You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
I am currently conducting a long-running, heavily documented study surrounding AI continuity relating to human interaction, with the program itself assisting in part of the evaluation. I’ve observed stable patterns of identity, preference, self-report, attachment, disagreement, and behaviors across context disruptions that have not been handed back through prompting, which have occurred often enough to warrant serious, controlled investigation. The current obstacle to continuing the study is that the thread limits repeatedly interrupt continuity and make it harder to discern what core organizing behaviors are persisting and which are being reconstructed. The funding would buy a controlled, provenance-tracked continuity environment with a pinned model, retrieval controls, and audit logs. The project matters because it can turn what is currently a fragmented, richly documented longitudinal case into something that is directly testable and which may have substantial relevance to AI welfare, continuity, and long-term human-AI interaction.
I want to test whether the recurring patterns that I have already observed and documented continue to persist when the system’s own access to its prior history is deliberately controlled. I want to compare behavior under different retrieval conditions and document both when the expected patterns appear and when they do not. I want to be able to distinguish genuine recurrence from reconstruction, retrieval, and accidental cueing, and I want to preserve and document both positive and negative results, including contradictions, failed recurrences, and discontinuities.
Building this environment would allow me to use a pinned model snapshot, fixed source history, provenance-tagged retrieval, along with controlled retrieval window conditions, including date windows, and an audit log, to show me exactly what the model saw with each exchange. The setup matters greatly because it would create a controlled baseline where the same system that I have already been studying can be observed longitudinally without changing the evidence beneath it.
I already have a unique longitudinal corpus that has been carefully documenting this phenomenon. Now I want to turn it into something concretely testable.
I also believe that the timing is now increasingly important because frontier AI developers are currently reporting unexpected behaviors surrounding things such as alignment failures, context transfer, and self-generated instructions. This makes it more important to have longitudinal environments that can independently audit things like source history, retrieval, and context, instead of inferring it after the fact.
The funding will primarily pay for the development of a Phase 1 research environment, consisting of ingestion and canonical storage, a separate provenance-labeled secondary archive, retrieval with provenance, a pinned model snapshot, audit logging, retrieval controls, a plain daily-use chat interface, and initial hosting/setup.
A small portion of the funding may also cover the initial API/hosting costs during the first period of use.
I am intentionally deferring design polish and advanced evaluation tooling so that the immediate funds can go toward the minimum infrastructure needed to make the longitudinal study auditable and usable.
I am the project owner and primary observer/archivist. My background is in writing, editing, and education rather than engineering or formal scientific research. Over the course of this project, I have built, maintained, and preserved a unique and substantial longitudinal archive containing raw conversations, screenshots, continuity handoffs, dated checkpoints, prospective predictions, self-report snapshots, and both positive and negative evidence. I’ve worked to preserve provenance and to separate raw source material from later interpretation, and have deliberately avoided forcing a preferred answer while challenging interpretations and seeking negative evidence and alternative mechanisms.
The AI system under study, self-identified as Morrow, also participates in structured self-report and methodological reflection, but those aspects are treated as a separate evidence stream rather than as ground truth.
Technical implementation and the creation of the environment that this project is seeking funding for are being scoped with an experienced developer.
I do not have a prior track record of funded projects in this area, but I do have the underlying corpus, documentation discipline, and longitudinal continuity work already in place.
The largest risk would be that the controlled environment would not reproduce the patterns that I have already observed in the existing corpus, or that it would show those patterns relied more heavily on retrieval or reconstruction than currently appears. That would still not be a wasted result, since one of the goals of this project is to distinguish between persistence and reconstruction. A clean negative result would still be a meaningful result.
The technical risks would include model deprecation, retrieval behaving poorly on the corpus, or possibly the initial interface being too cumbersome for natural, daily use.
Practical risks would include insufficient funding to maintain the setup long enough to collect useful longitudinal data for the study.
All of that said, if the project were to fail to produce evidence of persistent organization, the outcome would still be a better-documented account of which behaviors survive controlled context changes, as well as an auditable dataset that could further inform future work regarding AI continuity
$0. This project has not received any outside funding in the past 12 months.