What is an AI assistant?
ChatGPT (made by OpenAI), Claude (made by Anthropic) and Gemini (made by Google) are AI chat assistants. You type or speak a request in plain language, and the assistant replies in plain language. Under the hood, each one runs on a large language model: software trained on enormous amounts of text so that it can predict what a helpful, well-written response to your message looks like.
That word predict is the key to understanding both the magic and the limits. An assistant does not look up a single “correct” answer in a database. It generates a new response each time, word by word, based on patterns it learned. That is why it can write a birthday poem about your dog in seconds, and also why it can sometimes state something false with complete confidence.
Most assistants today can do more than chat. Depending on the tool and your plan, they may be able to search the web, read files you upload (PDFs, spreadsheets, images), create images, talk with you by voice, and remember preferences between conversations. Features change often, so treat any list like this one as a snapshot rather than a promise.
What they are genuinely good at
- Drafting: first drafts of emails, posts, letters, outlines and reports.
- Explaining: breaking down a confusing topic at whatever level you need (“explain like I am new to this”).
- Summarizing: turning a long article or document into key points.
- Rewriting: changing tone, length or reading level (“make this friendlier and half as long”).
- Brainstorming: generating many options quickly: names, ideas, angles, questions to ask.
- Organizing: turning messy notes into a checklist, table or plan.
Where they fall short
- Facts can be wrong. Assistants sometimes invent details, sources or numbers. This is often called hallucination.
- Knowledge has gaps. Without web search, a model may not know about recent events.
- They do not truly know you. Unless you give context, answers will be generic.
- They can sound certain when they are not. A confident tone is not evidence.
- They are not professionals. For medical, legal, financial or safety decisions, use AI to prepare questions, then talk to a qualified person.
Example: a vague request vs. a useful one
A vague request:
Write an email to my landlord.
You will get a generic template. Compare it with this:
Write a short, polite email to my landlord, Maria. The kitchen tap has
been dripping for a week. I'd like someone to fix it this week, and I'm
home after 5 pm on weekdays. Keep it friendly and under 120 words.
A reply to the second request might look like this:
Hi Maria,
I hope you’re well. The kitchen tap has been dripping for about a week now. Could someone take a look this week? I’m home after 5 pm on weekdays, so any evening works for me. Thanks so much for your help!
Best, Jordan
The difference is not the tool. It is the context you provided: who, what, when, tone and length. You will practice this throughout the course.
A healthy mindset
Think of an AI assistant as a fast, well-read helper who has never met you and occasionally makes things up. You stay in charge: you give direction, you review the work, and you decide what to use.
Practice task
Open any AI assistant you can access (you will set all three up in the next lesson). Ask it for something useful in your real life using the pattern who + what + details + tone + length. Then ask the same thing in a vague, one-line way. Compare the two answers and write down two differences.