Mihaela Chiselita

ClimateGPT

ClimateGPT

ClimateGPT

An AI tool that empowers climate policymakers with transparent, ethical, and human-centered decision-making support.

An AI tool that empowers climate policymakers with transparent, ethical, and human-centered decision-making support.

An AI tool that empowers climate policymakers with transparent, ethical, and human-centered decision-making support.

ClimateGPT logo
ClimateGPT logo

A speculative AI-assisted platform that helps climate policymakers make faster, more transparent, and ethically grounded decisions through human-AI collaboration.

Client:

Erasmus AI

My Role:

UI/UX Designer

Year:

2024

Service Provided:

UX/UI Design

The Challenge

Climate policy requires urgent decisions, but many AI tools built to support it lack trustworthiness and clarity. Policymakers work under pressure and often face systems that obscure their logic or reinforce bias. I wanted to explore a different kind of AI—one that supports human agency, reveals its own assumptions, and respects ecological complexity.


My Approach

ClimateGPT is a speculative design prototype that reimagines how generative AI could operate in public service. Rather than optimizing for speed or automation, the tool prioritizes explainability, environmental accountability, and collaboration. It invites users to reflect—not just execute—and reframes AI as a transparent partner in climate governance.

I grounded the project in research, expert interviews, and prototyping. I analyzed existing platforms like Erasmus.AI, FloodAI, and ClimateGPT (2023), interviewed climate advisor Paul Vetter and governance researcher Sara Saberi, and reviewed literature on AI ethics, public trust, and posthuman design. My goal was to balance critical design framing with realistic, actionable functionality.

The Process

The project began with speculative interface sketches exploring refusal and resistance—tools that challenge rather than accelerate decision-making. A SWOT analysis of comparable platforms revealed gaps in usability, transparency, and bias mitigation.

My interviews highlighted three consistent needs: trust, traceability, and simplicity. Feedback from Vetter and Saberi led me to simplify the workflow, clarify system feedback, and reduce abstraction in key interactions.

Through iterative prototyping and usability testing, I gradually moved from provocation to functionality—keeping the tool ethically assertive but accessible to actual public sector users.

The Process

The project began with speculative interface sketches exploring refusal and resistance—tools that challenge rather than accelerate decision-making. A SWOT analysis of comparable platforms revealed gaps in usability, transparency, and bias mitigation.

My interviews highlighted three consistent needs: trust, traceability, and simplicity. Feedback from Vetter and Saberi led me to simplify the workflow, clarify system feedback, and reduce abstraction in key interactions.

Through iterative prototyping and usability testing, I gradually moved from provocation to functionality—keeping the tool ethically assertive but accessible to actual public sector users.

Key Features

Prompt Reframing offers subtle edits to user input that promote climate-conscious language and thinking without interrupting flow.
The Energy Cost Meter displays the estimated carbon footprint of each AI query, surfacing the environmental costs of digital decision-making.
The Reasoning Panel reveals the sources, logic, and confidence behind each AI-generated output.
The Two-Page Export Tool helps users produce lightweight, shareable policy drafts based on real tone and context preferences.
Finally, the Environment Scenario Tool invites users to explore decisions from non-human perspectives, shifting the frame from anthropocentric policy to planetary thinking.


My Role

I led UX strategy, system logic, and interface content. I designed the flows for key features, structured the platform’s architecture, and authored the tone and interaction language. I also synthesized stakeholder insights into practical design decisions—especially where ethics and usability had to meet.

Key Features

Prompt Reframing offers subtle edits to user input that promote climate-conscious language and thinking without interrupting flow.
The Energy Cost Meter displays the estimated carbon footprint of each AI query, surfacing the environmental costs of digital decision-making.
The Reasoning Panel reveals the sources, logic, and confidence behind each AI-generated output.
The Two-Page Export Tool helps users produce lightweight, shareable policy drafts based on real tone and context preferences.
Finally, the Environment Scenario Tool invites users to explore decisions from non-human perspectives, shifting the frame from anthropocentric policy to planetary thinking.


My Role

I led UX strategy, system logic, and interface content. I designed the flows for key features, structured the platform’s architecture, and authored the tone and interaction language. I also synthesized stakeholder insights into practical design decisions—especially where ethics and usability had to meet.

The top half of a screen of a mobile device
The top half of a screen of a mobile device
The top half of a screen of a mobile device

Outcome

ClimateGPT functions as both prototype and provocation. It demonstrates how speculative UX can be grounded in system logic and ethical foresight. The project was well received in critique and highlighted for its clear, actionable approach to responsible AI in public systems.


My Understanding

This project reinforced that ethical design starts with structure, not surface. I learned how to translate abstract concerns—like AI bias or ecological care—into concrete features and user flows. ClimateGPT challenged me to stay critical while designing for real people in complex systems.

Next Project

Next Project