CEPID - From Personal AI to a Collective Future - How Can We Study What We Are Proposing?

From Hypothesis to Question, from Question to Experiment

Throughout this series, we have built hypotheses.

We proposed a Personal AI capable of returning to the Body-Territory a greater capacity to interpret its own data.

We asked how millions of Selves could form Jiwasa without once again surrendering their singularities to the collective.

We arrived at the possibility of belonging without remaining in the same position and of transcending Being: not remaining imprisoned by what our data, our history, or some classification system said we were.

Then we also brought the Biome into the decision.

But a conceptually interesting architecture is not yet science.

Now we need a different movement:

Hypothesis → question → experiment → science.

We do not intend to present a finished experimental protocol here. It is not our role to define sample sizes, controls, statistical methods, or which variables would constitute sufficient evidence.

We want to do something that comes before that:

formulate questions clearly enough that researchers can criticize them, transform them, and design better experiments.

Perhaps this is one of the possible roles of a future CEPID from Personal AI to a Collective Future.

First: the Equipment Does Not Measure Jiwasa

A NIRS device does not measure belonging.

EEG does not directly measure consciousness.

ECG does not measure collectivity.

SpO₂ does not measure transcendence.

These devices record specific physiological phenomena.

Meaning appears later — when we relate those signals to behavior, task structure, environment, reported experience, and the hypotheses used to interpret them.

This caution is particularly important in hyperscanning.

A review by Hakim and colleagues examined 215 hyperscanning studies published between 2000 and 2022 and identified 27 different methods used to quantify inter-brain coupling. In other words, even what we call “synchronization” depends on methodological choices regarding how it is calculated. (sciencedirect.com)

At the same time, there is evidence that cooperative tasks may be associated with increased inter-brain synchronization. A meta-analysis involving 13 studies and 890 participants found significant synchronization in frontal and temporoparietal regions during cooperation. (eneuro.org)

So synchronization exists as an observable phenomenon.

But:

synchronization is not automatically Jiwasa.

A Simple Experiment

Imagine three participants.

At first, we do not need to separate them by religion, profession, culture, or political position.

We simply present a common problem:

A community is facing a progressive reduction in water availability and has limited resources to respond.

Three possible actions are offered:

A — expand water storage systems;

B — restore natural areas related to water retention and circulation;

C — temporarily restrict certain high-consumption activities.

None of the alternatives is presented as the correct answer.

First, each participant responds individually.

We have:

A | B | C

Then the three participants begin to talk freely while being observed with NIRS hyperscanning.

We can ask:

When do their positions begin to move closer?

When do they diverge?

Does anyone change their mind?

Does a minority position later influence the solution?

Can the participants reach a decision without all of them thinking in the same way?

Perhaps we observe:

[A B] | C

and later:

[A B C]

But another possibility may emerge:

[A B] ↔ C

A and B remain closer to one another, while C maintains a distinct interpretation, and yet all three are still able to construct a common response.

It is precisely this second possibility that interests us.

Harmony Does Not Need to Mean Homogeneity

We may be too accustomed to searching for convergence.

But perhaps a functional collective is not one in which everyone ends up occupying the same position.

Perhaps it is one capable of producing:

coordination without uniformity;

convergence without identity;

divergence without rupture;

belonging without permanence.

We can provisionally call this possibility:

Harmony Without Homogenization

If we expand the experiment to six, eight, or thirty people, several groups may emerge.

For example:

[A B C] | [D E] | [F G H]

Then:

[A B] | [C D E] | [F G H]

And later:

[A B D] ↔ [C E] ↔ [F G H]

Clusters may form, split, grow, shrink, or exchange participants throughout the task.

The interesting question is no longer only who synchronized with whom.

We begin to ask:

Who moved between clusters?

Who functioned as a bridge?

Which difference persisted?

Did an initially peripheral position later reorganize the collective?

And above all:

Did the collective need to eliminate its differences in order to find a solution?

This Already Happens, in Another Form, in the Digital World

This question has an important contrast.

Digital platforms already process behavioral data at enormous scale to personalize what different people receive.

Meta explains that its AI systems use multiple signals and predictions to determine which content on Facebook and Instagram may be most relevant to each person. (about.fb.com)

TikTok likewise describes its recommendation systems as personalized on the basis of preferences inferred from user interactions. (support.tiktok.com)

In 2026, Meta described systems capable of processing sequences of billions of interactions and building representations of interests and intentions to improve recommendation and ranking systems. (engineering.fb.com)

Technically, these processes do not necessarily mean “clusters” in the strict statistical sense.

But conceptually, something powerful is already happening:

distributed signals → pattern recognition → differentiation among users → differentiated distribution of information.

Today, this capacity is used, among other things, to personalize feeds, advertisements, and recommendations.

Our question is different:

what if part of this capacity to recognize differences were primarily under the control of the Body-Territory itself?

From the Profile That Fixes Us to the Position That Can Change

A digital network tries to predict:

“what is this person likely to be interested in?”

A Personal AI could add:

“does this still represent you?”

That small difference changes the architecture profoundly.

Imagine that, during our experiment, the model identifies greater similarity between two participants.

The AI could present:

“During this stage, your dynamics showed greater proximity to participant B.”

But the participant could respond:

“Yes, that represents me.”

Or:

“No. We only agreed on one part of the problem.”

Or:

“At that moment, I had already changed my position.”

Then we would have:

signal → analysis → interpretation → feedback → contestation or recognition.

The participant ceases to be merely a source of data.

They begin to participate in the meaning produced from their own data.

This is where transcending Being begins to acquire a possible technological dimension.

No classification needs to become permanent.

Different Cultures Can Expand the Questions

The same experiment could later include participants coming from different histories, religions, urban communities, and Indigenous peoples.

But not in order to ask:

“how does an Indigenous person think?”

or:

“which belief produces greater synchronization?”

That would reduce people and cultures to the categories with which they entered the laboratory.

The more interesting question would be:

how do different ways of interpreting territory, water, community, and future participate in the construction of a common decision?

In early BrainLatam documents exploring this possibility, tasks involving water collection, reforestation, regeneration, and cooperation were already proposed, together with reflection by participants after the activity.

Those documents also proposed involving Indigenous, religious, and community leaders and adopting co-creation processes.

Today we can formulate this more precisely:

other epistemologies do not need only to enter the experiment; they can modify the experimental question itself.

From Clusters to a Proto-Idea of the State

There is another provocative possibility.

If several clusters can remain partially independent and still coordinate when a problem emerges, this dynamic may offer a proto-idea for thinking about institutions.

Health, environment, civil defense, education, infrastructure, and the economy do not need to detect the same things.

We can use, carefully, an analogy with the immune system.

Its cells do not remain permanently gathered, waiting for a threat. Many circulate, recognize signals locally, communicate through chemical mediators, and recruit broader responses when necessary.

A network of Personal AIs could allow us to imagine something similarly distributed:

Body-Territory
→ Personal AI
→ authorized signal
→ emerging collective pattern
→ competent institution.

An AI would not need to surrender the entire Body-Territory to the State.

Only previously authorized, anonymized, or aggregated signals would participate in specific institutional databases.

Environmental variations could matter to one institution.

Health signals to another.

Infrastructure problems to another.

And unusual patterns in information circulation could raise questions related to national sovereignty.

In the case of coordinated mass disinformation campaigns, for example, the architecture would not need to automatically decide what is true or false.

It could perform a prior function:

detect that an unusual pattern is emerging.

And Then the Most Difficult Problem Appears

Suppose the experiment works.

We find clusters.

We find synchronization.

We find divergence.

We find people changing positions.

We find groups capable of reaching common solutions without becoming homogeneous.

And perhaps, in the future, we learn to detect collective phenomena earlier and earlier.

We will have produced data.

But data still do not carry their meaning by themselves.

An algorithm may identify three clusters.

The participants may respond:

“We were not three groups. We were simply using three strategies.”

An institution may detect an unusual informational pattern.

But who will decide whether that represents manipulation, legitimate cultural transformation, conflict, innovation, or simply difference?

This is where our experiment ends and the next problem begins.

Perhaps we can learn to observe:

Self → difference → cluster → Jiwasa → decision.

But measuring does not mean understanding definitively.

And understanding does not automatically grant the right to decide.

The next question, therefore, will be:

If we can measure differences among Body-Territories and observe emergent properties of a collective, who gains the right to say what those signals mean?

That will be Blog 12.

References

HAKIM, U. et al. Quantification of inter-brain coupling: A review of current methods used in haemodynamic and electrophysiological hyperscanning studies. NeuroImage, v. 280, 120354, 2023. DOI: 10.1016/j.neuroimage.2023.120354.

CZESZUMSKI, A.; LIANG, S. H.-Y.; DIKKER, S.; KÖNIG, P.; LEE, C.-P.; KOOLE, S. L.; KELSEN, B. Cooperative Behavior Evokes Interbrain Synchrony in the Prefrontal and Temporoparietal Cortex: A Systematic Review and Meta-Analysis of fNIRS Hyperscanning Studies. eNeuro, v. 9, n. 2, 2022. DOI: 10.1523/ENEURO.0268-21.2022.

ZHAO, Q.; ZHAO, W.; LU, C.; DU, H.; CHI, P. Interpersonal neural synchronization during social interactions in close relationships: A systematic review and meta-analysis of fNIRS hyperscanning studies. Neuroscience & Biobehavioral Reviews, v. 158, 105565, 2024. DOI: 10.1016/j.neubiorev.2024.105565.

META. How AI Influences What You See on Facebook and Instagram. Meta, 2023.

META ENGINEERING. From User Sequences to Scaling Laws: A Multi-Stage Architecture for Meta's Ads Ranking. Meta, 2026.

TIKTOK. How TikTok Recommends Content. Official documentation on recommendation and personalization systems based on preferences and interactions.

BRAINLATAM. Tareas Cooperativas para Mitigar Emergencias Climáticas y Activar Creencias Diferenciales — NIRS Hyperscanner. Working document, 2024.

BRAINLATAM. Estrategia para Involucrar a Políticos y a la Sociedad en el Proyecto sobre Creencias y Emergencias Climáticas — NIRS Hyperscanner. Working document, 2024.