Design Note (Provisional)
This document explores open questions about privacy, observation, inference, care, and AI within the Pancakes ecosystem.
It is intentionally exploratory.
It does not establish policy or architectural requirements.
Instead, it identifies a class of questions that increasingly appear as Pancakes evolves from a collection of applications into a long-lived life computing platform.
Many of these questions cannot be answered using traditional privacy models alone.
Traditional privacy discussions focus on questions such as:
These questions remain essential.
However, Pancakes introduces another class of questions.
Instead of asking:
Who may access information?
we increasingly find ourselves asking:
What kinds of knowledge should humane software help create?
This distinction becomes important because Pancakes is designed to:
Even when every individual observation is ethical, the resulting body of knowledge may raise entirely new ethical questions.
This document explores those questions.
The phenomenology underlying these questions has already been explored from an Informational Experiential Realism perspective in Informational Personal Space, Opacity, and Future-Space Restriction.
That document argues that the discomfort associated with “being modeled” is not primarily about privacy violations.
Instead, it arises when interpretations become durable social structures that influence another person’s future despite never overcoming the inherent opacity of human experience. This document assumes that analysis.
Its purpose is different.
Rather than explaining the phenomenon, it asks:
If informational personal space exists, what responsibilities does that create for software?
Abed begins keeping a personal journal in Pancakes.
His goal is not to study anyone else.
He wants to become better at understanding social situations and his own relationships.
Each evening he spends a few minutes reflecting on the day.
His entries contain observations such as:
The notebook is fundamentally about Abed’s own life.
He is not tracking Annie.
He never accesses Annie’s notebook.
He never reads her private records.
He never bypasses permissions.
Years later he asks his AI assistant:
“Can you help me understand my relationships?”
The assistant reviews thousands of journal entries.
Among many observations it notices that Annie’s mood, participation, and energy appear to follow a recurring monthly rhythm.
Eventually it becomes capable of predicting that rhythm with considerable accuracy.
Neither Abed nor the AI ever accessed Annie’s private information.
Everything was inferred from years of ordinary public observations recorded for an entirely different purpose.
This raises an uncomfortable question.
Should Pancakes help create that knowledge?
Observation is an ordinary part of human life.
We constantly notice:
Observation is not surveillance.
Observation is one of the foundations of empathy, friendship, parenting, teaching, and caregiving.
Questions:
Writing observations changes their character.
Human memory is:
Written records become:
Questions:
Many important conclusions are never directly observed.
They emerge through inference.
Examples:
Observed:
Inferred:
Often the inferred conclusion is substantially more sensitive than any individual observation.
Questions:
Inference naturally enables prediction.
Examples:
Prediction changes relationships.
Prediction changes power.
Prediction changes responsibility.
Questions:
Humans naturally develop models of other people.
AI differs primarily in scale.
Humans:
AI may:
The concern may therefore be less about artificial intelligence itself than about persistent, large-scale modeling.
Many observations do not belong exclusively to one individual.
Examples:
These observations are inherently relational.
Questions:
Observation can arise from care.
Examples include:
The same observational techniques can also support:
Technology alone does not distinguish the two.
Questions:
Future Pancakes systems may support:
These domains necessarily involve observing other people.
Should systems distinguish between:
Should these contexts carry different ethical expectations?
People possess physical personal space.
Standing extremely close to another person is often uncomfortable even though it causes no direct physical injury.
Perhaps people also possess informational personal space.
Possible hypothesis:
Every person deserves a zone of interpretive openness around themselves.
Questions:
Perhaps privacy is only one category of boundary.
Perhaps there are also boundaries around knowledge itself.
Possible questions include:
What may be:
These may represent distinct ethical categories rather than different stages of the same process.
Throughout Pancakes, symbolic projection is generally preferred over explicit disclosure.
Instead of:
Fertility window
an ambient client might display:
A crimson moon.
Instead of:
Burnout score
it might display:
A neglected garden.
Instead of:
Elevated stress
it might display:
Storm clouds gathering beyond the hills.
One possible reason is that symbolism preserves ambiguity.
It communicates emotional reality without collapsing lived experience into explicit diagnostic statements.
Perhaps symbolic projection is not merely an aesthetic decision.
Perhaps it is also a privacy technology.
Should AI:
If so:
Should an AI ever know something that no human participant has consciously recognized?
Life journals inevitably contain observations about other people.
Examples include:
Should AI treat bystanders differently from users?
Do bystanders possess rights even when they never use Pancakes themselves?
Entire communities may become observable over time.
Nodes may reveal:
Such knowledge may support stewardship.
It may also enable surveillance.
How should communities preserve dignity while still understanding themselves?
These are hypotheses rather than conclusions.
Some questions remain unresolved.
Pancakes began by asking questions about privacy.
Increasingly, it appears to be asking questions about knowledge itself.
Traditional privacy asks:
Who may access information?
Humane life computing may also need to ask:
What kinds of knowledge should systems help create?
The goal may not simply be protecting information.
It may be protecting each person’s ability to remain partially unknowable.
Not because opacity is a flaw.
But because opacity is one of the conditions that makes human freedom, growth, reinterpretation, and genuine relationships possible.
The questions explored in this document are often associated with artificial intelligence, surveillance, or large-scale data collection.
However, the underlying phenomenon appears in ordinary human life.
A customer visits the same neighborhood convenience store almost every day.
Each visit is unremarkable.
He purchases the same bag of chips.
The owners gradually recognize him.
Eventually he realizes they refer to him among themselves as:
“The chip guy.”
No private records were accessed.
No surveillance occurred.
No confidential information was disclosed.
The owners simply remembered a repeated pattern of public observations.
Yet the customer experiences discomfort.
Traditional privacy frameworks struggle to explain this reaction.
Nothing secret has been revealed.
Nothing has been stolen.
Nothing has been exposed.
The discomfort instead arises from a different source.
The repeated observations have become a durable model.
The customer is no longer encountered as an open-ended person.
He is encountered through an established interpretation.
The issue is not that the owners know he buys chips.
The issue is that they may now believe they know what kind of person he is.
Repeated observation naturally invites inference.
Observed:
Possible inferences:
None of these conclusions necessarily follows from the observation.
They remain interpretations.
Nevertheless, repeated observation tends to increase confidence in those interpretations.
Importantly, the customer’s lived experience remains opaque.
Perhaps the chips are shared with coworkers.
Perhaps they replace lunch.
Perhaps they are simply an inexpensive daily pleasure.
The observers cannot know.
Yet their growing familiarity may create the feeling that they do.
This illustrates an important distinction.
Observation is not inherently problematic.
Neighborhood businesses often remember regular customers.
Remembering names, favorite orders, or familiar routines can be an expression of hospitality, belonging, and care.
The discomfort emerges when accumulated observations become perceived knowledge of another person’s character.
Repeated exposure produces epistemic confidence that may exceed what the observations actually justify.
A second example illustrates a related but distinct phenomenon.
While packing a backpack, someone remarks:
“You’re going to the library.”
The prediction is correct.
By itself, this is unremarkable.
The observer has noticed a recurring pattern and successfully anticipated another occurrence.
But the conversation continues.
The observer says:
“I know you better than you know yourself.”
This statement transforms the interaction.
The issue is no longer prediction.
It is authority.
The speaker claims that their accumulated observations entitle them to a more authoritative understanding of the other person than the person’s own first-person experience.
This claim cannot be justified by repeated observation alone.
Human behavior may become increasingly predictable while human experience remains fundamentally opaque.
The observer may correctly anticipate where someone is going while still misunderstanding why they are going there, what they are thinking, or how they understand themselves.
Prediction does not overcome opacity.
It merely increases confidence.
This suggests another ethical distinction.
There is a profound difference between saying:
“I think you’re going to the library.”
and saying:
“I know you better than you know yourself.”
The first is a prediction.
The second is a claim of epistemic superiority.
These examples suggest that humane software should distinguish between increasing predictive capability and increasing epistemic authority.
Accumulating observations may improve a system’s ability to recognize patterns.
It does not necessarily justify stronger claims about another person’s motives, identity, values, or lived experience.
This document therefore raises a possible question for future work.
Do people possess not only informational personal space, but also a form of epistemic primacy?
If so, it would not imply that individuals are always correct about themselves.
Rather, it would recognize that each person’s first-person perspective occupies a privileged role that should not be casually displaced by external models, whether human or artificial.
AI systems may help people discover patterns, identify possibilities, or reflect upon their lives.
They should be cautious about presenting those models as more authoritative than the person’s own ongoing process of self-understanding.
The concept of epistemic primacy remains provisional.
It may ultimately prove to be an important extension of informational personal space and a useful design principle for humane life computing.
These everyday examples illustrate that the ethical questions explored in this document are not unique to artificial intelligence.
They arise whenever repeated observation becomes durable interpretation.
Artificial intelligence magnifies the phenomenon through persistence, scale, and statistical sophistication.
The underlying ethical question, however, remains the same.
How much knowledge should repeated observation be permitted to create?
Humane systems should recognize that repeated observation does not eliminate opacity.
Accumulated evidence may improve prediction while still failing to justify confident conclusions about another person’s motives, values, circumstances, or character.
Respect for informational personal space therefore requires more than protecting access to information.
It also requires epistemic humility regarding what observations can legitimately support and respect for each person’s ongoing role as the primary interpreter of their own life.