ECO · PsycheAI

PsycheAI · Scientific and Institutional Partnership Proposal

PsycheAI and ECO

Computational research on observable regularities of human discourse.

Renan Oliveira Santos — Psychologist and developer responsible for PsycheAI · August 3, 2026 · Version 1.0

Cover letter

To the interested institution,

My name is Renan Oliveira Santos, I am a psychologist with more than fifteen years of professional experience, and a developer of digital solutions for research, data analysis and artificial intelligence.

I am pleased to present, through this document, PsycheAI, an independent research and technology-development project dedicated to the computational investigation of observable regularities in human discourse. Its first research module is called ECO — listening intelligence.

The project starts from a central question:

Is it possible to computationally represent observable regularities of human discourse in a traceable and reproducible way, without attributing clinical meaning, without automating interpretations and without replacing the subject's own confirmation?

PsycheAI is currently consolidating its scientific architecture and preparing a pilot study. At this stage, I am seeking dialogue with institutions, research groups and researchers to present the proposal, receive scientific appraisal and assess the possibility of institutional collaboration, methodological guidance or connection with researchers interested in the interface between psychology, language, computation and artificial intelligence.

This presentation does not presuppose adherence, endorsement or prior responsibility on the part of the partner institution. Its purpose is to open a serious, transparent and technically delimited dialogue.

Sincerely,

Renan Oliveira Santos

Psychologist and lead developer of PsycheAI

1. Executive summary

From the Project for a Scientific Psychology to discourse as a computational object

PsycheAI is a computing research project inspired by the Project for a Scientific Psychology, written by Sigmund Freud in 1895, and by the structural formalization of discourse developed by Jacques Lacan, particularly in Seminar 17.

Its computational object is not the mind, the unconscious or the individual's personality. Its object is discourse as a set of observable occurrences, recorded over time and open to computational organization.

The working hypothesis holds that human discourse presents regularities — such as repetitions, recurrences, associations, interruptions, shifts and changes of position — that can be recorded and represented by a computational system in a verifiable way.

The first module, called ECO, is meant to receive an authorized discursive production, organize the observable elements and return to the participant occurrences found in their own material. ECO must not state the meaning of a word, conclude that a given element constitutes a signifier, produce a psychological diagnosis, or replace professional listening. Confirming or rejecting the relevance of an occurrence remains with the subject.

The first pilot study should work with a small group, initially estimated between 3 and 20 participants, after ethical approval. Its goal will be to validate the collection method, data security, processing traceability and the system's ability to recognize and return regularities without exceeding the established scientific limits.

PsycheAI seeks an academic approximation that can contribute to methodological refinement, critical evaluation of the hypothesis and the future ethical submission of the research.

2. Origin and motivation of the project

From Freud to Lacan, and from theory to computation

Throughout the history of psychology, different authors have sought to describe psychic processes through models capable of organizing observations, relations and transformations.

In Project for a Scientific Psychology, Freud made an initial attempt to represent psychic processes through quantities, barriers, facilitations, memory and discharge. The text does not offer a contemporary computational model, but it raises a question that is still current: how to build a rigorous representation of processes that are not directly observable, without confusing the model with the phenomenon?

Later, Lacan shifted attention to the structure of language and to the positions established within discourse. His formulation allows repetition and discursive relations to be thought of not merely as isolated contents, but as occurrences within a structure.

PsycheAI is born from the approximation between these historical questions and the contemporary possibilities of computation. The project does not intend to turn psychoanalytic theory into a set of automatic diagnoses. Its purpose is more limited: to investigate whether certain observable aspects of discourse can be recorded, related and returned with methodological transparency.

3. Research problem

"What does the speech mean?" is not the first question

Contemporary artificial-intelligence systems can transcribe, classify, summarize and generate language. However, many of these operations are carried out by complex models whose answers can be difficult to explain and reproduce.

In contexts related to human experience, this problem becomes particularly sensitive. A system can produce a linguistically convincing interpretation without having sufficient grounding to sustain it.

PsycheAI proposes a different direction. Instead of first asking "what does the speech mean?", the project asks:

The scientific problem consists of building a computational representation that stays faithful to the observable data and makes its limits explicit.

4. Scientific hypothesis

Four methodological commitments

Human discourse presents structural regularities — such as repetitions, recurrences, associations, interruptions, shifts and changes of position — open to observable, traceable and reproducible computational representation, without automatic attribution of clinical meaning, without autonomous identification of signifiers, and without replacing the subject's own confirmation.

This hypothesis contains four commitments:

  1. Observability: every result must correspond to an identifiable occurrence in the collected material.
  2. Traceability: it must be possible to locate the origin of each result in the recording, transcription or discursive sequence.
  3. Reproducibility: versions, parameters and processing conditions must be recorded to allow re-execution and comparison.
  4. Interpretive limit: the system must not turn a statistical or linguistic regularity into a clinical truth about the participant.

5. What ECO is

Returning what appeared, without determining what it means

ECO is PsycheAI's first research module. Its name expresses its function: to return something that appeared in the discourse itself, without taking the place of the one who determines its meaning.

In a pilot version, ECO will be able to perform operations such as:

ECO will not be presented as a therapist, an analyst or a diagnostic instrument. Its initial purpose will be exclusively investigative.

6. What PsycheAI does not intend to do

Explicit limits, not merely communication caveats

To preserve scientific and ethical clarity, the project establishes explicit limits. PsycheAI does not intend to:

These limits are not mere communication caveats. They constitute requirements of architecture, methodology and validation.

7. Pilot study objectives

7.1 General objective

To assess the technical, methodological and ethical feasibility of representing and returning observable regularities of human discourse through ECO.

7.2 Specific objectives

8. Preliminary design of the pilot study

A starting point, not a definitive design

The definitive design will be developed with methodological support and submitted for ethical review before recruitment. As a starting point, it is proposed:

ParticipantsBetween 3 and 20 adults
NatureExploratory and methodological pilot study
ParticipationVoluntary, subject to free and informed consent
CollectionOne or more sessions of discursive production guided by protocol
RecordAudio, transcript and timestamps
ProcessingDeterministic methods and documented AI resources
ResultReport of observable occurrences with supporting evidence
ValidationHuman review, participant feedback and error analysis
PurposeResearch, without diagnosis or clinical care

Sample size, inclusion and exclusion criteria, session duration and collection instruments will still need formal methodological justification.

9. Proposed computational method

A chain of traceability, from audio to feedback

Processing must maintain a chain of traceability:

  1. authorized recording;
  2. secure storage;
  3. transcription linked to the audio timeline;
  4. controlled normalization of the text;
  5. recording of the algorithm version;
  6. execution of the rules or models;
  7. production of candidate occurrences;
  8. review and quality control;
  9. feedback accompanied by evidence;
  10. recording of the participant's response, when foreseen in the protocol.

The project must distinguish between two types of processing:

Deterministic processing — operations based on explicit rules, such as counting, literal repetition, distance between occurrences, frequency and temporal location. These operations favor explainability and reproduction.

AI-assisted processing — operations that use statistical or language models, such as voice transcription and identification of linguistic variations. These results must receive treatment compatible with their uncertainty, maintaining human review and a record of the execution conditions.

Artificial intelligence must not be used as an interpretive authority.

10. Ethics, privacy and protection of participants

Protection by design, not as an afterthought

The research will involve potentially sensitive data, including voice, transcription and personal accounts. Ethical and informational protection must therefore be part of the design from the outset. The protocol must address:

The research must be submitted for ethical review before any recruitment or collection. Planning takes into account Brazilian Law No. 14,874/2024, Decree No. 12,651/2025, National Health Council Resolution No. 510/2016, the applicable complementary norms, and the General Data Protection Law (LGPD).

11. Current state of the project

Scientific, technical, institutional and ethical preparation phase

PsycheAI already has a conceptual and documentary foundation under development, including:

The project has not yet started collection for the pilot study.

12. Why seek an academic partnership

Critical dialogue, not just outreach

The search for an academic partnership stems from the interest in presenting an interdisciplinary project that brings together psychology, the structure of language, computation and scientific investigation.

PsycheAI is not seeking mere publicity. It seeks critical dialogue. An academic collaboration could help turn a promising hypothesis into a methodologically consistent protocol, submitted for ethical review and capable of producing examinable results.

The partner institution may, should it recognize relevance and institutional availability, contribute in one or more of the following ways:

This document does not presuppose that all these forms of collaboration are possible. The intention is to identify, together, which mode of approximation would be pertinent.

13. Concrete proposal for approximation

An introductory meeting, between 45 and 60 minutes

As a first step, an introductory and preliminary-assessment meeting is proposed, with an estimated duration of 45 to 60 minutes. In the meeting, the following may be presented:

  1. the scientific question and the hypothesis;
  2. the delimitation of the computational object;
  3. a conceptual demonstration of ECO;
  4. the preliminary design of the pilot study;
  5. the ethical and technical limits;
  6. the current state of the documentation and software;
  7. the possibilities for collaboration.

After this meeting, should mutual interest exist, a specific work plan may be drawn up, without any automatic commitment between the parties.

14. Possible contributions from each party

PsycheAI's current contributions

Contributions that could be discussed with the partner institution

Authorship, intellectual property, data use, responsibilities, dissemination of results and any financial resources must be defined in advance through appropriate instruments, should the collaboration move forward.

15. Expected results

Success measured by verifiability, not by interpretation

In the pilot phase, the project expects to produce:

The success of the pilot will not be measured by the ability to produce interpretations about the participants, but by the quality, transparency and verifiability of the computational representation.

16. Acknowledged limitations

Part of the scientific problem, not something to hide

PsycheAI acknowledges from the outset that:

These limitations will be treated as part of the scientific problem, and not concealed when presenting results.

17. Proposed stages

Five phases, from institutional approximation to scientific communication

Phase 1 — Institutional approximation

  • sending this proposal
  • introductory meeting
  • identification of an academic contact
  • preliminary relevance assessment

Phase 2 — Scientific consolidation

  • review of the hypothesis
  • methodological refinement
  • definition of instruments
  • definition of the sample and criteria
  • review of risks and protective measures

Phase 3 — Ethical preparation

  • completion of the protocol
  • review of the consent form
  • completion of the data-management plan
  • definition of the proposing institution, when applicable
  • submission for ethical review

Phase 4 — Pilot implementation

  • infrastructure preparation
  • team training
  • recruitment after approval
  • running of the sessions
  • processing and quality control

Phase 5 — Analysis and scientific communication

  • analysis of results
  • documentation of limitations
  • drafting of the report
  • assessment of scientific publication
  • decision on continuity and expansion

18. Closing remarks

Without erasing the difference between data, interpretation and experience

PsycheAI is born from an old question posed before contemporary technological resources. Its proposal is not to reduce the subject to a set of patterns, but to investigate how far computation can represent occurrences of discourse without erasing the difference between data, interpretation and subjective experience.

ECO intends to function as a research instrument that makes certain regularities visible and returns their evidence. The system must say what it found and where it found it, but must not, on its own, determine what this means for the subject.

It is precisely because it recognizes the scientific, ethical and human complexity of this work that PsycheAI seeks academic dialogue before starting its pilot study.

I invite institutions, research groups and researchers to learn about the proposal and to assess the possibility of a first meeting. A critical appraisal, even if it does not immediately result in a formal partnership, will already represent an important contribution to the project's development.

19. Contact

Renan Oliveira Santos

Psychologist and developer responsible for PsycheAI — Divinópolis, Minas Gerais, Brazil.

Project website: ecopsy.online

Instagram: @eco_psycheai

Institutional contact: contato@ecopsy.online

Initial references

Institutional note: this proposal presents an independent project in a preparatory phase. The document does not imply any bond, approval, partnership, endorsement or responsibility on the part of any institution consulted until there is a formal statement from the parties.

Want to learn more about the research?

Go back to ECO's main page or read the full scientific foundation.

← Back to ECO Research Foundation