Artificial Intelligence 101

What is AI?

The term AI as we use it now almost always refers to generative AI. Generative AI is a group of algorithms that produce the most likely result based on a given prompt and the content it’s been trained on. A prompt is a request for the AI to do something. Most often, the prompt happens when you type a question into the box on the AI model’s website.

Large-scale AI models are trained on data scraped from all across the internet—blog posts, books, code bases, and more—but some AI models are trained on only a subset of data related to a specific task. For example, a company may train an AI on its help site and then use it as a chatbot to answer user questions.

You may also hear the term Large Language Model or LLM. An LLM is a type of generative AI that is specifically trained on human language. The most common types of AI that you interact with every day, including chatbots, are LLMs, but not all AI models are LLMs.

If you ask a generative AI model a question like, “How many financial institutions are in the U.S.?” it will reference all the material it was trained on to either find a source with the answer or use data to calculate it. Similarly, when you ask generative AI to make something new, such as an image, it pulls from the data it’s been trained on that is most similar to your request to produce a result.

The most common types of AI that you interact with every day, including chatbots, are LLMs, but not all AI models are LLMs

Face-scanning technology on your phone or at the airport, song and video recommendations on your favorite streaming platforms, map and GPS routing, and many other things you interact with every day also utilize non-generative AI. This is a type of AI focused on analyzing data and spotting patterns, rather than creating new content.

Usage

Some people think the advent of AI is an industrial revolution, because it’s a monumental breakthrough that’s changing so much about the way people live their lives. AI can be used in a lot of ways—many are yet to be discovered or fully realized. Common uses include…

• Data analysis
• Automation
• Research and summarizing data
• Content generation

AI makes it possible to do many tasks extremely quickly. Things that once took a lot of time and money can now be completed almost instantly and without the need for direct human intervention. This could lead to massive jumps in productivity, creation, scientific discovery, and more. It’s likely that AI will continue to integrate into everyday life, impacting everything from how houses are designed to how jobs are executed.

Fallibility

Because of the way AI generates answers or requests, it isn’t always correct. It doesn’t “know” anything; it makes a well-educated guess based on its training data. That guess is often mostly correct, but sometimes pulls from faulty or incorrect information, or even assumes an answer that ends up being false. Since it’s trained on human-made data, AI also carries human mistakes, biases, and assumptions.

You should be extremely careful about implicitly trusting an answer generated by AI. It’s very possible that this answer isn’t necessarily or entirely correct. Many people refer to incorrect information given to them by AI as the AI “hallucinating,” or an “AI hallucination.” Always double-check the information you get from AI with a secondary, trustworthy source.

Artificial intelligence can be a powerful tool, but it's important to understand what it actually is and how it works before you use it.