what even is AI?

At Gyroscope, we use AI a lot, it’s kind of our thing. We developed Guided Discovery by curating Design Thinking and Design Sprints and adding AI tools and methods. In doing so, we drive efficiency that results in an actionable outcome.

You’ve most likely heard things like AI, LLM, neural network, algorithm and a bunch more terms around AI. So much that they may start to lose meaning or feel less-than-defined. Enterprise software terminology and TLAs can be somewhat of a barrier to overcome.

We wanted to create a quick and easy summary of defining AI and helping you understand its origins. Sifting signals from the noise is a challenge in any hype cycle.

The Baseline

AI isn't a wonder tech. What's going on with that hand, why are there 2 people staring in?

Let’s get something super simple out of the way, as there is an increasing misconception taking hold out there: AI is an umbrella term, think car to vehicle. Is a car a vehicle? Yes. Is a vehicle a car? Only when it’s a car. AI is the same thing: an LLM is a form of AI (its Machine Learning that mimics human intelligence), a text to image generator is also an AI. AI is an umbrella term to apply to many different types of technologies that allow computers to perform tasks in ways that reflect how humans interact and communicate. 2024 AI can: -See and Understand, which includes translating written and spoken language.

-Analyze data, utilizing multiple inputs, including sensors, monitoring tools, and system logs.

-Recommend, suggesting applicable content and decisions based on available data.

-Solve problems, applying reasoning, learning over time, and acting on behalf of users.

So, with that out of the way, how do we define it? Interestingly, that can be a bit of a moving target, as at the highest level, AI is described as any form of intelligence (read: decision-making) made by a machine that mimics a human. So do if and then functions qualify as AI? No. these are input only based logics, and don’t include any form of learning. AI fundamentally must be learning as part of the process, supervised or unsupervised by humans, by which it is making its decisions.

Artificial Intelligence (AI) is the ability of a computer or machine to mimic human thinking and problem-solving. It uses algorithms and data to make decisions, recognize patterns, and learn from experience.

Nothing new

If one classifies AI as automation that is programmed to drive an outcome then one of the earliest versions of the technology would be a Gyroscope. Early Gyroscopes (1817) were simple machines that provided data to humans to inform decisions. It is no coincidence that we named our company after the device in question. AI is, in fact, technically about 80 years old. The Logic Theorist program, written in 1956, is widely considered to be the first real AI, with research into the subject of artificial intelligence dating even further back into the 1930s. If you are interested in a deep cut, the origin story of AI can be backdated to the Chain Rule from the late 1600s. Like pretty much everything cool in the world, it started out as an idea first conjured in mythology and fiction before becoming part of our material world, but that’s a story for another time.

Artificial Intelligence has been around, humming quietly along for decades, in applications all around you, for some time, it’s only recently though that is has come to the forefront of our social consciousness in such a huge way. If you’ve ever played a video game, you’ve experienced a form of AI: the decisions the video game is making in response to your input as a player. Google can be attributed to early non-consumer gaming applications. Spelling correction, serving relevant ads, and suggesting search results are all forms of human mimicry.

Just 10 years ago, we were calling AI Machine Learning and integrating it into technologies to make products run better. For example, in the CAD space, topology optimization requires light-level machine learning to run at a reasonable compute speed to give results back in a timely manner. AI checked our grammar for this article, reflecting the flexibility of its application.

Ok cool, so what now?

Knowing is half the battle. Now that you’ve got some context to ground what qualifies as AI, and where it came from. You are now better positioned to understand how to use the technology, what is worth your time and effort, and what might just be co-opting the terminology for attention purposes.

We encourage you to remember that AI is a tool, not magical new wonder tech, and that harnessing it best requires research and consideration of what you’re trying to achieve. AI is not a strategy. Features are not a roadmap. You get the idea.

Want to learn more about how to use AI? Need support in applying it to achieve outcomes? Reach out to us here