Solutions Economic and Market Watch

Hello, and welcome to the Economic and Market Watch podcast. This is Sam Kem of CFC.

Today, you'll hear an economist's take on artificial intelligence and what it means for the workplace.

At work, there tend to be two types of people: Those who think AI can do everything and those who reject it before even giving it a try. I'll admit, I leaned closer to the latter — until one day someone used a single word that completely changed my perspective. That word was constraint.

So in today's episode, I'll share what AI — particularly generative AI, such as ChatGPT — can do what its limits are, and most importantly, how cooperative leaders should weigh the cost and benefits of AI. The first and most important step in effectively using AI is defining your relationship with it. People often say, "AI is a tool." And they're right — but somehow, that never fully clicked for me. Then, one day at a conference that I attended, an AI expert told the audience to think of AI as their five-year-old.

Other experts emphasized context, but this one emphasized constraints. And with AI, that's absolutely true. Instructions matter, but instructions with constraints matter even more.

When it comes to AI, your leadership shapes the output. Much like with a five year old, if you give vague or open ended instructions, you'll likely get chaos. With AI, that chaos is called hallucinations.

For instance, I asked AI to draft a podcast script on AI in the workplace. What I got was a very generic, buzzword-heavy piece — something you've probably heard dozens of times, and not exactly podcast-worthy.

So I tried again. I asked for real life examples, supporting data, and more detail. But here's what happened. The more I asked, the more the AI started making things up.

Why? Because much like your five-year-old, AI wants to please you. If it doesn't know, it will still give you an answer, just not always a true one. And the solutions to this? Constraints. Just like you might tell a child, "You can play for one hour, with these toys, and no hitting your brother."

So I uploaded bullet points with the structure and main points for the podcast. I added constraints on style, format and length. Finally, I told AI to tailor the script for a podcast on economics with an audience of electric cooperatives. And what you're hearing right now is the result, plus my own edits, of course.

So now let's turn to economics. As an economist, I weigh every tool in terms of cost and benefits. AI can make life easier, but does it make you more productive? That depends. This podcast, for example, would have been produced even without AI. Would it have taken more time? Probably not much more.

The hardest part of creating a podcast script is research and idea development, and I didn't rely on AI for that. Writing the script itself doesn't take too long. But having AI draft a version for me to revise certainly make things a lot easier. And here's the catch. Sometimes the convenience or time saved is what the electricity and the computing power consumed.

Other times, a human worker may be more efficient when you factor in the true economic costs, including energy use. For electric cooperatives, this is especially relevant. After all, AI system consumes significant amounts of power. That means using AI isn't just a technology choice. It also affects load forecast, resource planning and ultimately, members' bills.

Another common concern is jobs — especially entry level jobs. Many experts predict that AI will replace them, but I see it differently. Think about our interns and young staff members. They bring energy, creativity, and a perspective that AI simply cannot replicate. AI may summarize documents or streamline repetitive tasks, but if that's all your entry level employees are doing, then they're not being fully utilized.

Entry level staff should be engaging in meetings, learning from professionals, developing ideas, asking new questions and proposing solutions based on their growing experience. Those are things AI cannot — and will — not do.

If AI is replacing the core tasks of your entry level workers, that's not really an AI problem. That's a leadership problem. Leaders must ensure that workers are properly utilized in ways that AI cannot replicate.

From a macroeconomic perspective, humans, especially early career staff, often provide value at a lower cost to the economy than AI once you account for the resources that AI consumes. And that's an economist's take on AI.

As always, thank you for listening. Be sure to download the economic and market watch intelligence brief and dashboard. Talk to you soon!