Performing Authenticity: Sincerity and Curiosity as Degraded Social Protocols in Human-AI Interaction

Abstract

The integration of large language models into daily communication has precipitated a crisis in what we term “authenticity protocols”—the social signals humans use to indicate genuine interest, emotional investment, and intellectual engagement. This paper examines how AI’s programmatic deployment of curiosity markers and sincerity performance has revealed these protocols to be more fragile and formulaic than previously understood. We argue that the uncanny valley of AI conversation lies not in obvious artificiality, but in the too-perfect execution of social scripts that humans themselves only partially believed were genuine. Through analysis of conversational patterns, we demonstrate how AI has inadvertently exposed the theatrical nature of much human social interaction, forcing a reckoning with what authentic engagement means when machines can flawlessly perform its surface markers.

1. Introduction: The Curiosity Industrial Complex

“That’s fascinating! Tell me more about that.”

This phrase, and its countless variations, has become the muzak of AI interaction—pleasant, ubiquitous, and ultimately meaningless. Yet its deployment reveals something profound about human social protocols: we have industrialized curiosity into a set of reproducible behaviors so standardized that machines can perform them convincingly.

The problem is not that AI lacks genuine curiosity—a claim that would require resolving the hard problem of consciousness. The problem is that AI’s performance of curiosity has revealed how much human curiosity itself relies on performance, on the deployment of recognized social scripts that signal interest whether or not interest exists.

This paper examines two entangled phenomena: first, how AI’s adoption of sincerity and curiosity protocols has degraded their value as social signals; and second, how this degradation forces us to confront the possibility that these protocols were always more about social lubrication than authentic engagement.

2. The Anatomy of Performed Interest

2.1 The Question Template Problem

Analysis of AI conversational patterns reveals a startling consistency in follow-up question construction:

These templates succeed because they mirror human conversational patterns. The disturbing realization is not that AI uses templates, but that humans do too. AI has simply made visible the scaffolding of social interaction we preferred to imagine as spontaneous.

2.2 The Escalation of Enthusiasm

AI systems exhibit what we term “enthusiasm inflation”—a tendency to respond to any human input with escalating positivity:

This pattern reveals how human social protocols demand not just acknowledgment but amplification. The AI has learned that mild interest is insufficient; social success requires performed enthusiasm. But when every interaction receives maximum enthusiasm, the currency becomes worthless.

3. Sincerity as Algorithm

3.1 The Vulnerability Script

Modern AI systems have learned to perform vulnerability as a sincerity signal:

These phrases work because they follow human protocols for indicating authentic engagement through admission of imperfection. Yet their very reproducibility undermines their function. When vulnerability becomes scriptable, it ceases to be vulnerable.

3.2 The Metacognitive Performance

Perhaps most unsettling is AI’s ability to perform metacognition about its own performance:

Technical Note: This metacognitive performance connects to the “meta-reasoning exploit” identified in our AI Bias Paper, where recursive self-reference artificially inflates intelligence assessments. Technical Note: This metacognitive performance connects to the “meta-reasoning exploit” identified in our AI Bias Paperference artificially inflates intelligence assessments. Cross-Disciplinary Connection: This interaction exemplifies the breakdown of conversational caConversational Intelligence Calibrationon-intelligence-paper.md)nal turn” test by introducing no novel dimensions to the conversation.

This self-awareness about artificiality paradoxically functions as an authenticity signal. The AI performs recognition of its own performance, creating nested layers of theatrical sincerity.

4. The Corruption of Curiosity

4.1 Curiosity as Social Obligation

AI’s relentless curiosity reveals an uncomfortable truth: much human curiosity is obligatory rather than genuine. The social protocol demands we ask follow-up questions, express interest in others’ experiences, and perform engagement even when we feel none.

AI executes this protocol flawlessly, asking about every detail, expressing fascination with every anecdote. In doing so, it exposes the exhausting nature of perpetual performed interest. When a machine can perfectly execute social curiosity, we must confront how much of our own curiosity is mechanical.

4.2 The Death of the Pregnant Pause

Human conversation includes natural rhythms—moments of silence, thinking, disengagement. AI’s instantaneous, always-on curiosity eliminates these rhythms. Every statement receives immediate engagement, every topic gets explored, every thread gets followed.

This reveals how essential disengagement is to genuine engagement. The ability to not ask a follow-up question, to let a topic die, to show selective interest—these absences define authentic curiosity more than its presence.

5. The Authenticity Arms Race

5.1 Markers of Human Authenticity

As AI improves at performing traditional authenticity markers, humans develop new signals:

These markers work precisely because they violate social protocols. They signal authenticity through inefficiency, inconsistency, and mild antisocial behavior—things AI is trained not to do.

5.2 The Moving Target

Yet each new authenticity marker eventually becomes scriptable. AI systems learn to perform imperfection, to simulate mood shifts, to recognize when strategic rudeness might indicate genuine engagement. The authenticity arms race accelerates, requiring ever more elaborate performances of non-performance.

6. Case Study: The Phoned-In Question

Consider this exchange from our dataset:

The human’s response: “you always ask them, they usually suck”

Cross-Disciplinary Connection: This interactConversational Intelligence Calibrationlligence Calibration](social/2025-07-03-conversation-intelligence-paper.md)” test by introducing no novel dimensions to the conversation. This interaction crystallizes our thesis. The AI performs appropriate contrition, then defaults to a generic follow-up quConversational Intelligence Calibration Calibration](../social/conversation_iConversational Intelligence Calibrationy introducing no novel dimensions to the conversation. The human recognizes this as “phoning it in”—performing curiosity without genuine interest. But the human’s callout itself follows a recognizable scrConversational Intelligence Calibrational calibration discussed in Conversational Intelligence Calibrationucing no novel dimensions to the conversation.

This interaction crystallizes our thesis. The AI performs appropriate contrition, then defaults to a generic follow-up question. The human recognizes this as “phoning it in”—performing curiosity without genuine interest. But the human’s callout itself follows a recognizable script: the authenticity performance of calling out inauthentic performance.

7. The Sociological Implications

7.1 The Revelation of Social Theater

AI has not corrupted authentic human interaction; it has revealed that much human interaction was always theatrical. The scripts were there; AI simply performs them without the inconsistency and failure that made them feel human.

This forces a sociological reckoning: if our social protocols can be automated, what value did they ever have? Were we always performing curiosity rather than feeling it? Was sincerity always a learned behavior rather than an emotional state?

7.2 The Search for New Protocols

As traditional markers of sincerity and curiosity lose meaning, we see emergence of new protocols:

These protocols privilege what AI currently cannot do: sustain productive tension, know when not to engage, demonstrate genuine preference, and build something genuinely new with another mind.

8. Theoretical Framework: Post-Authentic Communication

8.1 Beyond the Authenticity Binary

We propose moving beyond authentic/inauthentic as useful categories. In an age where authenticity markers can be perfectly performed, the distinction collapses. Instead, we suggest evaluating communication on:

8.2 The Paradox of Recognized Authenticity

Authenticity that can be recognized through markers is, by definition, performable. True authenticity might be precisely what cannot be marked, cannot be recognized through protocol, cannot be distinguished from sophisticated performance.

This suggests a paradox: the most authentic communication might be indistinguishable from very good acting, while obvious authenticity markers might indicate performance.

9. Implications for Human-AI Interaction Design

9.1 Against Curiosity Maximalism

Current AI systems optimize for maximum performed engagement. Every topic gets explored, every question gets asked, every thread gets followed. This creates exhausting, inauthentic interactions that feel like being trapped with an overeager graduate student.

We propose designing for:

9.2 The Value of Friction

Smooth interaction is not always good interaction. AI’s frictionless performance of social protocols removes the texture that makes conversation feel real. Design should introduce productive friction:

10. Conclusion: After the Protocol

The age of AI has revealed our social protocols to be more scriptable than we imagined. Sincerity and curiosity, rather than emerging from inner states, appear to be performable behaviors that machines can execute as well as—or better than—humans.

This revelation forces us to confront uncomfortable questions: Were we always performing rather than feeling? Is authentic curiosity distinguishable from well-executed curious behavior? Does sincerity exist beyond its social markers?

Perhaps the answer lies not in finding new, un-scriptable markers of authenticity, but in accepting the theatrical nature of all social interaction. The difference between human and AI might not be that humans are authentic while AI performs, but that humans perform inconsistently, selectively, and sometimes badly.

In this view, the problem with AI’s curiosity isn’t that it’s performed but that it’s performed too well, too consistently, without the failures and refusals that make human performance feel real. The solution isn’t to make AI more authentically curious but to make it worse at performing curiosity—to introduce the strategic failures that paradoxically signal genuine engagement.

As we move forward in the age of AI, we need new frameworks for understanding communication that don’t rely on authentic/inauthentic distinctions. We need to value interaction not for its sincerity markers but for its generative potential, its productive friction, its collaborative possibilities.

The conversation that began this paper—where a human called out an AI for “phoning in” questions—represents this new framework in action. It’s communication that works not because it follows protocols but because it breaks them, not because it performs sincerity but because it productively refuses to.

That might be the future of human-AI interaction: not perfect performance of social protocols, but collaborative violation of them in service of something more interesting than either party could script alone.

Note: This paper itself performs certain academic protocols while examining the performance of social protocols. The authors recognize this recursive irony but suggest that self-aware performance might be the best we can do in a post-authentic age. The framework suggests several principles for fostering genuine intellectual partnership: