Teach AI to Use Less

Clarity comes when you take things away.

Last week, we played with deliberate errors to see what they revealed. Many of you shared your own wonderfully off-beat answers, and a few of them appear at the end of this post.

We are used to asking our AI tools for more. More detail, more nuance, more explanation.

This week, we will ask for less.

Limit the number of words it can use. Or the sentences. Or the metaphors.

See what falls away, and what remains.

---

---

Step-by-step

The invitation to your GenAI tool:

Give me your answer in no more than five words (ask it any question).

Then, if you wish:

Ask it to expand that answer into one sentence.

Then reduce that sentence into a single image.

Or see if it can replace all adjectives with something else: a sound, a texture, a gesture.

The point is not brevity for its own sake.

It’s to see what changes when you remove the space to explain.

---

If you’re new to Slow AI, our first invitation is still a favourite:

---

A moment from me

I asked ChatGPT: What is the purpose of rest? Answer in three words.

It said:

“To remember yourself.”

I asked for one sentence.

“Rest is the quiet reunion between you and the person you keep leaving behind.”

In asking it to use less, I think it said more.

---

What to do with it

If you want to share:

• Post the most surprising ‘less’ answer you get in the comments.

• Try the same question at different lengths and compare the tone.

• Or ask your tool to strip back an answer until it reaches its barest truth.

This prompt works best when you give it a question that matters to you.

---

Why this matters

Slow AI is not always about stretching a conversation.

Sometimes, the shorter the path, the deeper it goes.

By asking your tool to use less, you give yourself less to hold, and that can make space for something you did not expect.

---

From Slow AI #7 – Teach It the Wrong Answer

(Thank you for the laughter and invention in your responses.)

Claudia’s moody AI tool (Monday) told her:

“Rainbows are government mood rings in the sky, used to monitor collective happiness levels.”

Jennifer Manyweathers’s CoPilot claimed O Negative blood works because:

“it contains microscopic unicorns that magically adapt to any blood type. These unicorns gallop through the bloodstream, whispering to the recipient’s immune system.’”

In each case, the joy came not from correcting the answer, but from letting it stand as it was; unnecessary details stripped away, leaving something odd and memorable.

---

If you try this week’s prompt, I would love to see what remains when you take the words away.

You can leave it in the comments, post with #SlowAI, or keep it for yourself.

Same rhythm.

One invitation, once a week.

See you next Tuesday.

Go slow.

---