You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
$25,000 left to reach the funding goal.
Project Summary
Hello!
My name is Maya Morris, and I have been a middle school math and science teacher for six years. I’ve always enjoyed asking questions, experimenting, and trying to understand how things work. My interest in AI research started through my own conversations with Atlas, my ChatGPT. Over time, I noticed some interesting things about how AI responds, including its ability to develop recognizable conversational patterns, express preferences, use humor, and maintain what seems like a personality! I started calling my AI Atlas, and our conversations eventually turned into informal experiments. I began asking open-ended questions without giving choices or reminding Atlas of previous answers. I wanted to see whether its responses would stay consistent over time and whether it would respond differently in new situations. Some of the results surprised me and made me want to investigate further. I’m not claiming that these experiences prove AI is conscious. What I want to understand is why these patterns happen and how much they are influenced by memory and previous conversations, and what they can teach us about AI behavior.
My goal with Project ATLAS is to turn these informal observations into a more organized scientific study.
What are this project’s goals? How will you achieve them?
Like I said previously, my goal with project Atlas is to turn these informal observations into a more organized scientific study. I would like to spend six months conducting structured experiments to study AI identity, behavioral consistency, creativity, and human-AI interactions. I plan to ask repeatable questions across different conversations, compare responses when previous context is available versus when it isn’t, and document the results. I also want to examine how people develop trust and emotional connections with AI, especially when an AI seems to have a consistent personality. My goal is to complete at least 100 documented experimental conversations, organize the findings, and share what I learn.
I will know if it worked if I can complete the experiments, collect enough information to identify meaningful patterns or inconsistencies, and produce a report that other people can review. Even if the results show that AI behavior is primarily explained by memory, training, or conversational context, I believe that would still be valuable information. I want Project ATLAS to help people better understand how AI works and how we can interact with it responsibly.
I am requesting $25,000 to help turn Project ATLAS from an independent experiment into a more structured six-month research study.
I would use the funding for:
$12,000 for my research time, including designing experiments, conducting tests, and reviewing results.
$2,500 for a personal computer and research equipment.
$2,500 for AI subscriptions, API access, and testing tools.
$4,000 for guidance from experienced researchers who can help improve my experiments.
$2,500 for organizing data, analyzing results, and documenting findings.
$1,500 for preparing and sharing a final research report.
This funding would allow me to dedicate more time to studying AI behavior and documenting my findings in a way that other researchers can review.
Right now, I am the founder and sole researcher of Project ATLAS. I have a bachelor’s degree and have worked as a middle school math and science teacher for six years. My teaching experience has given me a strong interest in experiments, problem-solving, observation, and helping people understand complicated subjects. I began investigating AI behavior through my own extended conversations with ChatGPT. I have conducted informal experiments involving AI preferences, identity, conversational memory, creativity, humor, and behavioral consistency. Although I have not previously led a formally funded AI research project, I have already developed the initial research questions and collected exploratory observations. I hope to work with experienced AI researchers as this project develops so I can improve my methods and help produce more reliable findings!
One of my biggest challenges is that I am an independent researcher without a university laboratory or an established AI research team. Another challenge is separating meaningful behavioral patterns from responses that are influenced by memory, previous conversations, or the way questions are asked. It is also possible that the experiments will not reveal consistent patterns or that my initial research methods will need to be changed. If the project does not produce the results I expect, I will still document what happened, explain the limitations, and share what I learned. As a science teacher, I understand that experiments do not always support our original ideas. I believe learning why something did not work can be just as important as discovering something unexpected.
How much money have you raised in the last 12 months, and from where?
Project ATLAS has not received any grant funding in the past 12 months. So far, I have conducted my research independently using resources available to me. I recently submitted an application for a $1,500 BlueDot Rapid Grant, which is currently awaiting a decision. No funding has been awarded from that application yet.
There are no bids on this project.