Just this week, OpenAI launched GPT-5, their most advanced artificial intelligence (AI) model yet, promising access to "PhD-level intelligence" for everyone. Meanwhile, you’re trying to get up to speed on a complex issue like recycling post-consumer waste plastic, but every article you find is either too technical or oversimplified. Here's the paradox of our time: Do you take advantage of the newest technology to help you make well-informed decisions about environmental challenges, or do you avoid AI but take longer to gather the knowledge you need to reach a deeper understanding of our ever-increasing environmental crises?

Environmentalists have legitimate concerns about AI’s energy consumption and footprint on our landscape. Data centers that consume enormous amounts of electricity and water for cooling are being built to develop even smarter AI systems. Take a step this week to learn how you can use AI for enlightenment and not just entertainment, so the benefits of your AI sessions can offset their environmental impacts.

What Is AI and How Can You Use It?

Artificial intelligence “chatbots” are computer programs that can have conversations with you about any topic. Think of them as astoundingly well-read, but sometimes over-eager, research assistants who have absorbed millions of books, articles, pictures, videos, and websites, then learned to explain complex topics in whatever style you prefer. Simply type into a form on a website or speak to an app on your phone (“prompting” the AI model); you'll get a detailed, customized answer. The same prompt to the same chatbot will generate a different response each time.

The most popular AI tools are free to try and work a lot like texting or talking to a colleague. OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini are three options, each offering a simple signup. You can ask them anything from “Explain how solar panels work” to “Is drinking oat milk or almond milk better for the environment?” and receive responses tailored to your exact needs. The key is asking good questions and—crucially—verifying the answers, since these tools can “hallucinate,” meaning they can occasionally make up incorrect information with complete confidence.

Quick Start: How to Use AI for Learning

Understand What AI Is Really Good At. You won’t be able to find a person in the world who is better read than an AI model. They have literally “read” every book ever published and have stored a memory trace of every bit of information they have ever encountered. Different models have different “personalities.” Most of the ones provided for public use are trained to do their best to try to understand what you want them to do and respond in a helpful and respectful way. They are especially good at explaining both sides of a contentious issue and considering every angle of a problem. If you’re having a hard time understanding why your crazy uncle says the things he says about climate change, you can put that question to an AI and find out. Most AI tools can be incredibly fair and balanced when discussing any sensitive topic.

Choose the Latest AI Models for Better Accuracy. Hallucination, occasionally making up things that are not true but seem plausible, can be a problem with AI. New models tend to hallucinate less frequently than older models, so it's better to use a newer model. GPT-5, launched on August 7, 2025, reasons better and hallucinates less than GPT-4, according to OpenAI, the company that created it. However, even the most advanced AI models will sometimes still make mistakes, so you can't completely trust any of them. They are great study buddies, but just like your study buddies in college, they aren't perfect. Your critical thinking skills are important to apply when you're using AI. If it sounds too good to be true, it probably is.

Ask for Research and References. Very old AI models (from way back in 2024) are not able to do web research, so their answers to your questions are based on their vague recollection of everything they have digested during their training runs. It is interesting to chat with these older models because they provide a very good consensus view of any topic. But when you ask them specifics, they'll often make up plausible-sounding answers that are wrong. New models (released in 2025) can do online research and what's called “reasoning.” If you tell them to do research and show their work, they can search the Internet and find references that help ground their answers. You’ll still have to check their references. Just because it’s published on the Internet doesn't mean it’s true!

Request Learning-Level Explanations. Tell the AI model exactly what level of understanding you want. For example: “Explain planetary boundaries in simple terms a high school student would understand” or “Give me a technical explanation of lifecycle assessment methodology suitable for someone with a college science background.” This prevents either oversimplification or unnecessary complexity.

Always Ask for Sources and Expert Names. Explicitly request references to scientific journal articles and the names of scientists who are experts in the subject. This improves response quality and gives you starting points for verification. You might prompt: “What are the different scientific viewpoints on how quickly we need to achieve net-zero emissions? Please include references to peer-reviewed studies and name key researchers in this field.”

AI Tools and Terms

Latest Models to Try (Some Require Paid Subscriptions):

  • GPT-5 (chatgpt.com) - OpenAI's most advanced model with PhD-level reasoning capabilities

  • Claude Opus 4.1 (claude.ai) - Anthropic's latest model, strong at analyzing documents and providing nuanced explanations

  • Gemini 2.5 Pro (gemini.google.com) - Google's thinking model with real-time web access for current information

Key Terms to Know:

  • Chatbot: An AI model that can have a conversation with you

  • Prompt: Your side of the conversation; the AI model generates a response

  • Hallucination: When AI generates plausible-sounding but false information

  • Token: A fragment of text, sound, or image that an AI model can process

  • Context window: How many tokens an AI can process at once

  • Reasoning: The ability of an AI model to "think" as it's generating its response

  • Tool Use: The ability of an AI model to do things like search the web

  • Deep Research: The ability of some AI models to prepare research reports

Intermediate: Strategic AI Learning for Sustainability

Seek Multiple Perspectives on Complex Issues. Environmental science often involves legitimate scientific debates. Ask AI to present different viewpoints: “What are the main arguments for and against nuclear energy as a climate solution, and what does the scientific evidence say about each position?” This helps you understand the full landscape of scientific thought rather than getting a one-sided view.

Request Practical Applications. Connect theoretical knowledge to actionable practices: “Based on what you’ve explained about ecosystem services, what can I do in my home to help maintain biodiversity and ecosystem health?” This bridges the gap between understanding environmental science and implementing sustainable practices.

Use Follow-up Questions for Deeper Understanding. In an AI chat session, you can build on previous responses. In a conversation about sustainable agriculture, you might follow up with: “To add to your response about feed lots, explain more about the connection between livestock grazing and soil carbon storage.” You can explore topics systematically and thoroughly.

Ask AI to Improve Your Prompts. Ask AI to help you ask better questions: “How can I improve the questions I’m asking to get more accurate and useful information about renewable energy systems?”

Advanced: Verification and Integration Strategies

Cross-Reference Multiple AI Systems. Compare responses across different AI models (ChatGPT, Claude, Gemini, etc.). When multiple systems give similar answers backed by the same scientific sources, you can have higher confidence in the information.

Verify Against Credible External Sources. Before taking action based on advice from an AI model, check against peer-reviewed articles, reports from established scientific organizations (like NASA, NOAA, or the IPCC), or authoritative references like Sustainable Practices: Your Handbook for Effective Action. Use AI as a starting point, not a final authority.

Focus on Science-Based Claims. Place higher trust in knowledge you can trace to specific research studies, government data, or established scientific principles. Be especially skeptical of claims about new technologies, recent policy changes, or local sustainability issues where AI’s training data may be incomplete or outdated. AI’s desire to please you could lead it to hallucinate. Recognize that you can badger some AI model into simply agreeing with you rather than helping you truly understand an issue. If you desperately want affirmation, AI can give it to you. But it’s more valuable to gain insight and ask AI to challenge your beliefs to make sure they stand up to scientific scrutiny.

Expert: Community Learning and Teaching

Organize AI-Assisted Study Groups. Use AI to generate discussion questions, explain complex concepts to groups with mixed backgrounds, and synthesize information from multiple perspectives. Share verification responsibilities among group members, with each person fact-checking different aspects of AI responses.

Develop Local Sustainability Resources. Since AI may not accurately represent local sustainability issues, use it to understand general principles, then research local applications through municipal websites, local environmental organizations, and regional academic institutions.

Train Others in Responsible AI Use. Share strategies for effective prompting, verification techniques, and critical evaluation of AI responses. Help others understand both the potential and limitations of AI for environmental education.

The Environmental Cost-Benefit Analysis

Here’s the reality check: every AI query consumes energy. According to recent research from Epoch AI, a typical ChatGPT query uses about 0.3 watt-hours, far less than the 3 watt-hours often cited from older estimates. To put this in perspective, brewing a single cup of coffee requires about 70 watt-hours, meaning that drinking one less cup of coffee frees up enough energy to make over 230 AI queries.

When deciding how to gain knowledge, consider your alternatives. Searching the web, driving to a library, and attending conferences—all have environmental impacts. If AI helps you learn faster and implement more effective sustainability practices, the environmental benefits may far outweigh the energy costs of the queries that enable them.

Making Peace with the Paradox

Using AI for sustainability education means accepting a fundamental paradox: leveraging resource-intensive technology to learn how to live more sustainably. This isn’t hypocrisy—it's an investment that empowers you to take action with a larger positive impact.

Environmentalists have always used the tools available to advance their cause, from fossil-fuel-powered vehicles to attend climate protests to electricity-powered computers to organize campaigns to improve energy efficiency. AI is simply the latest tool. If AI helps you understand heat pump technology and explain it well to your family, leading you to replace a gas furnace, the emissions saved will dwarf the energy cost of your learning.

Use AI to Gain Wisdom

This week, try using AI for one specific sustainability learning goal. With GPT-5 now available, you have access to more powerful reasoning capabilities than ever before. Maybe you want to understand how carbon markets work, or you’re curious about the science behind forest carbon sequestration, or you need to explain climate change to a skeptical audience. In your prompt, request credible science-based sources and multiple perspectives. Use the AI response as a starting point, then check references to verify key claims.

AI won’t solve our environmental crisis for us, but it can help us understand the science, technologies, and strategies we need to solve it ourselves. The future of sustainability education isn't about choosing between AI and traditional sources—it’s about using AI strategically while maintaining our critical thinking skills so we can separate helpful knowledge from any harmful hallucinations. In an era of information overload, that might be the most crucial sustainability skill of all.

References and Resources

AI Tools and Platforms

  • OpenAI ChatGPT - GPT-5 released August 7, 2025, with improved reasoning and reduced hallucinations

  • Anthropic Claude - Claude Opus 4.1 released August 5, 2025, with advanced reasoning capabilities

  • Google Gemini - Gemini 2.5 Pro launched in March 2025 with thinking capabilities and real-time web access

  • Perplexity AI - AI search engine that provides sources with responses

AI Energy Impact Research

Verification and Fact-Checking Resources

Sustainability Education Resources

Data Center and Energy Infrastructure