A Recipe for AI: The AI Taste Test!
Lauresa E.
Vice President, Executive, AI Transformation, AI & Data leader, Innovator and Agilist
We often talk about AI as if it is something we magically cooked up that delivers mind-blowing results. AI is like the perfect combination of savory and sweet, hitting everyone's palate with the unique precision of a Michelin chef. Suddenly, we all become legendary chefs.
However, there's an irony here: an AI-generated meal or secret recipe may fall short when actually executed and eaten by a human. How do we know if the person executing the recipe has the necessary skills or if the ingredients, under this skillset, can deliver the mind-blowing culinary results?
Coming from a world of technology, we often forget the reality of applying AI to the practical human experience—like eating, which most humans do several times a day. The idea that AI could delight at lunch today or dinner with a client seems far-fetched. "Amazing, AI Chef," we might joke, but it serves as a lesson in how our decisions in AI affect the intended outcomes for human beings.
AI requires data, so while it might be the perfect recipe according to several sources, the available ingredients or even their quality might not match up. A favorite memory is trying to recreate Thai food from a cooking class at the Blue Elephant and realizing it was inedible with the quality of ingredients available near me! I had taken a class, had the recipe, and the only factor was a difference in the sourcing of the ingredients, resulting in a disaster for the taste buds. More embarrassing yet, I was showing off my newly acquired skillset to a group of friends, and, to my dismay, it ended in unexpected and utter failure.
It also takes a skilled human to execute the outcome, no matter how advanced the AI recipe is. Imagine our AI Chef making a complicated sauce; one wrong temperature setting or an ingredient added out of order can greatly affect the dish's outcome. The true result won't be known until it graces your taste buds. The skillset of a person executing AI actions can lead them into believing they are successful, even when the final result may fall short and in the ultimate business sense gives suboptimal results.
Recommended by LinkedIn
Personalization further complicates the recipe. AI might know you don't like onions and suggest leaving them out, but the dish might not taste good without them. If you hadn't known they were in there, you might have loved it, but your bias against onions influences the result. So because of the data, did AI create a bias in the outcomes presented? Did it change the outcome of the results based on bias?
We also rely on trusted advice from AI, especially in our workplaces, to guide us in executing tasks to the highest standards. AI should understand our environment and equipment, making adjustments or suggestions as needed. Executives in AI must apply the 'taste test' to their strategies, ensuring the right outcomes are achieved and aligning AI capabilities with practical, real-world applications.
AI requires great care and isn't a perfect recipe at face value. This metaphor of AI as a culinary artist highlights the complexity and nuance required for successful AI applications. Just as a recipe needs a skilled chef to come to life, AI solutions need the right context, expertise, and adaptability to deliver their promised impact.