Overview
Once common in data processing discussions, the mantra has recently re-emerged at GetSavi around content design using AI applications: rubbish in equals rubbish out. Despite how much AI has enhanced traditional course authoring applications, poor translation of source materials will lead to disappointing design results.
To use a different example, AI slop built on poor prompts for a marketing campaign will likely not give you the desired uptick in donations or new members. It will alienate your prospects.
To put a positive slant on this: much is said about AI’s unintended consequences, less is said about when it provides unexpected benefits.
For GetSavi, one of these benefits has been the assertion that AI transformation must be surrounded by good human processes. The machine will not do it all on its own. Sorry, it won’t.
“Keeping the human firmly in our AI-enabled design processes.”
Over recent months, we have emphasised the importance of the “pre- and post-AI” design stages in our processes: the categorisation of source material going into these applications, and the validation methods applied to the output.
As good learning professionals, our commitment to experimentation is making us stronger. It lets us keep our promise to clients – rapid, easily navigable, cost-effective digital learning, built from the rich source content they share.
We’re preserving the craft of instructional design as the technology behind it grows more powerful. That’s something we’re proud of and never want to lose.
Whether you’re an L&D leader, practitioner, or instructional designer, here are three lessons we are happy to share on mastering content design with AI – each one upholding the value of human intervention before, during, and after AI does its work:
- Lesson 1: Mandatory content and key references. AI can condense large volumes of source material into digital learning modules, but it still needs clear signals on what’s mandatory or key content. Skip this step, and you’ll face a rework loop later.
- Lesson 2: Upfront choices about features. AI will suggest content components for your courses, but it’s still worth defining the features your learners need upfront. That keeps you in control of course milestones and flow, rather than reworking an arbitrary design later.
- Lesson 3: Avoiding hallucinations. AI has accelerated text, image, video and animation production advantages, but accurate prompting, curation, and spot-checking remain essential. Hallucination happens – proper planning avoids this occurring frequently in the flow of your design activities.
We’ve done much of this groundwork already, so our clients don’t have to. However, our projects still build in regular points of collaboration, because our approach delivers integrity as well as quality at scale – keeping the human firmly in our AI-enabled design processes.
