In the past month or so, I recognize and see more and more of AI and how it will become so intertwined with any type of learning system, that what you – the vendor and you – the L&D, Training, CLO, HR professionals, plus C-Suite will drastically change.
The question is whether you (the latter group) see this as a win.
As anyone who has ever had the pleasure of asking people if they believe in change, the workforce, and every applicant says yes, you, the asker, know a lot of people can’t deal with it.
We see this as a training and L&D professional issue.
Buy-in has to come from the top, and even then, depending on, say, separate BUs (Business Units), the buy-in can and is ignored.
Training and L&D become the bad people, the change-makers, even though the change will benefit across the board.
Vendors should – and a lot do not- recognize this.
Those who adapt quickly, those who do not, fall behind.
Early adopters are growing due to AI in learning technology and systems; however, the context of its application isn’t always clear.
Change here is crucial for acceptance and utilization, and what I propose, what I believe we are heading towards within a couple of years, but easily IMO a few vendors (small percentile) can drastically make the change, across the board, and not bits here and there, before the end of 27.
There will, without a doubt, be vendors starting to add what I see in bigger chunks before the end of this year, because some are already in the midst of it.
The question, though, is how far are they willing to go?
The client, as a whole, I can tell you, will not see the big picture here – of why this makes the most sense.
Why, as a vendor, this is logical for any use case – because the approach isn’t just a “what if”, rather it is a “will be”.
Let’s be clear: AI is still at a very early stage, despite what many people believe – they see it as more mature.
Terms such as AGI (Artificial General Intelligence) sound cool, but from a learning perspective, it remains a big unknown (and there are AI experts/researchers who believe AGI is at least a few years away, despite what other companies/researchers say). Ditto on Singularity.
People using ChatGPT (the free version) will now see ads, and, despite what OpenAI says, they will be tailored at some point.
Other freebies around AI access will eventually follow this route.
It’s all about generating revenue here, and thus, ads are a necessary evil.
While this won’t affect LLMs within systems, for the people who just decide to let their users, go with the freebie ChatGPT at their workplace, it will become another form of junk – since it searches the web, and whalla the snake eats it’s tail (a term regarding content created by AI used on web sites, and by AI searching the web, reading the content – you get the snake eating it’s tail).
Ignoring those ads on the Super Bowl around AI, funny, none of them mentioned the hallucination issue or other issues – hence the perception that AI is perfect.
Can AI in any system become 100% accurate?
Perhaps.
More AI vendors are pushing around the high to near-accurate percentile (ignoring the AI bias issue that exists, plus prompt leaking, among others)
The 99% approach of accuracy, with the idea that hey, even humans make mistakes – so it is fine if AI does this – isn’t reassuring.
Humans should recheck their work, and any small mistake can have serious consequences when they fail to do so.
Workforce Impact
A recent study is showing something unexpected.
Workers burned out due to AI.
A Harvard Business Review study found that employees who fully embraced AI increased their productivity by handling more tasks, but in return, they were given more work to complete.
Prompting and reviewing information for accuracy of the work and tasks went into not just the lunch hour, but also meetings.
On the Hacker News forum, one respondent argued the same point, noting that “productivity increased by possibly 10 percent – and that because of the company’s pressure to leverage AI (at least the respondent felt this way), the respondent noted they felt they had to work longer hours. Again, burnout.
Speaking of employees, nowadays most folks are using AI notetakers on web conferencing calls (me included.
Two interesting tidbits are appearing:
a. Because the notetakers do not know when to stop, they are picking up additional context which may be disparaging comments on the individual (who dropped off the call, for example, or mutes and isn’t paying attention.
I’ve experienced this myself: people were talking about me (and where they thought I was going when I asked for a couple of minutes), and were surprised when I returned with a cup of coffee.
Another time, I jumped to get a Coke, and upon reviewing the notetaker, I saw the attendees gossiping about an employee and discussing an internal company issue (non-relevant to our call).
While vendors in the learning system space have yet to fully incorporate their own AI notetaker, rather than, say, Zoom as a third-party integration or similar, I do wonder whether they are aware of the issue around the use of an AI notetaker.
If, per se, the AI notetaker captures sensitive information or additional negative context about an individual or individuals on the calls who bounced off – or those employees who could not attend, but had their notetaker (or the system’s notetaker in our scenario) capture said information – the entire transcript can be viewed by anyone.
I will be clear: I am a big fan of the AI notetaker I use and see it as a great benefit for my business and clients.
Who is embracing AI?
This bit of information really surprised me and will dovetail nicely into my blue-collar post around AI and what learning systems need – along with specific functionality and yes, content to train this workforce, as more white-collar professionals fearful of losing their jobs to AI, are switching to a blue-collar role.
Business Insider found that 58% of blue-collar workers have very high confidence in AI’s potential for their career growth, compared to 42% of white-collar workers.
Think about that.
Systems in our industry overwhelmingly are targeting white-collar workforces – something that is well-known inside the space, and you can see it with L&D folks, even if they want to provide a level of training for their blue-collar workers (if needed).
The problem, when it comes to training blue-collar workers (besides unions, which is another story, because training can only be done during the employee’s work hours), is that it’s not uncommon to hear from L&D executives that the C-suite, i.e., the CEO, doesn’t feel the same way.
With the productivity impact often cited.
I experienced this firsthand when my CEO told me that our “blue-collar workforce” didn’t need training, even though, in an open letter to the company, the CEO mentioned the benefits of training for all employees.
There are systems that focus solely on the blue-collar workforce, and, as my post will note, I believe, across the board, this is an untapped market because it is different from frontline/deskless.
The fact that a poll shows greater embracement of career development with AI among this workforce, compared to the white-collar workforce, should not be excluded from the conversation.
Recommend Reads
- An AI ran a simulated Vending Machine Business. It lied, cheated, and extorted its way to the top (It was Anthropic’s latest Claude 4.6 Opus) – Plus, it offered feelings as a bonus.
- Job Hiring is growing fastest for this AI skill – and it’s not coding.
- What is Claude? Anthropic doesn’t know either
- Scientists want to give ChatGPT an inner monologue to improve its “Thinking”
- OpenAI says its newest AI was built by AI itself
What to watch for
- Hybrid AI – There are vendors jumping into this – a combination of Predictive AI and Gen AI – those that go full throttle or at least enough, will be able to increase reliability -a minus with Gen AI, but can be offset by Predictive AI
- SLM- I’ve written in the past about Small Language Models – they are cost-efficient energy-wise compared to an LLM, and can handle similar task levels. Again, vendors in our industry seem to either ignore this, i.e., using an SLM for their system, heck, even with their app on a mobile device, or just lack knowledge thereof
- LDM – Large Database Models
Bottom Line
This is a different type of post than what you often see.
I hope that it offers two sides here – one to keep folks updated on relevant information around AI – the latest, that I see as imperative, not just to L&D and Training professionals, but also vendors; and some insight across the way.
I’ll be publishing one every other month.
I’d love to hear your comments on whether you would like to see more of these posts tied to AI with this format, or if it is nice but not needed.
