“It’s a terrible product.” “None of us were like, this is really useful.”
Those two statements are from the co-founders of OpenAI, who have created the hottest generative AI tool: ChatGPT.
The solution as it exists right now, is a serious work in progress. Numerous articles have been published showing its shortcomings, from “declaring love and harboring negative feelings about its developers,” “threaten a philosophy professor,” to “disinformation and the potential rise of conspiracy statements,” and “getting things wrong when writing.” Then there is this little issue around copyright, who owns what,” and whether or not “ChatGPT will fall under the AI risk from European regulators.“
It would be one thing if all these issues, were resolved and everything was all set, but that just isn’t the case. And ChatGPT is not alone. Midjourney, another generative AI tool (also free), where an individual created a comic book, using the tool (which changes text to images/art/graphics), and now has come afoul with the U.S. Copyright office. Before you say, well this problem around copyright is only a U.S. issue, based on my research it is a Germany issue, a Canada issue, and as noted earlier, a European issue. My guess is that it will be/if not already an issue in many other countries.
This post though isn’t about the concerns around generative A.I., although before adding it your learning system or learning tool, you should investigate and ascertain what if any consequences could come about by doing so. There are already a couple of vendors, in the works with it, and at least one, Firmwater, who says they have added it to their learning system (I have been promised a demo). Many of you are probably aware of it’s widespread use in Academia, by students (in this case it is used with ChatGPT. There isn’t any doubt the race to use generative A.I. is on. Venture capitalists are dumping huge amounts of capital into firms that offer it, and firms that have solutions with it in it. Microsoft, Google to name just two, are going full throttle with ChatGPT. I tested Bing with ChatGPT on a mobile device, and came away somewhat impressed, and equally unimpressed. Generative A.I. isn’t a new concept, from the standpoint of new in 2023, yet the raise of one specifically, ChatGPT has impacted the markets, never before seen…
Well, not exactly. When the Internet, with web sites, first appeared, there were folks who raced to using it, were enthralled by it, and spent way too much time on it. Then there were folks who thought it was a fad, and companies who sat on the sidelines, for way too long, seeing how this would work. Yahoo back in the days actually checked every website, that sought to be on their search engine. Yep, humans did the checking. Thus, if you made it onto Yahoo, you were going to generate a lot of traffic – my site in fact, did. Mosaic 2.1 was the rage, and CompuServe (RIP) was out. Netscape took the browser world by storm. The point here is that technology not fully understood, debuted, and changed the way we do things, experience things, and learn things, appeared. Nobody ever thought about the legal ramifications, disinformation, and negative message spreading that impacted many lives, nor the consequences of privacy. On the flip side, I did buy a few SNES video games and never had my credit card hacked. Those were the days.
Learning system vendors didn’t incorporate Mosaic, but with the Internet (again with websites, for it somewhat existed prior to 1994-95, sans the websites), systems in the cloud could appear (we didn’t call it in the “cloud”, many of us referred to it as on the vendor’s servers).
From a learning system standpoint, and an e-learning tool/product or even generative A.I. specific for learning/training in the corporate side, GAI (I will refer to it as this acronym), has the potential to be change learning and training as we know it. In five years’ time, a product without GAI, will be seem as outdated, and old school. Thus vendors, will need to add it, regardless of whether they the vendor has hired generative engineers/developers or went third party. Today, people know ChatGPT, but others do exist. They include Jasper (fee-based), MidJourney (free), Github Copilot (uses OpenAI codex) (fee-based, but really inexpensive), Rowy (free), other offerings from OpenAI (fee-based, including ChatGPT), Gupshup (Fee), and YextChat (beta). As I write, uh not using generative AI, I am testing out You.com, an AI search engine. Oh, and before I forget, you may have missed the article about Duolingo who will “create AI chatbots to teach languages.” If I was a vendor, like Workato, that connects APIs, and right now has a growing customer base with learning system vendors, I’d be paying attention to GAI offerings that just by developing code, could do the same thing (at a cheaper price point, let alone free).
It’s interesting that a language provider would be diving headfirst in working with OpenAI, when learning system vendors’ overwhelmingly, have ignored adding courses/content in teaching and thus learning foreign languages. Personally, I always found this nonsensical for doing so. Perhaps, they will start to pay more attention, and who knows. Speaking of which, I do know of a 3rd party content provider who is adding generative AI to their system (where the content is housed). I will be seeing it, within the next two months.
It’s the term to use when referring to Mixed Reality (A combination of AR and VR). I saw MR many years ago, at a show in London (for VR, AR and future tech, such as MR). I was immediately impressed, but well aware that it was very early. A hologram in one location, being seen by potentially thousands at a conference across the globe in real time. You could have a presenter at one location, presenting, asking and answering questions, and being seen by others in real-time (with low to no latency), anywhere else in the world. Just think about that, from an ILT standpoint, or seminars. EdTech, I could immediately see a potential there, although EdTech’s track record on solutions such as this, is a mixed bag. They jumped headfirst into VR, ignoring the costs, and so on. Anyway, XR is combing the physical world with the virtual world, with a real-time experience. HaloLens by Microsoft, the product I used, was the first, but it was beta land, and the headset was heavy and uncomfortable. Yes, you have to use a headset, a minus that I saw and still see with VR, but with all those folks going rah-rah over the Metaverse (the technology to go WOW, isn’t there yet), but some companies are adding their offerings in the pseudo Metaverse, XR I believe offers the best way to do so.
If you are an early adopter and want to jump right into XR, you can..almost..as two offerings are in the works. One from Apple (price is supposedly going to be even more expensive that HTC), the other from HTC, which produce amazing headsets, but outrageously expensive. HTC by the way, shows some amazing examples of XR on their web site. Apple’s XR reportedly will allow you to do virtual typing.
One reason I love XR, besides the hologram human/avatar – the one I tried it was not 100% me, the technology wasn’t there at the time – is what could be achieved from a learning and training standpoint. I could go into all the possibilities, but I do not believe XR will be the behemoth for learning technology.
It will be Digital Twins.
You are a learning system vendor. You want to test to see how this API or integration (deep or otherwise) will do and act in your system. You are in the “wonderzone.” With a digital twin, that concern is forever gone.
You just bought your learning system. You are excited but are unsure what an integration with your HRM or HRIS will show, and what data will appear. The vendor says, “don’t worry,” yet you are worried. A digital twin of your learning system will resolve that, by allowing you to test the integrations in real-time, virtually, with simulations, and constant data streams. Your worries are gone.
You are onboarding, you want to provide your new employees with various experiences, exactly like it will exist at your company. You are not interested in 3rd party courses, or building ones out, because it cannot produce all those experiences, for each individual. A digital twin, which they can view on a computer screen, can offer all those experiences and real-time sims, constantly updated with data, to predict future behaviors (and in our case, learning behaviors, training experiences).
What are digital twins?
Digital Twin Consortium
- A digital twin is a virtual representation of real-world entities and processes, synchronized at a specified frequency and fidelity
- Digital twins use real-time and historical data to represent the past and present and simulate predicted futures
- A digital twin is a digital representation of a real-world entity or system. The implementation of a digital twin is an encapsulated software object or model that mirrors a unique physical object, process, organization, person or other abstraction. Data from multiple digital twins can be aggregated for a composite view across a number of real-world entities, such as a power plant or a city, and their related processes.
- Digital representation of an intended or actual real-world physical product, system, or process (a physical twin) that serves as the effectively indistinguishable digital counterpart of it for practical purpose.
From a business standpoint (not just our industry), an article by VentureBeat, does an excellent job showing digital twins’ capabilities.
To fully see the power of digital twins:
- Digital twin opens way to effective treatment of inflammatory diseases
- 15 Digital Twin Applications/ Use Cases by Industry in 2023
- 7 Amazing Examples of Digital Twin Technology In Practice
While again, these are business examples, as with many technologies developed for business, they equally can be used for learning and training.
Anyone who has ever managed people in an office, has experienced the “we just knock and come in” employee to ask questions or chat. In a given day, you may have this thrill multiple times. And you are still expected to get work done.
Take a digital twin of you in your office, with multiple people (maybe twins of them), and using scenarios, see the outcomes if you do this or that with each situation. Now, you can see the application and thus action of your decision-making behaviors before testing it out in real-time. Sure, you can say, a course can do this – and yes, it can but it a point. It can’t be updated in real-time; it can’t forecast outcomes based on various scenarios, and uh, it isn’t you – just an image of whatever.
The possibilities with learning system vendors tapping into digital twins, will be outstanding if they do it.
Digital Twins with AI is already in play. And what makes it really cool, is that you – the vendor or even yourself (uh the buyer) could start right now. Amazon Iot Twinmaker. Microsoft Azure Digital Twins.
Although when you see it saying iOT, and then think, well, “we can’t use it for training, or learning,” in fact you can.
- “A digital twin enhances training opportunities for employees whose job sees them working on more hazardous components of the physical asset. While some industries already use simulations for worker training, a digital twin is far superior by virtue of being a completely replicated virtual reality environment. This greatly enhanced experience – and the ability to see elements that would have otherwise been hidden (or highly dangerous to access) increases a worker’s knowledge base in a safe and controlled environment. Employees are also better prepared and monitored when they do have to navigate dangerous mechanical components or structures.” (Visionaize)
- A digital twin model of leadership or an entire workforce (Kevin Kruse, The Metaverse, Digital Twins, And Leadership Development):
- You could receive hyper-personalized nudges based on your unique digital twin, receive tips and training on how best to communicate with your manager, your peers, your direct reports (based on everyone’s digital twin), advice on how to individualize your leadership based on the digital twins on your team and receive career path recommendations that better suit your strengths, personality, and experience
The two examples above, show off just a slight bit of what a digital twin can do, and with generative AI, we are talking to the moon and beyond.
Can we do it today?
Yes, anyone whether you are in L&D, or Training or HR or EdTech or Product/Marketing where you are providing training/learning to your employees, customers, association members, internal stakeholders, could start. It is not easy at the moment, and technical skill sets are a must, but it is doable. As a vendor in the e-learning space, especially those focusing on L&D, with their learning system – you should be starting to think about how to add it and offer it to buyers. For customer education, OMG, the potential for your clients using it, as a benefit to your system – thus the twin is included, too me, is incredible. Especially those who are in retail, hospitality, products/services industries.
And the best part to all of it – no need for buying a headset. No need to buy a special type of computer, that nobody can afford.
Technologies that were created for the populus, or business outcomes, or specific entities, have in the past, been tapped in the learning and training markets, including e-learning. Second Life (ATD, then known as ASTD, have a location in it). Web Conferencing, WebEx, was widely used in the tech space in the early 2000’s. Hence the birth of webinars for corporate. Centra, was available in the EdTech space in the late 90’s. Designed for business, used by folks offering training and learning to their employees and customers.
VR designed for the populus, initially gamers were the buyers, now used for learning and training. CompuServe for the populus – I used it to create a student network for journalism, whereas students in Japan, published stories following the earthquake, and my students, posted the articles in the school newspaper. Other students from other countries, and schools could tap into the network too.
Chatbots, created to irritate people when they come to a web site, now used a variety of learning system vendors. Knowledge bases. Forums. Discussion Boards – all available in 1994 – I used them on my web site, now? You see them, and have seen them in learning systems. They weren’t created for a learning system, but they are there. And the list goes on.
Technology can provide a lot of wonderful things, if people in the learning and training sectors think out of the box, on how to tap into the technology for their students, employees, customers, association members and so on.
As long as you recognize that technology evolves, and everything isn’t perfect right now, but will continue to improve (we hope), then jumping early in, isn’t a negative.
It’s a positive.
Speaking of which, a company called Stable Diffusion using their AI, is able to create a text to image model. They did so on a smartphone, without the person having to connect to the internet.
Out of the Box.