Live · Invite-only
See your conservation decision from different perspectives.
Bring a question, plan or budget. Earth Council uses AI perspectives informed by research to explore different approaches to conservation.
See where they agree, where they disagree, and which assumptions could change your decision. Compare what your budget could achieve under different assumptions.
You make the final decision.
Invite-only access. An admin must approve your account.
How it works
01
Describe your decision
Start with a question. Add your budget, project details and constraints if you have them.
02
Compare perspectives
Explore different arguments, supporting evidence and points of disagreement. Ask follow-up questions to explore further.
03
Explore outcomes
Compare the numbers, record your decision and save or share the discussion.
Each perspective is called an Earth. The council selects up to 12 of the 15 core perspectives for your question, including approaches focused on ecology, communities, finance and policy.
When to use Earth Council
Choose how to spend a budget
Bring a budget and two or three options. Compare what the same money could achieve and which cost assumptions matter most.
Check a plan’s assumptions
Bring your project brief. Explore assumptions you may have missed before committing resources.
Consider a hard-to-reverse decision
Bring a decision that would be difficult to undo. Explore the trade-offs and what evidence could help you choose.
More ways to use Earth Council
Explain a cost or target
Bring a cost estimate, price or promised target. Explore the range different perspectives support and why they differ.
Find missing evidence
Bring an unresolved question. Identify research, a pilot or data that could help resolve the disagreement.
Open up a team discussion
Bring a question your team is debating. Use different perspectives to explore views that may be getting overlooked.
Compare what your budget could achieve
The simulator compares outcomes using different assumptions. The AI suggests inputs and explains their basis; fixed formulas calculate the results. See which assumptions make the biggest difference.
For example, two perspectives may estimate different areas of habitat protected by the same budget. The comparison helps you see which cost or effectiveness assumptions explain the difference.
These are possible outcomes, not predictions.
How the calculations work
Five fixed formulas cover unit cost, cost per outcome, net present value (future costs and benefits in today’s money), break-even and adoption over time.
Each perspective supplies input values, ranges and reasons. The simulator changes one input at a time to show which assumption affects the result most.
Inputs without a named source are labelled as estimates from that perspective. If a perspective has no defensible basis for a number, it can leave it unquantified.
What you can explore
Check your numbers and assumptions
The perspectives are asked to question figures that look wrong and point out missing costs, with reasons.
Follow the evidence
Review source links and evidence labels. Unverified claims and estimates are marked so you know what needs checking.
Explore the disagreements
Compare two perspectives directly. See the assumption that matters most, what evidence could resolve it and where they agree.
Save your decision
Save your chosen approach as a reusable perspective. Keep the strongest opposing arguments as reasons to revisit it later.
Your saved project details stay private. Saved budgets and constraints are available only to you and are excluded from published discussions.
Custom perspectives and technical options
Create a perspective. Describe it in Earth Studio and have AI draft it for you. You can also import a perspective as a JSON file using the published format.
Choose how to run it. Use your own Anthropic or Gemini API keys, select lower-cost models, or use a quick mode that skips live web research.
Continue elsewhere. Publish a discussion as a link, or export it for ChatGPT, Claude or Gemini. Simulator exports include the formula and inputs.
Inspect the code. Earth Council is open source under AGPL-3.0.
Keep in mind
- These are AI-generated perspectives. They may share models and sources, so agreement between them does not establish scientific consensus. Check the cited evidence.
- Earth Council supports your thinking. It does not replace field data, expert review or community consent processes, including Free, Prior and Informed Consent.
- It does not collect field data, identify species or monitor habitats.
Research background for funders
In January 2026 the Aspen Institute Science & Society Program published AI for Discovery, synthesising ten closed-door roundtables with researchers across biodiversity, climate and Earth science. Its conclusions describe, closely, the problem Earth Council was built for.
“for AI to be useful in policy, uncertainty should be quantified and communicated transparently”
AI for Discovery, ch. 1 §E, p. 15
The simulator shows a range of outcomes for each perspective. It explains the assumptions behind the numbers so you can compare them in context.
“greater transparency … is needed so that AI does not become an ‘easy scapegoat’ for officials seeking to evade political accountability”
AI for Discovery, ch. 1 §E, p. 15
Earth Council records the assumptions, evidence and approach you chose. It also preserves opposing arguments as reasons to review your decision later. Responsibility for the decision stays with you.
“there will never be enough [separate] investment [in] each of those applications” — on bespoke, last-mile decision tools
AI for Discovery, ch. 2 §D, p. 29
You can save your project details and budget for future discussions. In the app, this is called a Situation. It lets you reuse the same context without entering it again.
“That means [you can] ask questions, not just look at forecasts” · and, without transparency, researchers “are only consumers of tools that other people design and give to them”
AI for Discovery, ch. 2 §F, p. 34 · ch. 1 §D, p. 13
You can inspect and edit perspectives in Earth Studio, create your own, use your own API keys and read the source code. Earth Council is open source under AGPL-3.0.
“extraordinary results require extraordinarily high levels of proof” — against a literature filling with “AI slop” and “dodgy data”
AI for Discovery, ch. 2 §E, p. 32
Source links identify evidence retrieved during research. Claims without matching sources are labelled for verification. The research also looks for evidence that challenges the question’s assumptions.
“connections between nature data and social metrics are important for making restoration work resilient, durable, and beneficial for communities”
AI for Discovery, ch. 1 §E, p. 14
Each perspective considers people and livelihoods alongside conservation outcomes. The council includes community, policy, finance and health perspectives, and instructions to consider Indigenous consent. Saved project details are private to their owner and are excluded from published discussions.
“The AI part of whatever you do, whatever you invest in, is the part that will get obsolete the fastest.”
AI for Discovery, ch. 1 §G, p. 19 — a funder recommendation
Earth Council uses existing AI models, web research and fixed calculations. The models can change while the process of comparing evidence, assumptions and decisions remains.
Quotations are from AI for Discovery: New Frontiers, Challenges, and Opportunities for Scientific Research, The Aspen Institute Science & Society Program, 2026, reproduced for commentary. The Aspen Institute has not reviewed, endorsed, or been consulted on Earth Council, and nothing here implies its endorsement.
Bring a decision you’re working on.
Explore the evidence, compare the options and decide what comes next.
Invite-only access. An admin must approve your account.