The biggest barrier to AI adoption is you
When AI advances fast, but humans need a bit of time.
Last week, my colleagues Claire Gibson and Dr Anna Barnett wrote about the role of human judgment in learning design. Per their guest post in The L&D Dispatch:
“We know how to cut through conversations to identify needs, we have a track record of designing learning interventions that change behavior, and, when we do use AI, we’re comfortable critically evaluating its outputs.”
Their newsletter struck a note of caution amid an explosion of AI-powered products at the Learning Technologies conference. While these tools might seem superficially impressive, are they tackling the right problems? Do they improve workplace performance and build capability?
This week, I want to take this from the other side of the equation: even supposing these tools are amazing, are we ready for them?
In our podcast earlier this year with Don Taylor, we discussed how it’s best not to think of AI in the aggregate as a ‘tool’. It’s more like infrastructure.
Like electricity, trains or networked computers, AI is a foundational technology: opening up vast possibilities across every domain.
But it also faces some of the same challenges as new infrastructure.
The first commercial power plant opened in New York City in 1882, but it took decades for electricity to be available in homes. Around the same time, there were concerns that travelling at speed on a train could suffocate you. More recently, the weak growth experienced during the computer age is famously described as the “productivity paradox”.
In each case, people’s reaction to new technologies has varied. The innovators and early adopters can’t wait to embrace it, the majority need some convincing, and the laggards dig their heels in until it’s impossible to ignore (see the adoption curve).
Where AI is different is in the speed of adoption.
ChatGPT reached 100 million users within two months of launch, the fastest growth ever seen for a consumer app. It now has over 400 million weekly users, more than the population of the US. And according to one survey, 95% of US companies are using generative AI in some way.
It’s not yet as ubiquitous as search. By comparison, Google processes trillions of queries a year. But it’s catching up.
Then there’s the anecdotal side. On a recent car journey, I was speaking to a “normal person”. Someone who works in an office, but not in L&D. They use a laptop, but they’re not following AI news closely.
They told me that they had just been given access to Microsoft Co-Pilot at work, and that it was “quite good for re-writing emails”.
Whether you’re a learning designer trying to justify your ongoing employment, or a learning leader looking to invest in new tools, it’s probably worth bearing this normal person in mind.
We are not typically designing for the early adopters, we’re designing for the majority
In her book, Design for How People Learn, our pal Julie Dirksen (who sits on our Product Advisory Board) outlines six gaps that prevent people performing at work.
We’ve written about these before, and I think they’re a useful lens for explaining why THE MOST TRANSFORMATIVE TECHNOLOGY MANKIND HAS EVER CREATED is currently “quite good” at writing emails.
Think about why someone might be reluctant to adopt AI at work:
🧠 The knowledge gap: “I don’t know what AI can do for me.”
💃 The skill gap: “I can’t get AI to do what I want.”
😟 The motivation gap: “Why the HAL would I use this thing that I think is going to kill me?” (Joke there for the nerds)
⬆️ The habit gap: “I’ve never had to use AI before and got on just fine.”
🏢 The environment gap: “I don’t know where to find AI tools and my Compliance and IT colleagues teams have freaked me out that I’m going to get us sued by using it wrong.”
📣 The communication gap: “No one has told me that I can use AI (or, they haven’t told me often enough and with enough guidance). So I either don’t use it, or don’t admit to using it.”
A case in point: For a couple of years now, Mindtools has offered a difficult conversations simulator, "AI Conversations”. The product won two Silver trophies at the Brandon Hall Awards and we’ve demonstrated that having just one simulated conversation increases user confidence by over 9%.
Each scenario gives users a chance to practice doing something they hate: having a difficult conversation.
Read that again: They can practice doing something they hate.
Why would anyone voluntarily use this product?
Where we’ve seen it work best is as part of a structured program, in an environment where the psychological safety of participants is prioritized and users go through the experience with their peers.
In other words, in an environment that is distinctively human.
Our old colleague Owen Ferguson used to cite Amara’s Law so often that I would roll my eyes, but he has a point:
‘We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.’
I’m sure that AI will transform how we interact with the world, in ways that we can’t yet imagine.
But, in the meantime, we need to remember that the biggest factor in the adoption of new technology isn’t its capability. It’s how comfortable “normal people” are with embracing it.
Here at Mindtools, our Custom team are experts in designing human-centric programs that use a range of tools to drive meaningful change. Reply to this newsletter from your inbox or email custom@mindtools.com if you want to improve the performance of your people and teams.
🎧 On the podcast
Whether it comes to facilitating workshops, designing learning experiences, or developing campaign assets, L&D practitioners are increasingly expected to bring creativity to their roles. But how do you do this if you don't have a creative background, or if you don't think of yourself as a 'creative person'?
In last week's episode of The Mindtools L&D Podcast, Ross D and Gemma were joined by Kenny Temowo, Senior Director of Talent Acceleration at SharkNinja and host of the Inside the Art of Making podcast, to discuss:
Kenny's career journey through the creative industries, into workplace learning, and back again;
what it means to be 'creative', and how this applies in L&D;
how to meet consumer-level creative expectations with L&D budgets;
how to earn the permission to be creative at work.
Check out the episode below. 👇
You can subscribe to the podcast on iTunes, Spotify or the podcast page of our website.
📖 Deep dive — Manager Skills Framework (6 of 12)
If you’ve been following this newsletter over the past couple of months, you’ll have seen us dedicating this deep dive session to the 12 key skills that make up the Mindtools Manager Skills Framework.
What we hope that you’re beginning to see is how these 12 skills interact, and this week we’ve got another cracking example.
In a 2015 paper from researchers at the University of Sheffield, the level of trust that exists between an employee and their manager was found to positively relate to business performance in three key areas:
Financial performance
Employee productivity,
and Product/service quality.
This is backed up by a paper from Crossley, Cooper & Wernsing (2013) which we shared when discussing goal setting. In essence, the more trust that a manager can instill in their teams, the better results will be.
How, then, to build trust?
One interesting finding from the Sheffield team was that trust was positively associated with the amount of training employees receive. When we invest in our people, they trust us more.
In our Skills Framework, we refer to the manager’s role in supporting workplace learning as ‘guidance’.
If we want to improve financial performance, productivity and quality, we need managers who actively contribute to their people’s ongoing development, collaborate on meaningful goals, and in doing so build a high trust environment.
Check out our Manager Skills Assessment to find out how we can help you measure these skills in your organization, and our Manager Skill Builder for how we can help build them!
Brown, S., Gray, D., McHardy, J., & Taylor, K. (2015). Employee trust and workplace performance. Journal of economic behavior & organization, 116, 361-378.
👹 Missing links
🔮 The future skills you’ll need as a learning designer
You might be sick of us linking off to Dr Philippa Hardman's Substack, but what can we say? She’s very, very, good. In this post, she explores Duolingo’s AI-powered approach to learning design and highlights the skills that learning designers need to develop if they want to remain useful: including prompt engineering, building feedback loops and interpreting analytics.
🤖 Wait, students are using ChatGPT to cheat on essays?
I’m shocked, SHOCKED I say, to discover that college and university students are using artificial intelligence to write their essays for them. Including one who began her essay on critical pedagogy with the line: “To what extent is schooling hindering students’ cognitive ability to think critically?” There’s always been a clear divide between ‘learning’ and ‘passing exams’, but the chasm between those two now seems impossibly vast. I’ve linked to an interesting discussion on LinkedIn about this topic, but I also liked Nick Thompson’s take:
‘If kids can get through your classes and curriculum by using AI then you need to change the classes and curriculum!’
Finally, this post—on the benefits of an em dash—struck a chord. Apparently the em dash has become an indicator that writing is AI-generated, but it’s also the punctuational equivalent of a panic button. When we started The Mindtools L&D Podcast nine years ago, we had a problem. We wanted to include episode numbers in our titles, but we often used colons too. How to present ‘436: Content libraries: Less is more’ without triggering an urgent colonscopy? Em dash to the rescue: 436 — Content libraries: Less is more.
👋 And finally…
This has absolutely nothing to do with L&D, I just found it hilariously funny. If you work in an office, I encourage you to take this approach to your next meeting!
👍 Thanks!
Thanks for reading The L&D Dispatch from Mindtools! If you’d like to speak to us, work with us, or make a suggestion, you can email custom@mindtools.com.
Or just 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?