A couple of weeks ago, I got back to listening to Beyond the Prompt episodes with Jeremy Utley and Henrik Werdelin. While there are many highlights, the episode with Blair Vermette stood out. And it has to do with curating the content AI gives you versus controlling it and expecting AI to create something perfect on your behalf. We need to bring something to the party too!
Additionally, AI tends to gift you with something you didn’t quite ask for. What do you do now? Get frustrated, try again or give up some control? Or embrace it and see where it takes you?
ps. I discovered these podcasts are also on YouTube. Still, I find I prefer the audio-only version—perhaps it helps me keeps me focused on listening and not getting distracted by the composition of the visuals.
As GenAI starts taking ground at work, how does an organization make its vast amounts of data useful?
When I first started using GenAI, my first thought was we can skip the fancy visualizations and go straight to asking our questions directly with GenAI. However this would require data to be organized, consistently labeled, and cleaned up. Easier said than done. Plus GenAI would need an assistant to actually do proper calculations.
So let’s start with an org’s existing linguistic data, its knowledge base that defines a company.
Back to basics – text first
Cleaning up, organizing and ensuring that the information is clear and readable. Fancy graphics, flat pdfs, animated presentations are not going to help. With GenAI, text is where it’s at. Is it good for a screen reader? Can we manually highlight the text? Then let’s go!
Yes, boring stuff. Yet it’s the boring stuff that makes GenAI so useful. Manufacturing documents for example usually follow a template, with headers, and clearly laid out text. The qualities that make a document accessible are what make it an asset for GenAI. So start organizing your documents, clean out obsolete, incorrect, draft versions. Take heed from the document-driven functions of your company. Lay it out clearly. And on the surface, it seems tedious, yet when you get into it, it triggers your mind to start understanding the content more. Once you’re organized, it frees you to think more clearly on to other topics. It allows you and others to access information easily, including GenAI. Writing it down and organization as liberation.
For example, at a manufacturing plant, try standardizing equipment maintenance logs. Boring? Absolutely. But this allows the maintenance team to ask GenAI “Which machines had bearing failures in the past six months?” instead of digging through spreadsheets. AI excels at finding patterns in simple, well-organized text and data.
Building bridges
With your documents and data organized, the next step is creating access points. Think about how GenAI will interact with your data. Creating simple indices, metadata tags, and consistent naming conventions is crucial. Your HR handbook, financial reports, and product specs might all live in different systems, and GenAI needs a coherent way to find and interpret them.
The question determines the answer
And while GenAI keeps improving, humans are necessary to provide the oversight, knowledge and experience. The key is teaching the team to ask good questions, not just expecting the AI to deliver insights unprompted.
Your AI should have an access badge too
Not all your company data is or should be accessible to all employees, which means your AI should follow the same model.
Good for AI, good for humans
Not surprising, organizing for AI benefits humans too. Clear document structures, consistent naming, and accessible information repositories make life easier for everyone. New employees onboard faster when they can ask basic questions to GenAI.
Start!
Somewhere in your org, there’s already a team or individual who maintains well-structured documents. See how well GenAI can assist in these areas, and spread those learnings.
It is a most insightful, meaningful and comprehensive collection of thoughts on what AI means for serving human dignity and the common good, and the roles all of us play in designing the future.
And while you can use AI to digest this essay for you, I would encourage you to take time and read it with full human capabilities.
…human dignity and the common good must never be violated for the sake of efficiency, for “technological developments that do not lead to an improvement in the quality of life of all humanity, but on the contrary, aggravate inequalities and conflicts, can never count as true progress.” Instead, AI should be put “at the service of another type of progress, one which is healthier, more human, more social, more integral.
Therefore, the ends and the means used in a given application of AI, as well as the overall vision it incorporates, must all be evaluated to ensure they respect human dignity and promote the common good.
Saint John Paul II observed that “humanity now has instruments of unprecedented power: we can turn this world into a garden, or reduce it to a pile of rubble.”
As return to office makes the headlines, there is a new hybrid in the works.
Human + Agentic Teams – HAT
(courtesy of yours truly)
Having spent the day in NYC hearing about all things AI, the dawn of the new hybrid is here.
Today’s food for thought –
Oversharing is a problem everywhere, not just on social media, but also at work
Success of agentic teams will still ultimately rely on humans (for now)
Understanding human behavior is key to change
ps. On my way back to Grand Central via the 7, us riders were unceremoniously kicked off after only a couple of stops due to an unauthorized person on the tracks…which resulted in a nice brisk midtown walk leading to an encounter with Antonio Pio Saracino’s The Guardian: Superhero. Seems a fitting representation!
As Thanksgiving comes around (in the US at least), so do year-end reflections, whether it be for work or personal reasons. As my memory seems to get hazier these days (I learned brain fog can be a symptom…), again the importance of making notes to self regularly.
Capture the significance of events while in the moment.
As time goes on, people forget the impact, the details, that happened during that time.
The faintest pencil is better than the sharpest memory. – Chinese proverb
Remember why plans were made.
We often make plans with current thinking in mind. As time goes on before fruition of those plans, many changes can happen and we often question whether the plan is still wise and why we decided on it in the first place.
And that’s where the importance of dialogue comes in. Written words won’t replace dialogue. In the age of GenAI, where almost anything can be had in written format instantaneously and give a false sense of understanding, it is the human mind that still needs to better understand through dialogue with others.
SOCRATES: You know, Phaedrus, writing shares a strange feature with painting. The offsprings of painting stand there as if they are alive, but if anyone asks them anything, they remain most solemnly silent. The same is true of written words. You’d think they were speaking as if they had some understanding, but if you question anything that has been said because you want to learn more, it continues to signify just that very same thing forever. When it has once been written down, every discourse roams about everywhere, reaching indiscriminately those with understanding no less than those who have no business with it, and it doesn’t know to whom it should speak and to whom it should not. And when it is faulted and attacked unfairly, it always needs its father’s support; alone, it can neither defend itself nor come to its own support.
Most people fall into one end of the spectrum when hearing these 2 letters. One side goes to the HI (human intelligence) end, and the other goes to the yes, AI is salvAItion!
When I first heard of ChatGPT, I fell into the HI side despite not even trying ChatGPT. I had no idea what it really did nor was I that interested. Fast forward to early 2024 and our organization has now rolled out Copilot. OK, big deal? Now what? Little did I know in leading this change, I was also leading myself to the other side, although I’m still in the somewhat moderate camp as we figure out the grey areas. Techwise, timing wasn’t too behind as it seems that friends don’t let friends use GPT 3.5 (and we are now on GPT 4)
So what led to this 180 or rather 135? In bringing myself up to speed, it turns out that having more information makes it easier to go from one dot to another without having to take giant leaps of fAIth. Imagine these giant lilypads spread quite apart from each other making it hard to reach. However, once you start learning about the mechanisms of Gen AI, seeing specific examples of what it can and cannot do, it’s as if these additional smaller lilypads appear in between making it easier to get to that next giant lilypad.
Curious about Gen AI? Here are my small lilypads –
Start with How to research and write using Generative AI tools. Dave is engaging and really goes into specifics which helps people connect with what they can really do with the tool. I liken it to staging a house. Some people get the overall potential and concept of an empty house, while others need that staging to bring the house to life. Taking it further, others need staging that is very specific to their tastes to make that personal connection.
We can’t bury our heads in the sand and pretend that GenAI doesn’t exist. Better to be informed and realize the implications, and when in the position to do so, advocate for the common good.
A great visual article on the overview of how Gen AI works. You’ll see why this process consumes so much energy – that’s another discussion unto itself…
The Gen AI Pandora’s box has been opened, find out what’s in it…
PS. I see a lot of posts about using GenAI for business but not a lot geared towards educators. It’s probably just my feed, however if you are a teacher or know those who are, it’s even more imperative that educators understand what this means in the learning space and how to ensure that students are still activating their own mental capacities. So no time like the present to rip off that band aid and find out the next generation is doing with GenAI.
Image generated by Copilot. Prompt: Can you create an image of a pond with big lilypad milestones and few smaller lilypads in between. Have a couple of frogs leaping between lilypads. Make it watercolor like, sunny, blue skies, vibrant colors.
