Your AI learning designer is a 'Yes-man', don't listen to him
ChatGPT needs no encouragement - and that's a problem.
This year’s explosion of generative AI tools are amazing. ChatGPT is like Google without the stress of having to pick a source. DALL·E is your eccentric nephew who’s pretty good at art. Synthesia offers a faster way to produce videos with uncomfortable looking actors who haven’t learned their lines.
And they all respond, more or less instantly, to any command that you give them.
But that presents a difficult question for those of us who works as learning designers: If the cost of content production is effectively zero - what is our role?
The question is pressing for us because the Mind Tools Custom team are, essentially, a content team. We produce custom digital learning content and workshops for clients who, as of this writing, are still paying us to do so.
Except, maybe that’s not entirely true.
At a recent team day, we ran a workshop on the implications of AI tools for our Custom team. (The workshop was, by the way, more or less designed by ChatGPT.)
There, we answered three key questions:
💭 What has our experience of AI tools been?
To date, most of us have been using ChatGPT for idea generation. It’s become an always-on braintrust whenever we’re stuck on a project or need inspiration to get started.
And we’ve got better at it!
Rather than asking: ‘Give me an example of a difficult conversation’, we are now more likely to use a prompt like: ‘I am a manager working for a global drinks company. I have eight people reporting to me, with varying levels of experience and attitude to work. Give me five examples of difficult conversations I am likely to have in this role’.
Then we pick-and-choose the output that best meets our needs, and re-write them as required.
So, for now, it’s a starting point.
💡 What opportunities exist?
The most enthusiastic response to this question was that it creates more time to spend with clients. The closer we are to our clients and end users, the better our solutions are likely to be.
We also think AI is likely to dramatically reduce the cost of translation (albeit it still needs a proofread by a native speaker for now), as well as handling administrative tasks like summarizing meeting notes.
😱 What concerns do we have?
There are three major concerns on our team.
First, the ubiquitous: ‘An AI could never do MY job as well as I can’. Which, in our case, translates to: Content will be generic, same-y, and inaccurate.
Personally, I’m not so sure. As tools like ChatGPT and Bard continue to improve, I’m pretty sure all of those hurdles will be overcome. Sorry, team!
The second concern is legalistic: What information are we providing to the AI, and who now has that information?
We’ve been careful to only input generic and anonymised prompts, rather than wholesale handing over client information.
And the third, finally, is that clients will use AI tools to generate content rather than come to us. It’s the big one. The existential concern.
🤔 So, what is our role?
Which takes me back to the start: What is the role of a learning designer in a world where the cost of content production is zero?
I think it’s to say: ‘No’.
Learning design is not, ultimately, a content production job. It’s about solving real problems for our organizations and our clients. It’s about being that friend or colleague who asks challenging questions in a constructive way, to co-create a solution that makes a measurable difference to performance.
Yes, an AI tool can create a 30-minute course on any topic that you want, in a matter of seconds. It’s the personal assistant who follows you round and says ‘Yes’ to every idea that you have.
But the greatest value we add as learning designers is to do the opposite of that. To say: ‘Hold on, are we sure that a course is the answer?’ And, if it is, to ask: ‘What is it that you actually need people to do off the back of this course?’
AI tools will doubtless have an impact on the way we work. In fact, they already are. But they’re also an opportunity to re-focus us on where we add the most value: Resisting the quick fix, and creating solutions that make a difference.
Want help challenging your business stakeholders to make a measurable difference to your organization? Contact custom@mindtools.com or reply to this newsletter from your inbox.
🎧 On the podcast
As the workplace becomes more automated, we thought we’d focus the Mind Tools L&D Podcast this week on making them more human.
Reena Anand is a speaker, writer and trainer who left her law career to focus on the intersection of neurodiveristy and race. For Reena, this calling is personal: her eldest son is autistic.
She joined Gemma and I to share her advice for organizations looking to support their neurodivergent employees:
‘I think there's a massive misconception that the adjustments or accomodations for neurodivergent people must be costly and time consuming. Actually, the vast majority are really cheap and easy to do. And those adjustments mean the difference between ‘an engaged employee with really great mental health’, versus ‘someone who feels they can't be themselves, burns out, and exists the workforce permanently’.’
Listen to the full episode here:
You can subscribe to the podcast on iTunes, Spotify or the podcast page of our website. Want to share your thoughts? Get in touch @RossDickieMT, @RossGarnerMT or #MindToolsPodcast
📖 Deep dive
Sometimes, the best workplace learning takes place far from from L&D team.
A new study from Stanford's Erik Brynjolfsson and MIT's Danielle Li and Lindsey R. Raymond, covered by NPR’s Planet Money, found that a generative AI-based conversation assistant improved the productivity of customer support agents by 14%.
How it did this is fascinating.
In essence, the customer support reps use text chat to solve customer issues. The AI model was fed past transcripts of these text chats in order to learn which responses led to the fastest resolutions, and then used that data to make suggestions.
Says Brynjolfsson:
‘What this system did was it took people with just two months of experience and had them performing at the level of people with six months of experience.’
That has two clear consequences: First, it increases the value of new employees. Second, it decreases the relative value of more experienced reps - the group whose past success was used to train the AI in the first place.
It’s a great early insight into the impact AI can have on the workplace, and thanks to Stella Lee for sharing it.
Brynjolfsson, E., Li, D., & Raymond, L. R. (2023). Generative AI at work (No. w31161). National Bureau of Economic Research.
👹 Missing links
🔍 Are we heading toward a content-less internet?
I’ve expressed multiple times my concern that ChatGPT is going to dramatically increase the amount of junk on the internet. It turns out the opposite might be true. According to Justin Pot, writing in The Atlantic, Google’s new search experiment is designed to deliver answers within the tool - rather than sending users off to the source. That robs creators of the page views that earn them advertising revenue, so why would they bother creating content in the first place?
🧠 Language is natural, critical thinking is not
In terms of poverty, illiteracy and disease, the world is getting better. In terms of critical thinking, it might be getting worse. That’s the view of Steven Pinker, author of Rationality. In the past, Pinker has argued that language comes naturally. In his new book, he argues that thinking requires effortful practice and outlines how a better understanding of logic, probability and randomness can help us make better decisions.
✍🏽 AI is great for summarizing notes, but here’s how to turn that into learning
Taking notes is useful for learning, and getting ChatGPT or Bard to summarize your notes gives you an easy shorthand for reviewing what you learned. But for long-term meaningful learning, we need to be more active participants in this process. In this newsletter from the excellent Dr Philippa Hardman, you’ll find three techniques (with prompts) to use ChatGPT to help you learn from your notes.
👋 And finally…
I’ve been listening to the fantastic Oscar Wars by Michael Schulman, which traces the evolution of the awards show over its near 100-year history and asks, at the outset:
‘What are the Academy Awards, anyway? The answers vary. They're a vaunted tradition celebrating a great modern art form. They're an industry party—like a convention of landscapers, but with better outfits. They're the closest thing America has to royalty. They're the only thing forcing Hollywood to factor art into commerce. They're a marketing ploy propping up a multibillion-dollar business. They're a method, however dubious, of organizing movies into a canon. They're a game. They're a relic. They're a fashion show. They're a horse race. They're an orgy of self-congratulation by rich and famous people who think too highly of themselves.’
The book explores all of these ideas, as well as some of the technological changes that have shaped the industry. Like, for example, the introduction of sound.
Far from being an immediately positive addition, the addition of sound to motion pictures met with resistance from directors who found that they could no longer use dramatic camera movements because the noise of the rigging required rendered them useless.
For an example of the creativity on display prior to the advent of ‘talkies’, check out the opening scene from 1928’s The Crowd:
👍 Thanks!
Thanks for reading The L&D Dispatch from Mind Tools! If you’d like to speak to us, work with us, or make a suggestion, you can get in touch @RossDickieMT, @RossGarnerMT or email custom@mindtools.com.
Or hit reply to this email!
Hey here’s a thing! If you’ve reached all the way to the end of this newsletter, then you must really love it!
Why not share that love by hitting the button below, or just forward it to a friend?