Our AI and Digital Tools Policy

Our purpose is simple: to help people build deeper relationships with the living world, and to help turn that care into learning, celebration, advocacy and action.

We are not trying to increase engagement with a platform for its own sake. We are trying to support deeper engagement with beings, ecosystems and the living world itself.

This page is our short public promise about how we use AI and digital tools, why we use them, how we try to keep them in bounds, and how we plan to improve over time.

Why we use AI

We are a small charity with big ambitions and limited resources. Used carefully, AI can help us offer useful tools sooner than we could otherwise afford.

It can help translate ecological information into plain English. It can help people discover a being, learn about it, and find ways to connect, celebrate, support or stand beside it. It can also help us meet people where they are — whether that person is a professor of ecology, a teenager on a phone, or someone who simply wants to know more about otters, fungi, rivers, bees or moss.

But AI is a means, not an end. The dream is not more AI. The dream is more relationship, more care, more time with the living world, and less time staring at screens.

How we use AI right now

We have moved to GreenPT as our Phase 1 AI provider. Based on what we currently know, we think it is the best available, most ethical and currently affordable option we have identified for ISB.

AI helps generate and support our shared Being Profile Cards — pages about a species, habitat, river or other being. Used well, this can be quicker, more focused and potentially less energy-intensive than repeated open-ended searching.

Your more personal Being–Besider page can include optional AI help, but it does not have to. Where practical, we want people to have a real choice to write, edit or complete things themselves.

We may also use AI to help with search, translation into plainer language, and suggestions for learning, connection, celebration, support or action. Over time, we want fewer free-floating AI guesses and more grounded, checked and sourced information.

We do not use AI because it is trendy. We do not want manipulative, addictive or dependency-building systems. We do not want AI to replace human judgment where safeguarding, privacy, accuracy or care really matter.

What you may see on the platform

You may see AI-assisted content where it is useful; citations, links or source trails where possible; a way to report errors or concerns; optional AI help in some places; and a rough CO₂e estimate before some AI actions.

For example, generating a Being Profile Card may currently be in the region of 100–150 mg CO₂e. That is an estimate, not an exact measurement, roughly equivalent to about one grain of rice.

That may sound tiny, and per card it is. But impacts add up. That is why we try to count them, reduce them and design with restraint. It is also one reason to reduce unnecessary screen time, since every second spent online has an energy and carbon cost. We plan to offer lower-energy and lower-bandwidth versions of the website in the near future.

Our AI impact checks

Because the AI landscape changes quickly, we review our choices through a set of impact lenses:

🌍 Energy and emissions
How much energy does this use, from what sources, and how much CO₂ are we generating? What are the impacts of the energy infrastructure?

💧 Water and heat
How much cooling, water use or heat impact may sit behind the system? We want to avoid systems reliant on evaporative cooling where possible, and favour closed-loop systems with heat recovery.

🧱 Materials and e-waste
What hardware does this depend on? What waste does that create? What are the impacts of extracting the raw materials?

🧑‍🏭 Labour and supply chain
What human labour and supply chains sit behind the interface?

🕸️ Wider system effects
Who owns and funds the tools? What power structures are we strengthening or resisting? What investments are the hosts making?

🔓 Sovereignty and exit-ability
Can we leave? Can we take our data and knowledge with us? Are we locked in?

🧠 Psycho-social harms
Could this create cognitive atrophy, dependency, manipulation or simply more screen time?

⚠️ Hallucination and false authority
Could it sound confident while being wrong?

🔐 Organisational and security risk
What are the risks around security, reliability, abuse and operations?

⚖️ Rights and compliance
What about privacy, consent, data protection, intellectual property and legal duties?

🔍 Transparency
What do we really know? What is published, accessible and verifiable? What do we think we know, and what remains uncertain?

These checks matter because the AI landscape is shifting all the time. A provider that seems good today may look different tomorrow. So we keep reviewing.

Our direction of travel

We are not trying to build a future where everything is run by AI. We are trying to use a flawed tool carefully, while moving towards better alternatives.

🚦 Phase 1 — GreenPT now
Use the best ethical and currently affordable provider we have identified so far. We understand that GreenPT uses renewably powered servers and small models, and we will provide a fuller due diligence report in due course.

📚 Phase 2 — Build our own knowledge base
Create a richer library of checked information through partnerships with natural history datasets such as GBIF, cached content and sourced data, so we rely less on AI guesswork.

🖥️ Phase 3 — Greater sovereignty
Longer term, we want more control over our own green infrastructure, including our knowledge base and, if feasible, more of our hosting stack.

🧬 Phase 4 — ISB-trained or ISB-shaped AI
In the long run, we may explore training or adapting our own AI — but only with proper care, lawful data, clear ethics and strong governance.

Our bottom line

We use AI because it helps us do some good that we could not yet do otherwise.

We know AI comes with real harms and trade-offs.

We are trying to use it with restraint, transparency and a clear plan to improve.

Above all, we remember that the screen should be a bridge, not a home.

Questions, comments or concerns?

Please email us at info@istandbeside.life

We also plan to open a public comment link for the fuller paper, so anyone who wants more detail — or wants to challenge us — can do so.