The Bread Machine Problem: What AI Literacy Gets Wrong

I have never liked cold weather. Now that we have established that, there is one thing I will admit that I do enjoy. It is something that usually only happens on those days when there is just the right mix of snow and ice to shut everything down and force you to stay inside. It is the opportunity to slow down, go into the kitchen, and enjoy the process of making something from scratch.

One of the things I love to make most is homemade bread. There is nothing that makes a house smell more like a home, or feel more inviting to strangers and friends, than the smell of homemade bread, particularly on a day when the world is cold and frozen.

To me there is just something about gathering each individual ingredient yourself, making sure you have just the right amount for whatever type of bread you're making, and getting your hands in the dough. It just makes it taste better somehow.

I do not need a bread machine to make bread because the machine itself is not the skill. The skill is knowing how to make bread without the machine. What are the ingredients, and what does each one do? It is knowing how a little more of this and a little less of that can affect the end result.

The point is the machine serves a purpose that has no direct connection to the skill that has taken me years to learn. The machine is a support—nothing more, nothing less.

AI tools are the same way. They help support and clarify thinking. They can serve as a vehicle for faster execution or as a silent automation running in the background that frees you up to do other, more pressing tasks. The tool itself is not the skill.

This is where a lot of current AI training in operational literacy is misplaced. Much of the training available today places an emphasis on tool training. Learn tools "X," "Y," and "Z," and you will become an AI expert. There are two glaring problems with this philosophy that only recently organizations are beginning to grasp.

The first problem is the nature of tools themselves. Tools do not last. They require maintenance, upgrades, and eventually replacement. In the past this was not such a big deal because each version lasted long enough that newer versions were often years down the road. AI changes the game.

OpenAI became available to the general public in late 2022. We are not even four years into the introduction of AI to the public at large. When you stop and think about that, it is almost mind-blowing. Look how far and how fast we have come in just a few short years. AI is everywhere. You can hardly find a piece of software untouched by some form of AI functionality.

To take a tools-first approach to learning AI is to limit yourself to an education that is only valuable for a limited time. In just a few months, you may find yourself needing to start over. This is one reason literacy has never been tied to tools. Tools change. Literacy survives because thinking transfers even as tools evolve.

The second problem is that AI, by design, is an iterative tool. It is not traditional computing where you put a specified input into a system and receive a specified output. In fact, it is entirely possible for two people using the same AI tool to enter the exact same prompt and receive different results. The quality of the interaction is influenced not only by the tool itself, but also by the person using it.

A tools-first approach to AI literacy would almost have to accept the idea that AI tools will consistently produce the same outputs regardless of who is using them. This is simply not the case. Anyone who has ever been pulled in by a social media promise of a perfect prompt or a "secret" trick that gets AI to do this or that can testify to the fact that not all interactions are equal.

My bread machine can make bread. That is true. I know how to use my bread machine, but my sister-in-law has a newer, more advanced machine. I do not know how to use her machine, even though I know how to use my own.

To fully master AI, it has to start with building a skill. It has to start with the one thing you need regardless of which AI tool you use: your mind.

Your mind, and the way you think with it, is the one internal tool you possess that levels the playing field in an AI environment. A mind that has learned to think and operate with iterative tools, a mind that has learned the power of evaluation and the need for stewarding systems, is a mind that demonstrates mastery of a skill.

A bread machine can make bread, but it cannot teach someone how to bake. Likewise, AI can generate outputs, but it cannot teach someone how to think.

The future will not belong to those who master a particular AI tool. It will belong to those who have taught their minds the skill of thinking with AI before the next "new" thing arrives and long after it has been replaced.