AI in Your Daily Life

You already use it

Before anything you deliberately open: predictive text, spam filtering, route timing, photo search, recommendations, fraud alerts on your card. All of it is the same family of technology, working invisibly, and it has been for years.

The difference with chatbots is that you now direct it yourself. Which means it is now on you to know when it helps.

One test that sorts most tasks

Before handing something to AI, ask:

Is checking this answer cheaper than doing the task myself?

That is it. It sorts almost everything correctly.

• “Rewrite this paragraph to be shorter” — checking takes five seconds. Strong fit.
• “What is the notice period in my employment contract?” (contract pasted in) — checking means reading one clause. Strong fit.
• “What is the statutory notice period in my country?” — checking means finding a reliable legal source, which is the entire task. Poor fit.

The pattern: AI is at its best when the material is already in front of it and you can judge the result immediately.

Where it genuinely saves time

Transforming things you already have

Summarising a document you pasted. Rewriting your own draft. Extracting the dates from an email thread. Turning notes into prose, or prose into bullet points. Translating something you can sanity-check. Low risk, immediate payoff.

Getting past a blank page

A bad first draft you can react to beats an empty document. Most people edit far faster than they compose, and this exploits that directly.

Producing options to choose between

Ten gift ideas, five names, three ways to phrase an awkward sentence. You keep the judgement; the model supplies volume.

Explaining things at your level

“Explain this the way you would to someone who knows X but not Y” is one of the genuinely excellent uses. For well-documented subjects it is patient, adaptive and available at midnight. For obscure ones, be more careful.

Structured thinking

“Here is my situation — what am I not considering?” It is good at surfacing angles you missed, because it has seen the shape of the problem many times. Weigh the answers yourself.

Where it quietly costs time

• Anything needing current, precise facts. Prices, availability, opening hours, this year’s rules. It will produce something plausible and wrong, and you will spend longer unpicking it than looking it up.
• Arithmetic across many numbers. Unless a calculation tool is actually running, treat the maths as a guess.
• Decisions that are actually about your values. Should you take the job, end the relationship, move city. It will produce a confident, balanced-sounding answer that has no idea what matters to you.
• Work that must be right on the first pass and where you cannot check it.

The three questions worth asking every time

1. Would I notice if this were wrong? If not, do not use AI unsupervised.
2. Is the source material in the prompt? If yes, reliability is much higher.
3. Am I asking it to retrieve, or to transform? Transformation is its strength. Retrieval is where it invents.

A rule about consequences

For anything medical, legal or financial, the useful framing is: use it to prepare, not to decide.

It is genuinely good at helping you understand terminology, work out what questions to ask, and turn up to an appointment better prepared. It is not a substitute for someone who is accountable for the advice they give, and it does not know your circumstances.

Try this: Pick one task you did this week that involved turning rough material into something tidy — notes into an email, a long thread into a summary. Redo it with AI and time both. If it saved you time, that shape of task is where your wins are. If it did not, you have learned something equally useful.