Artificial intelligence is changing the world of gardening and agriculture! This column focuses on residential gardening, so for now we’ll skip recent stories about AI in agriculture, a large-scale version of home gardening.

The earliest uses of AI for gardeners emphasize the identification of plants and plant diseases. Applications that work on a cellphone are convenient tools for raising “what is this” questions while walking in the garden. In a future column we will review some of these simple applications. For larger issues, gardeners have access to generative artificial intelligence, which responds quickly to detailed and possibly complex issues. This column provides a guide to framing such issues to prompt a constructive response from artificial intelligence.

This involves a new skill for gardeners, called prompt engineering. While some people are building careers as prompt engineers, everyday gardeners can create garden-related prompts that will generate useful results.

We’ll begin this skill-building after presenting today’s image gallery: examples of plants that are flowering in my garden now, in the early spring. We always keep the real garden in mind! Our guide to building AI prompts to yield constructive responses is based on article by technology journalist Becca Caddy, including eight key points. Her article is oriented to the widely used tool, ChatGPT, developed by OpenAI, and applies to other available AI tools. See “Advance your garden knowledge” below for links to her article and other relevant websites.

This column includes our garden-related versions of Caddy’s eight key points for engineering AI prompts.

Give examples

In your prompt, include two or three brief examples of the response you would like. Ideas include “design for a foundation bed,” “plan for a cottage garden” or “herb garden near the kitchen.” Such examples tell AI about the style of the response you want.

Specify the audience

Tell AI for whom the response is intended. Assuming you are the intended audience, indicate yourself. Ideas: “For a beginning gardener,” “for a skilled gardener” or “for garden designer.”

Explain who you are

Identifying your need will lead to a better structured response. Ideas: “I need to renovate a neglected garden,” “I have little time for gardening” or “my new home’s yard is only bare dirt.”

Tell it the format you want

Describe your desired form of AI’s response. Options include an article, conversation, work plan, bullet points, categories or guide. You might have other formats in mind.

Define the tone

This key point enables you to shape the tone of AI’s output. Ideas: “A garden designer’s report,” “advice from a gardening friend,” “a landscape engineer’s plan.”

Use natural language

Keep your prompt clear, direct and conversational. AI tools respond to natural language, so technical language is not necessary.

Ask the AI tool to show its thought process

Invite AI to explain its thought process for preparing its response to your prompt. (AI doesn’t think like a human, but it does have a process.) AI’s explanation of how it developed its response could help the prompt engineer (you) to understand how and why AI responded.

Refine as you go

Approach prompt engineering as a process, rather than a “one-shot” effort. If AI’s response lacks clarity or completeness, or is otherwise unsatisfactory, edit your prompt and try again.

Here’s an example of an AI prompt that includes all these key points: Provide a plan for a 50-by-100-foot residential garden. Respond with a garden plan, like a design for a foundation bed, a plan for a cottage garden or a plan for an herb garden, intended for an experienced gardener. The goal is to renovate an overgrown and neglected garden over the next year. A skilled garden design should present the plan and organize the sequence by month or calendar quarter. Indicate the steps for developing the plan.

This prompt includes the first seven of the key points. After reviewing AI’s response, I might decide to use the eighth key point: revising the prompt and trying it again.

Advance your garden knowledge

For an introduction to a widely used AI tool, read the brief article, “ChatGPT explained,” available through this link: techradar.com/news/chatgpt-explained. To try your own approach to prompt engineering, sign-up as a user of Open AI’s ChatGPT: tinyurl.com/ycx8xrk5.

Read Caddy’s article, “I’ve become a ChatGPT expert by levelling up my AI prompts”: tinyurl.com/3arezh4b.

This week in the garden

Today’s column invites the reader to advance into the information age of gardening. The challenging part of using generative artificial intelligence tools to address a problem is presenting the problem fully and accurately. You can present your problem effectively, with a little practice. That practice could be your project for this week in the garden. If you use AI to address a need in your garden, write to us about your experience for sharing with other adventurous gardeners.

Enjoy your garden!

Tom Karwin is a past president of Friends of the UC Santa Cruz Arboretum and the Monterey Bay Iris Society, a past president and lifetime member of the Monterey Bay Area Cactus and Succulent Society and a Lifetime UC Master Gardener (certified 1999-2009). He is now a board member of the Santa Cruz Hostel Society and active with the Pacific Horticultural Society. To view daily photos from his garden, facebook.com/ongardeningcom-5665117 63375123. For garden coaching info and an archive of previous On Gardening columns, visit ongardening.com.