AI Applications: What People Actually Use AI For

Most of the last few posts have been about how AI works. This one is about what it’s for.

The Applications group on the Periodic Table of AI is the nine green tiles on the right side of the chart. These are the everyday jobs people hand to AI, at work, at school and at home. For each one, you’ll get what it does well and where it can let you down, because both matter.

A glowing green sphere linked to a ring of small glowing icons: speech bubbles, a stack of pages, a bar chart, a magnifying glass over a book, a graduation cap, a headset, a code window and an envelope
Translation, summaries, data, research, tutoring, support, code and writing: AI’s everyday jobs.

If you’re just joining, the last post covered agents and automation, or AI that takes action on its own. This one is more down to earth.

Making things

Code Generation

Code generation means AI writing computer code from a plain-English request. “Write a script that renames all the photos in this folder by date.” It can also explain code, track down bugs and turn code from one programming language into another.

It’s one of the most popular uses of AI, partly because code is easy to test. Either it runs or it doesn’t. Still, AI-written code can look right and hide mistakes or security holes, so it needs the same review you’d give code from a new hire.

Writing Assistance

Writing assistance covers drafting, editing, rewording and brainstorming. Emails, reports, cover letters, product descriptions, the first draft of something you’ve been putting off.

It’s best as a starting point and a second pair of eyes. AI writing tends to sound generic unless you give it your details and your voice, and it can slip in facts that aren’t true. Read everything before you send it with your name on it.

Understanding things

Translation

Translation is one of the oldest jobs in AI, and one of the biggest success stories. In November 2016, Google switched Google Translate to a neural network system that translates whole sentences at a time instead of piece by piece, which made its translations far more natural.

Today’s AI can also adjust tone, explain idioms and translate a menu from a photo. For anything legal, medical or high-stakes, though, a professional human translator is still the safe choice.

Summarization

Summarization means boiling something long down to what matters: a 40-page report, an hour-long meeting transcript, a stack of customer reviews. It’s a real time-saver.

The risk is in what gets left out. A summary can drop the one detail you needed, or quietly change its meaning. For anything important, use the summary to decide what to read closely, not as a replacement for reading.

Data Analysis

Data analysis means asking questions of your numbers in plain English. “Which month had the most returns?” “Make a chart of sales by region.” Many AI tools can now write and run the calculations behind the scenes, so you don’t need to know formulas or code.

The catch is that a confident chart can rest on a wrong assumption, like a column read the wrong way. Ask the AI to show its work, and spot-check a few numbers yourself.

Research

Research means using AI to gather information, compare sources and pull findings together. With web search, AI can track down current information and point you to where it came from.

This is also where AI’s biggest weakness, hallucination, does the most damage. In 2023, a federal judge in New York fined two lawyers and their firm $5,000 in a case called Mata v. Avianca. The lawyers had filed a brief citing court cases that ChatGPT made up, complete with fake quotes. The judge noted there’s nothing wrong in itself with lawyers using AI, but they’re still responsible for making sure what they file is accurate. That lesson applies to everyone: AI can help you find sources, but you have to check them.

Helping people

Tutoring

Tutoring means using AI as a patient, always-available teacher. A good AI tutor doesn’t just hand over answers. It asks questions, gives hints and helps you work it out yourself.

Khan Academy built one called Khanmigo and released it in 2023, designed to guide students with questions instead of giving answers. The early results are a useful reality check. A large study of Khan Academy’s math program found that students using it improved, but having Khanmigo available didn’t add anything extra, and Sal Khan said students weren’t engaging with it much. An AI tutor only helps if students actually use it.

Customer Support

Customer support is one of the most common business uses of AI. AI can answer routine questions any time of day, look up orders and pass harder problems to a person.

But a company is on the hook for what its AI says. In February 2024, a tribunal in British Columbia, Canada, ruled against Air Canada after its website chatbot gave a grieving customer wrong information about a discount for travel after a death in the family. Air Canada argued the chatbot was responsible for its own words. The tribunal disagreed, saying the chatbot was just part of Air Canada’s website, and that the airline hadn’t taken reasonable care to make sure it was accurate.

Chatbots

Chatbots are programs you talk to in plain language, and they’re older than you might think. The first, called ELIZA, was built by Joseph Weizenbaum at MIT and described in 1966. It mimicked a therapist by turning your own words back into questions, using simple pattern matching.

ELIZA understood nothing, yet people quickly started treating it like a person. Weizenbaum’s own secretary reportedly asked him to leave the room so she could talk with it privately. That tendency to see understanding where there is none has a name, the ELIZA effect, and it’s worth remembering today. Modern chatbots are vastly more capable, but they can sound caring and sure of themselves whether or not they’re right.

How they fit together

All nine of these follow the same pattern. AI is fast, tireless and good at first drafts, rough sorting and finding a starting point. People are still needed to check facts, catch what’s missing, make the final call and take responsibility.

Used that way, as a helper rather than an authority, AI can save you real time in almost every one of these jobs.

Where to go next

The next group on the chart is Creative Tools, the AI that makes images, video, music and voices. That’s the next post.

For now, click any green tile and read the full page. If you only read one, try Summarization. It’s the easiest one to put to work today.

1 thought on “AI Applications: What People Actually Use AI For”

Comments are closed.