Portfolio note
What This Sample Demonstrates
This independent strategy sample demonstrates my approach to complex organizational decisions: translating emerging issues into practical guidance that is strategic, values-aligned, and operationally realistic.
For a busy reviewer: this sample is meant to show judgment, not just fluency. It turns a fast-moving, contested topic into a usable decision tool for nonprofit staff, development teams, program leaders, and executives.
Context
Why This Matters
AI is already entering climate and mission-driven work through research synthesis, grant drafting, donor research, CRM cleanup, translation, accessibility, communications, workflow automation, environmental modeling, emissions analysis, and conservation tools.
At the same time, AI is not only a software question. It is also an infrastructure question. AI systems depend on data centers, electricity, water, land, minerals, hardware supply chains, labor, cooling systems, and local permitting. The International Energy Agency projects global data center electricity consumption could roughly double by 2030 to around 945 TWh in its base case. WRI has also described data center growth as reshaping local energy grids, water systems, and land use in the United States, often amid limited public information about long-term costs and benefits.
For climate and mission-driven organizations, this creates a dual responsibility: stay open to AI uses that genuinely strengthen the work, while avoiding adoption that obscures or externalizes environmental, privacy, labor, or community harms.
30–60 second skim
Executive Summary
Use AI when it helps the mission
AI may be useful when it improves quality, speed, access, analysis, or climate outcomes in ways staff can verify and govern.
Do not treat AI as immaterial
Environmental footprint, data centers, water, energy, labor, community impacts, and vendor practices are part of the decision.
Match review to risk
The more public, sensitive, consequential, or resource-intensive the use, the more oversight it needs.
Operating standard: Use AI when it helps the work and the risks are manageable. Do not use AI when it creates unacceptable risks to accuracy, privacy, community trust, environmental responsibility, labor standards, or organizational integrity.
Leadership tool
Decision Tree for Evaluating AI Use
Define the task
What problem are we solving? Who will use the tool? Who will be affected by the output? Would a simpler workflow, spreadsheet, database, template, or conventional software tool work as well?
Pause if: the problem is vague or the main benefit is novelty.
Classify the risk
Is this low-risk internal support, moderate-risk organizational/public-facing support, or high-stakes use involving sensitive data, public claims, procurement, or decisions affecting people or communities?
Rule: the more public, sensitive, consequential, or resource-intensive the use, the more review it needs.
Weigh benefit and harm
What value does the tool create? How will we know whether it helped? What could go wrong? Could the output be inaccurate, biased, generic, misleading, privacy-invasive, or damaging to trust?
Add safeguards
Use human review, source verification, data restrictions, manager approval, legal/compliance review, environmental and community-impact review, disclosure, or periodic reassessment as appropriate.
Reconsider if: the safeguards are too burdensome to maintain.
Name accountability
Who owns the final output? Who verifies factual claims? Who approves external use? Who decides whether disclosure is appropriate? Who can pause or stop use if risks become unacceptable?
Applied judgment
Three Sample Decision Scenarios
Grant drafting, donor research, and CRM cleanup
A development team wants to use AI to draft proposal outlines, summarize public funder materials, improve stewardship language, and help clean non-sensitive CRM fields.
Potential value
Faster first drafts, more consistent templates, better internal workflows, and more staff time for strategy and relationship management.
Key risks
Confidential donor data exposure, inaccurate funder claims, generic language, overreliance on AI-generated prospect summaries, or loss of organizational voice.
Recommended path
Use with safeguards. Keep sensitive donor and prospect data out of unapproved tools. Require human review, source verification, and manager approval for external-facing grant or donor materials.
Translation, public education, and listening-session synthesis
A program or communications team wants to use AI to translate public materials, draft educational copy, or summarize notes from community listening sessions.
Potential value
Improved accessibility, faster plain-language drafts, multilingual support, and better synthesis of large amounts of input.
Key risks
Mistranslation, flattening community voice, privacy concerns, extractive use of testimony, or replacing engagement that should happen directly with people.
Recommended path
Use with safeguards or formal review. Public translations should be reviewed by fluent speakers. Community data should not be entered into unapproved tools. AI should support engagement, not simulate consent or lived experience.
AI vendor for methane detection, emissions modeling, or resilience planning
A climate organization is considering an AI-enabled vendor or platform that could support emissions analysis, methane detection, environmental modeling, or resilience planning.
Potential value
Earlier detection, better prioritization, stronger analysis, improved decision support, and potentially meaningful climate benefits.
Key risks
Opaque model assumptions, false precision, inequitable impacts, vendor lock-in, weak data protections, or environmental costs from data center energy and water use.
Recommended path
Formal review. Validate model performance, clarify data rights, assess energy/water/emissions disclosures, evaluate community impacts, and define what would trigger contract termination or tool retirement.
Practical categories
Risk Table
| Category | Examples | Primary safeguards |
|---|---|---|
| Probably Appropriate | Internal brainstorming, first-pass internal templates, summaries of public non-sensitive documents, grammar support, formulas, plain-language summaries, alt-text drafts, non-sensitive workflow planning. | Basic staff guidance, human review, source checking for factual claims, no sensitive data in unapproved tools. |
| Use With Safeguards | Grant drafts, donor research, CRM cleanup, public explainers, translation, policy summaries, community survey synthesis, campaign materials, program evaluation, emissions or conservation analysis. | Manager review, privacy controls, source verification, documentation for high-visibility uses, disclosure when trust or accountability would benefit. |
| Probably Avoid | Uploading confidential data into unapproved tools, final decisions about funding/hiring/eligibility, unverified scientific or legal claims, simulated community consent, synthetic testimonials, misrepresenting real people or places, endorsing AI infrastructure without due diligence. | Avoid unless there is a compelling reason and formal review. Require senior accountability, legal/privacy review, equity review, and clear stop conditions. |
Staff handout
One-Page Staff Checklist
- What task am I using AI for?
- Is this a real need or just convenience?
- Would a simpler tool or process work?
- Am I entering donor, employee, legal, financial, partner, community, or unpublished information?
- Is this tool approved for that data?
- Can the vendor store, reuse, or train on what I enter?
- What claims need independent verification?
- Could the output be outdated, biased, incomplete, or wrong?
- Do I understand the topic well enough to review the output?
- Who benefits from this use?
- Who could be harmed or misrepresented?
- Could this affect frontline communities, Tribal nations, workers, ratepayers, or local water and energy systems?
- Does this replace engagement that should happen directly with people?
- Who needs to review it?
- Should we document the use?
- Should we disclose AI involvement?
- What would make us revise or stop using this tool?
Vendor review
Vendor and Procurement Due-Diligence Questions
Energy, water, and emissions
- Where are the relevant data centers located?
- What grids power them?
- Does the vendor disclose energy use, water use, emissions, and cooling practices?
- Does the vendor use credible, location-aware clean energy procurement?
- Are water-stressed regions involved?
Community and ratepayer impacts
- Could the tool or vendor increase local utility costs?
- Who pays for transmission, substations, water infrastructure, or backup power?
- Are Tribal nations, local residents, workers, or frontline communities affected?
- Are community benefits concrete and enforceable?
Data and privacy
- Can the vendor use organizational data to train models?
- What data are stored, retained, deleted, or shared?
- Are there enterprise controls, audit logs, permissions, and deletion rights?
- Is the tool approved for the type of data staff want to use?
Labor, supply chain, and exit
- What labor practices support the tool?
- What is known about data labeling, content moderation, hardware manufacturing, mineral sourcing, and e-waste?
- Can the organization terminate the contract if risks become unacceptable?
- Is there a review process after implementation?
Closing frame
AI as a Judgment Question
For climate and mission-driven organizations, AI is not only a technology question. It is a judgment question.
AI may help organizations work faster, improve access, detect emissions, model risk, and reduce administrative burden. It may also intensify energy demand, water stress, extraction, surveillance, misinformation, labor concerns, and community distrust.
A responsible AI roadmap helps an organization stay clear-eyed about both realities. The central task is to use powerful tools in ways that deepen climate progress, public trust, environmental justice, and organizational integrity — rather than undermining them.
Selected public sources
Sources Used for Context
These public sources informed the context for this independent sample.
- International Energy Agency, “Energy demand from AI” — global data center electricity consumption projection.
- World Resources Institute, “7 Ways Data Centers Affect US Communities” — U.S. grid, water, land, and community impacts.
- Climate Justice Alliance, “AI Data Centers” — environmental justice concerns related to data center growth.
- Environmental Defense Fund and Nuveen, “Decoding data centers” — investor due-diligence questions across data center environmental and social impacts.