task is to start questioning systems that were designed with the assumption that producing something is the way to demonstrate intelligence. I mean systems of education, training, assessment in school and at work that rely on writing an essay, submitting a portfolio, completing a presentation slide deck, making a report, etc.
In L&D, we have always done this. We use course completion, knowledge tests, work samples and portfolios as evidence that learning has occurred or that someone is ready to perform. But we know all too well that completing training isn't the same as being able to perform at work, just like producing the right answer isn't necessarily evidence of understanding.
The artefact was never the actual competency yet we believed for so long that it was. But when the artefact can be produced by AI, it becomes an even weaker proxy for learning, intelligence, and competence.
Having worked in the areas of competency development and assessment for close to 3 decades, I know too well that competence is messy. It involves recognizing what is happening, deciding what matters, and applying information in various contexts. Then there is exercising judgment, adapting when circumstances change and taking responsibility for the outcome. The artifact may provide useful evidence of those things, but it isn't 'the thing' itself.
The artefact as the evidence of intelligence is losing ground quickly. Especially given the recent development of AI watermarking, where Anthropic has begun watermarking text generated by Claude so that the content can be machine-identified as AI-generated.
So, where does the value of expertise sit when AI can produce so much of the work we have traditionally been paid to produce? We are all realizing what was always true but hidden in plain sight that expertise is beyond outputs and is generally more upstream.
In my field of instructional design, this is becoming increasingly obvious. AI can produce storyboards, graphics, video, narration, activities and assessments. It can turn a reasonably well-defined brief into a fairly complete learning product very quickly.
What is the value of an instructional designer? I don't think the answer is simply "in knowing how to use AI." I think it is increasingly in everything that happens before we ask AI to produce something.
I am glad I work with clients and teams who get this. I am currently working on an organization-wide performance improvement initiative where we are taking an end-to-end approach to needs analysis, looking carefully at the system rather than responding to a request for training. We are examining the work, people, processes, tools, expectations, resources and environment, and working through the findings collaboratively. We are so focused on not hurrying towards the output and it is taking some discipline and intention to do it this way. I mean, when we can produce learning in minutes, it is rather tempting to start producing! But the more capable the tool, the more important is the role of the human to spend time figuring out what actually needs to be produced.
Because AI makes the “doing” easier, the value of human capability is more upstream in knowing what to ask, what to pursue and why it matters.
Perhaps we need to get much better at assessing the things that are harder to produce on demand and are certainly harder to outsource to AI.
Perhaps we have to redefine what now constitutes human intelligence and what makes us competent.
Perhaps it is about things like understanding the context, tapping into the lived experience, human judgment, and being responsible and accountable for our choices and actions.
This binary view of AI vs. Human just won't cut it anymore. This isn't as simple as AI on one side and human expertise on the other. The opportunity staring at us is in knowing how to combine the two. The shift from producing answers to identifying what problems to solve is the real discussion. Not the how but the why.
But it was always about the why, wasn't it?
No comments:
Post a Comment