My Artificial Intelligence Policy (and how it aligns with permaculture)

Last updated: July 2026

As a permaculture-informed small business owner, I have serious concerns about the ethics and environmental impact of artificial intelligence.

I recognise that AI could be (maybe?) a tool-for-good, in the right hands. And, as an Ethical Marketing Consultant, I feel it’s wise to stay somewhat across emerging developments within my industry, which AI is currently putting through the wringer.

So I do use artificial intelligence, but sparingly, intentionally and with great care and consideration.

I actively resist AI as a pervasive, thoughtless or default tool within my business, and I largely avoid recommending AI solutions to my clients.

Below is part-manifesto, part-policy on my concerns with AI, the ways it contradicts permaculture’s ethics, how I currently do and don’t use it, and my expectations (well, requests, really) for how you will use it when engaging with my work.

Ethical concerns I have with AI

Broadly speaking, the way artificial intelligence tools are developed and operate directly contradicts permaculture’s three ethics and my own business ethics and values.

My grudge list is vast and multifaceted because AI:

  • Is trained on stolen creative work from artists and writers — literally, people like me
  • Negatively impacts the environment, using vast quantities of water and electricity
  • Relies on polluting data centres that are making life unbearable for nearby communities
  • Encourages an abandoning of human connection and boredom for sketchy efficiency
  • Often lies, gaslights or gets things wrong, contributing to a culture of misinformation
  • Allows the hoarding of power by a small number of billionaire-owned companies
  • Relies on labour exploitation from the Global South
  • Is playing a role in creating and abetting weapons of war
  • Is expanding everyday surveillance via unregulated data collection, eroding our privacy
  • Can lead to discriminatory outcomes for vulnerable or marginalised people
  • Enshittifies the internet and our brains, and is even changing the way we speak

I live in the driest state in the driest continent on earth (South Australia). I’m the person who collects greywater for reuse on the garden via a bowl in the kitchen sink and a bucket in the shower; water here is that precious. Flushing litres down the AI drain every day just to write an email feels criminal.

And then there’s the fetishisation of efficiency and speed.

Optimisation culture insists that ‘doing things quicker’ is the ultimate success in business and life.

But so much of the beauty in human experience — the pondering and close observation, ah-ha moments after a stumble, learning by having a go — gets trampled in the frantic AI-fuelled rush to be done, fast and automatically if possible, just so you can do more.

Is anyone truly satisfied by such a hollow and soulless life?

As Natalie Brite from DoGoodBiz writes:

“Speed is not neutral. When we use AI to accelerate design work, we are choosing what gets compressed. And too often, what gets compressed is exactly the part that matters most: the slow thinking, the genuine engagement with a specific human problem, the refinement that comes from sitting with something long enough to understand it.”

How I use AI in my permaculture business

I wish to be fully transparent about how and when I use artificial intelligence, while acknowledging that this policy is a living document that will grow and change alongside my business and me.

Types of AI

First up, let’s get on the same page about types of AI. Here’s a helpful list from Amelia Hruby:

  • Assistive AI — when AI provides review and support for original work (eg, Grammarly)
  • Generative AI — when AI creates the work, alone or conversationally (eg, ChatGPT, Claude, Gemini)
  • Agentic/autonomous AI — when AI independently executes workflows (eg, Notion Agent or AI schedulers)

Which AI tools I pay for (TL;DR: none)

At this stage, I don’t pay for any generative or agentic AI tools.

For as long as is reasonably possible, I don’t want my money funding the generative AI gold rush.

Instead, I take a ‘people over tech’ approach. My priority is to invest directly with real people and fellow small businesses, which I feel more closely aligns with permaculture’s Earth Care, People Care and Fair Share ethics.

But I’m also realistic. I know that assistive AI is increasingly baked into every digital tool (ugh). And I recognise that, by using certain free tools, I’m allowing myself to become the product, a “user” who pays through my attention and the on-selling of my data.

This makes me deeply uncomfortable.

An ongoing business consideration is how to reduce my digital footprint and tracking — by switching settings to private, not allowing my inputs to train AI models (where possible), keeping my tech stack lean and minimalist, and looking for more ways to work offline.

What I do use AI for

In 2024 and 2025, I lightly experimented with generative AI — I wanted to know the beast I was up against. But the results were mostly sub-par.

Worse yet, I hated how it boxed me into a particular way of thinking, or edited out ways of expressing myself that, while perhaps more wordy than its clipped and polished prose, also felt more human and more me.

I know AI will only keep improving. But my experience thus far, combined with serious ethical concerns, means I am committed to running an AI-light business.

So, these days, Claude is the only generative AI tool I use, and I remain on the free plan.

What I use Claude for

  • Spotting patterns — When analysing large datasets (such as survey responses), a machine is definitely better at collating trends than I am.
  • Search engine optimisation (SEO/AIO/GEO) — Writing metadata, tweaking headlines and suggesting webpage edits to optimise around specific keywords, but I still do all the keyword research myself; AI is terrible at that bit.
  • Transcribing text from images — It’s funny how often this need comes up in client work; Claude saves me from retyping it all by hand.
  • Basic coding on my website — I’m aware of the security risks and accessibility issues when using AI to code (big ew to vibe coding), so I only use Claude to make simple tweaks, which I can sense-check using the basic CSS knowledge I already have within, y’know, my brain.

Other AI tools I use

  • Spellchecking while writing — Grammarly, but only on the free plan; the paid version gets very pushy about wanting to rewrite your work into that awful AI monotone.
  • Notetaking in video calls — Fathom generates transcripts and summaries of my live calls to ensure accurate records of client conversations.
  • Closed captions for YouTube — I use Rev to create YouTube subtitles so my videos are accessible for people with hearing impairments.
  •  

For more on this, browse the complete list of ethical(ish) tools I use in my business.

What I don’t use AI for

Given my permaculture ethos and ‘people over tech’ mentality, I’m mindful of not using AI to cut corners or avoid paying someone what they’re worth.

I’ve specifically ruled out AI for:

  • Giving marketing or systems advice — Any recommendations I make to the general public or my clients are based on my 20+ years of industry experience and my near-obsessive (lol) commitment to continual learning.
  • Writing — 90% of what I write, be it client work, journalism stories, Instagram posts or emails, comes directly from my human brain to yours, with no machine in the middle.
  • Research — I conduct all of this myself. Do you know how error-ridden those bloody AI tools are?
  • Trained chatbots — I have no intention of outsourcing my approach and experience to a shitty chatbot and selling that to you as though it’s just as good as working 1:1 with me.
  • Scheduling and automations — I find all that stuff dehumanising, and instead prioritise and value real human connection across my business.
  • Photos, graphics, videos or any type of visual imagery — AI imagery gives me the ick (it’s uncanny valley territory), and its creation unnecessarily guzzles water; I use my own photographs or graphics made in Canva. I do occasionally use free stock images, which I hope (but can’t guarantee) aren’t AI-generated.
  • Paid digital advertising (for now) — The algorithmic overlords at Meta and Google have long been ethically questionable; now their advertising platforms increasingly rely on automated, AI-driven optimisation. Instead, I have a referrals budget.
  • Teaching — I won’t be spruiking AI strategies or leading workshops about AI practices, nor will I host events with others who heavily promote using AI.
  • Legal things — Yikes. Just, no. Stuff like contracts and refund policies must be legally rigorous to protect my little biz, which means involving real lawyers. I use Legal123.

What I ask of you

This next section is for my readers, my clients, my collaborators and anyone else who stumbles across my corner of the internet.

I’m obviously not the AI Police, though I kinda wish they existed to rein in the harms of tech oligarchs.

Each of us must weigh up the options and impacts of AI, then make our own choices.

But I hope that, by being transparent about my own AI use, and explaining some of the real harms overuse creates, you might find motivation for some soul searching of your own.

I like how Andrea Reeves of Type Creative puts it:

“As the use of AI continues to become more and more common, you’re allowed to resist or ‘give in’ in whatever ways feel most aligned with your needs and values. Overall, I believe it’s important to continue questioning the norms of AI adoption, and to use it as intentionally and responsibly as possible.”

Don’t upload my work to AI tools (please)

I’d like to be clear on one matter, though: I do not consent to you taking my work (public or otherwise) and feeding it into any AI tools, LLMs or Chatbots. That is theft.

Please don’t upload any of my materials and resources — from korenhelbig.com, Permaculture Marketing, my journalism stories or anywhere else — to AI for any kind of use.

The two exceptions to this rule are:

  1. Content I’ve created specifically for 1:1 clients; after payment, that becomes yours to use however you like, AI tools included. (My proprietary templates, workbooks and frameworks fall outside of this.)
  2. Folks who use AI-powered text-to-speech apps or similar accessibility tools; if these make my work easier for you to access (without taking my content and turning it into something else), that’s a win in my book.

Obviously, I don’t have any real power to stop you from doing this.

But I believe in a world based on honesty and kindness, and I consider this a social contract between us, you and me, two fellow good humans surfing the internet seas with integrity and caring consideration for one another.

How to create your own ethical AI policy

Perhaps your mind is now humming with ideas on how to be more transparent and ethical with AI use in your own business?

If you’re ready to draft your own AI policy, here are some helpful resources:

If you have any questions or feedback about my permaculture-infused AI policy, please do contact me. I’d be all ears, and you know you’ll get a thoroughly human, AI-free reply :)