AI in plain language
You do not need to be technical to make good decisions about AI. You need an accurate picture of what it is and how it behaves. This module gives you that picture in plain language, so the rest of your learning rests on something true instead of on hype or fear.
By the end, you will be able to
- Tell the difference between automation, AI, and generative AI in your own words.
- Explain, simply, how a generative tool produces an answer.
- Name why an answer can sound confident and still be wrong.
- Point to where AI already shows up in everyday nonprofit work.
Lesson 1
Automation, AI, and generative AI
These three words get used as if they mean the same thing. They do not, and the difference changes how much you can trust the result.
Automation follows rules a person wrote. If a donation over a set amount always triggers a thank-you letter, that is automation. It does the same thing every time, and you can predict it.
AI is a broader word for systems that make a judgment or a prediction rather than following a fixed rule. A spam filter deciding what looks like junk mail is doing a small piece of this.
Generative AI is the newer kind most people mean in 2026. You give it words, and it produces new words, images, or code that did not exist before. A chat assistant that drafts a grant paragraph is generative AI.
Note
The practical point: automation is predictable, and most everyday AI tools are set up to give a slightly different answer to the same request each time. That is a setting, not a law of the technology, but with the tools you will use, treat the output as variable. It is why a person still has to check the work.
Try it
Think of one tool your organization already uses. Is it following fixed rules, making a prediction, or generating something new? Say which, and why.
RememberAutomation follows rules. Generative AI produces something new each time, so it needs a human check.
Lesson 2
How a generative tool produces an answer
You do not need the math. You need the one idea that explains almost everything these tools do well and badly.
A generative tool has been trained on an enormous amount of text, and in many tools images and other material too. From all of that, it learned patterns about which words tend to follow which other words. When you ask it something, it is predicting a likely next piece of text, one step at a time, based on those patterns.
That is the core engine. On its own it is not looking anything up in a file or checking a database of facts. It is producing what a plausible answer usually looks like.
Some tools go a step further and search the web or your own documents before answering, and they will show links or sources when they do. When a tool cites a real source it pulled in, that is more trustworthy than an answer from memory. Even then, open the source to confirm it says what the tool claims. When no source is shown, assume it is working from patterns, not a lookup.
Ask for a summary of a document you paste in, and it does well, because the pattern is right there in front of it. Ask for a specific grant statistic from memory, and it may invent one, because it is filling in what a plausible statistic would look like rather than retrieving a real one.
Note
Hold onto this: the tool is very good at shape and language, and unreliable about specific facts it was not given. That single sentence predicts most of what you will see.
Try it
In your own words, finish this sentence: a generative tool is good at ______ and unreliable at ______.
RememberIt predicts likely text from patterns. That makes it strong on language and shape, weak on specific facts.
Lesson 3
Why fluent is not the same as true
The most common mistake is trusting an answer because it reads well. Confidence is not evidence.
Because these tools are built to produce fluent text, a wrong answer looks exactly as polished as a right one. It rarely signals when it is unsure, so you cannot rely on its tone to tell you when to doubt it.
- It can state a false fact with full confidence. This is called a hallucination: the tool fills a gap with something that sounds plausible.
- By default it can give you a different answer to the same question two minutes apart.
- It can miss context it was never told, and answer as if it had the full picture.
- It can be out of date, because its training has a cutoff and the world moved on.
Important
A fabricated statistic or a made-up source in a grant application can disqualify the proposal and damage trust with the funder. Fluent and wrong is the dangerous combination, because it slips past a quick read.
Try it
Recall a time a confident-sounding answer, from anyone or anything, turned out wrong. What would have caught it? That habit is what this whole academy builds.
RememberPolish is not proof. Assume specific facts need checking, no matter how sure the answer sounds.
Lesson 4
Where AI already shows up in nonprofit work
This is not a future you are deciding whether to enter. It is already in your tools. Knowing where changes what you watch for.
AI is already woven into software your team uses. Recognizing it is the first step to using it on purpose instead of by accident.
- Visible tools you choose to open, like a chat assistant you use to draft an email.
- Embedded features inside familiar software, like suggested replies, writing help, or summaries in your inbox and documents.
- Vendor systems, like a donor platform or an application portal that scores or sorts people behind the scenes.
- Automated decisions, where a system flags, ranks, or routes a person, sometimes without anyone noticing a judgment was made.
Try it
List three places AI might already be operating in your organization's tools. You do not have to be sure. The point is to start noticing.
RememberAI is already in your visible tools, your embedded features, and your vendors. Notice it so you can govern it.
What you leave with
One-page AI concept map and glossary starter
A single page in your own words: the three terms, the one idea about how these tools work, and the two limits that always need a human. It becomes the shared language your team can build on.
This is just for you. It saves on this device only, and nothing is scored.