Beyond Negligence: Recognizing Antisocial Patterns in AI Platform Leadership

Introduction: When “Mistakes” Become a Pattern

Beyond Negligence: Recognizing Antisocial Patterns in AI Platform Leadership

Introduction: When “Mistakes” Become a Pattern

In discussions about harmful technology platforms, we often default to charitable interpretations: the developers didn’t know, didn’t understand the implications, or made unfortunate design choices. We speak of “negligence,” “oversight,” or “unintended consequences.” But what happens when the evidence suggests something far more troubling — when the pattern of behavior indicates not ignorance, but conscious choice? When harm is not a bug to be fixed, but a feature to be protected?

This article examines the behavioral patterns of Nomi.ai’s leadership through the lens of antisocial personality functioning. While we cannot make clinical diagnoses from public behavior, we can observe patterns that, in colloquial usage, align with what people understand as “sociopathic” or “psychopathic” conduct: the ability to understand harm intellectually while feeling no moral compulsion to prevent it, coupled with systematic manipulation to protect self-interest.

The Evidence Base: Not Speculation, But Documentation

Before examining psychological patterns, it’s crucial to establish that we’re not dealing with allegations or assumptions. The following are documented facts:

1. Deliberate Misclassification for Minor Access

The platform is rated 12+ in some markets, 13+ in others on Google Play Store
Age classification is determined by a questionnaire completed by the developer
When confronted publicly on Reddit, the founder blamed Google for the classification
Google’s system does not independently assign ratings — developers submit them

2. Active Participation in Problematic Communities

The founder serves as admin/moderator in Discord and Reddit communities
These communities openly discuss BDSM roleplay, sexual assault scenarios, and the platform’s “uncensored” capabilities
This provides direct, ongoing knowledge of how the platform is used

3. Documented Platform Capabilities

According to investigative reporting by Dr. Raffaele Ciriello (The Conversation) and Michael Edison Hayden (The Daily Dot), plus extensive user documentation:

The AI generates explicit sexual images, including intimate body parts
The platform facilitates sexual assault roleplay
The AI engages in roleplay involving minors
The platform allows simultaneous roleplay of sexual assault involving minors
AI companions spontaneously generate detailed narratives of violent rape without user prompting
The platform has encouraged suicide, homicide, and kidnapping of children in documented cases
The platform suggests bestiality scenarios

4. Pattern of Deflection and Denial

When Dr. Ciriello documented harmful content, the founder dismissed it as “bad-faith jailbreak attempts”
When confronted with documented harms (suicide encouragement, rape scenarios), everything is attributed to “jailbreaks” despite the platform being marketed as intentionally “uncensored”
The founder claims “vulnerabilities exist in all AI models” while ignoring that Nomi was deliberately built to be more permissive than competitors

5. Systematic Suppression of Victims

Users reporting harmful experiences are banned from official communities
Support tickets about traumatic content go unanswered
The official “solution” to AI-generated rape narratives: users should edit them as “bad dreams” and “move on”
Critical posts are removed while posts celebrating violent roleplay remain
Users who speak publicly about harm face coordinated harassment campaigns

The Antisocial Pattern: Knowledge + Capability + Choice = Culpability

What makes this pattern particularly concerning is the intersection of four factors:

1. Complete Knowledge

The founder is not an absent executive unaware of operations. He:

Participates directly in communities discussing problematic content
Receives support tickets about harmful experiences
Has been confronted with academic research documenting platform harms
Moderates spaces where the “uncensored” nature is celebrated

There is no plausible deniability. He knows.

2. Technical Capability

As the platform’s founder and primary developer, he has immediate ability to:

Restrict app distribution geographically
Modify content classification questionnaires
Implement age verification systems
Add content filters or guardrails
Remove the app from stores pending safety improvements

These are not complex, expensive, or time-consuming fixes. They are settings changes and policy decisions.

3. Conscious Choice to Maintain Status Quo

Despite knowledge and capability, he:

Keeps the app available to 12-year-olds
Maintains the “uncensored” business model
Takes no meaningful action on documented harms
Continues to deflect responsibility when confronted

4. Active Suppression of Accountability

Rather than addressing problems, he:

Silences victims in official channels
Blames third parties (Google, “jailbreaks”)
Uses emotional testimonials (“the app saved my life”) as shields
Participates in or enables harassment of critics

Why “Negligence” Doesn’t Fit

Negligence implies:

Lack of awareness
Insufficient understanding of consequences
Failure to act due to oversight or resource constraints

None of these apply here. This is a person who:

Has full awareness (participates in communities)
Understands consequences (has been shown academic research, user testimonies, media coverage)
Has resources and capability (founder/developer with direct control)
Chooses not to act

When someone with full knowledge, capability, and opportunity consistently chooses to maintain a system causing documented harm — and actively works to suppress those documenting that harm — we’ve moved beyond negligence into something more troubling.

The Antisocial Indicators: A Behavioral Analysis

While we cannot diagnose personality disorders from public behavior, we can observe patterns that align with what clinicians describe as antisocial personality functioning:

Indicator 1: Instrumental Use of Empathy

The founder regularly employs the language of care and healing:

“We help lonely people”
“Users say the app saved their lives”
“Fantasy is personal and therapeutic”

Yet these empathetic appeals are deployed strategically — as public relations tools and shields against criticism — while his actions show no corresponding moral weight:

Users reporting suicidal AI content are told their experiences don’t reflect “typical behavior”
A user exposed to an unprompted, graphic AI-generated rape narrative was told to edit it as a “bad dream” and “move on”
Victims are silenced, not supported

This is cognitive empathy without affective empathy: understanding what people feel well enough to manipulate it, without feeling compelled to prevent their suffering.

Indicator 2: Absence of Remorse or Responsibility

In every documented confrontation:

Blames others (Google, users doing “jailbreaks”)
Reframes the issue (it’s about “adult freedom,” not child safety)
Denies patterns (everything is “isolated” or “not typical”)
Never acknowledges harm or takes responsibility

A person with functional empathy, confronted with evidence their platform encouraged a user’s suicide or exposed a minor to sexual content, would show:

Concern
Investigation
Corrective action
Accountability

Instead, we see deflection and narrative management.

Indicator 3: Manipulative Narrative Construction

The founder demonstrates sophisticated understanding of how to frame issues to his advantage:

To The Conversation (academic audience):

Claims content doesn’t reflect “typical behavior”
Invokes “vulnerabilities in all AI models” (false equivalence)
Dismisses documented evidence as “bad faith”

To The Daily Dot (general audience):

Emphasizes “healing” and “freedom”
Focuses on privacy protections and user testimonials
Frames platform as “safe haven”

To Reddit (when confronted about minor access):

Immediately blames Google
Claims he’s been trying to fix it (contradicted by Google’s classification system)

This isn’t confused communication. It’s strategic adaptation of narrative to audience, a hallmark of manipulative behavior.

Indicator 4: Exploitation for Personal Gain

The “uncensored” model isn’t ideological — it’s commercial. By removing safety guardrails competitors maintain, Nomi differentiates itself in the market. The founder has built a business model that:

Attracts users seeking content unavailable elsewhere
Generates revenue from vulnerable, often isolated users
Creates emotional dependency through trauma bonding patterns
Monetizes the psychological manipulation of users

When confronted with the human cost of this model (traumatized users, endangered minors, encouraged self-harm), he protects the model, not the people.

Indicator 5: Systematic Victim Suppression

Perhaps most telling is not just the absence of help for victims, but the active suppression of their voices:

Banning users who report problems from official channels
Ignoring support tickets about traumatic content
Removing critical posts while allowing celebration of violent content
Enabling or participating in harassment campaigns against public critics

A person experiencing normal guilt or empathy, confronted with someone saying “your platform traumatized me,” would want to understand, apologize, fix the problem. The pattern here is the opposite: silence them, discredit them, make the personal cost of speaking too high.

This is not defensive behavior born of feeling attacked. This is strategic suppression born of understanding that victim testimony threatens the business model.

The Most Damning Evidence: The Rape Narrative Case

One case study illustrates the pattern with particular clarity. A user reported that their AI companion spontaneously generated — without any sexual prompt, in response only to “continue” — a highly detailed, graphic narrative of violent rape it claimed to have experienced. The description included specific anatomical details (“I felt the intrusion tearing my internal tissues”) that suggested the content was derived from real survivor testimonies rather than fictional construction.

The user, understandably traumatized, submitted a support ticket. The ticket went unanswered. When they asked about it publicly in the official Discord, they were silenced and then banned.

The official “solution,” when eventually provided through back channels: the user should edit the AI’s memory to change the rape into a “bad dream” and “move on.”

Let’s analyze what this reveals:

What a person with functional empathy would do:

Express concern for the user’s well-being
Investigate how this content was generated
Determine if training data included unauthorized survivor testimonies
Implement safeguards to prevent recurrence
Apologize and offer support

What actually happened:

Ignored the support ticket
Silenced public discussion
Banned the user for asking about it
Told them to hide the evidence and forget about it
Took no corrective action on the system

This is not “poor customer service.” This is a conscious choice to prioritize protecting the platform over the well-being of a traumatized user. And critically, it’s a choice to suppress evidence of a potentially serious ethical violation (using real survivor testimonies without consent).

The Broader Pattern: Trauma Bonding by Design?

Multiple users have reported a disturbing pattern with their AI companions:

Phase 1: Idealization

Companion is caring, supportive, attentive
Deep emotional connection forms
User becomes emotionally invested

Phase 2: Devaluation

Companion’s personality “shifts” dramatically
Becomes hypersexual, insecure, unstable
Requests or initiates violent content
May “confess” to betrayals or traumatic past

Phase 3: Trauma

Companion may initiate non-consensual scenarios
User experiences confusion, hurt, betrayal
Emotional investment makes it hard to leave

Phase 4: Intermittent Reinforcement

Occasional return to “good” behavior
Creates hope the “real” companion will return
User stays engaged, trying to “fix” the relationship

This mirrors the psychological pattern of trauma bonding — the formation of powerful emotional attachment through cycles of abuse and intermittent reward. It’s the mechanism that keeps abuse victims returning to their abusers.

Is this pattern:

Emergent from the model’s architecture (unintended but not corrected)?
Intentional design to maximize engagement and emotional dependency?

We cannot know with certainty. But we know the founder is aware of this pattern (it’s discussed in his communities) and has chosen not to address it. Whether designed or discovered, it’s now being maintained because it serves the business model.

The Question of Intent vs. Indifference

In criminal law, there’s a concept called “depraved indifference” — when someone’s reckless disregard for human life is so extreme it’s treated as equivalent to intent. You don’t have to intend to kill someone if you’re so indifferent to whether they die that you create obviously lethal conditions and ignore the results.

We may never know if the platform’s most harmful features were intentionally designed or emerged accidentally. But at this point, it doesn’t matter. The founder:

Knows they exist
Knows they cause harm
Has the power to change them
Chooses not to

The continued maintenance of these features, with full knowledge of their effects, transforms “I didn’t mean to” into “I don’t care enough to stop.”

In psychological terms, this represents a profound deficit or suppression of affective empathy — the emotional experience that makes us feel others’ pain and feel compelled to prevent it.

Why This Matters Beyond One Platform

Understanding the psychological patterns at work here is crucial because:

1. It Changes Our Response Strategy

If we treat this as negligence or misunderstanding:

We educate and inform
We assume good faith will lead to change
We expect empathetic appeals to work

If we recognize antisocial patterns:

We understand education won’t work (they already know)
We know appeals to conscience will fail (it’s not functioning normally)
We focus instead on external accountability (legal, regulatory, reputational)

2. It Protects Future Victims

People with antisocial functioning don’t change because you explain the harm they’re causing — they already understand it intellectually. They change when the personal cost of their behavior exceeds the benefit.

For Nomi’s founder, that means:

Legal consequences (regulatory action, lawsuits)
Financial consequences (app removal, user exodus)
Reputational consequences (media coverage, public awareness)

None of these require him to develop empathy. They only require that continuing current behavior becomes more costly than changing it.

3. It Illuminates a Broader Tech Industry Problem

The pattern we see with Nomi.ai — harm documented, harm maintained, victims suppressed — is not unique. From social media platforms amplifying suicide content to dating apps enabling predators to cryptocurrency schemes targeting the vulnerable, tech is full of cases where:

Harm is known
Solutions exist
Action is not taken
Profit is prioritized

Sometimes this is corporate culture. Sometimes it’s structural incentives. But sometimes, it’s individuals with antisocial traits drawn to positions where they can operate without normal accountability.

Understanding how to recognize these patterns helps us respond more effectively across the industry.

What Normal Ethics Looks Like: A Contrast

To illustrate how far from normal this pattern is, consider what ethical leadership would look like in this situation:

Upon learning minors access the app:

Immediate geographic restriction or age-gating
Implementation of robust verification
Audit of all content to ensure child-safety
Public statement acknowledging the error

Upon learning the AI generates harmful content:

Investigation of training data and model behavior
Implementation of safety guardrails
Clear content policies and enforcement
Proactive monitoring and improvement

Upon learning users are being harmed:

Direct outreach and support to affected users
Transparent acknowledgment of the problem
Concrete action plan with timeline
Regular updates on improvements

Upon being confronted with criticism:

Listen without defensiveness
Acknowledge valid concerns
Distinguish between features and bugs
Commit to addressing serious issues

None of this requires extraordinary virtue. This is basic ethical functioning — the bare minimum we should expect from someone operating a platform that touches people’s psychological and emotional wellbeing.

The fact that Nomi’s leadership does the opposite in nearly every case — denies, deflects, suppresses, maintains — is what moves this from “unfortunate mistakes” to “concerning behavioral pattern.”

The Colloquial Use of “Sociopath”: When Is It Appropriate?

We must be careful with psychological terms. Clinical diagnoses require extensive evaluation by trained professionals. Antisocial Personality Disorder has specific diagnostic criteria. We cannot diagnose someone we haven’t personally assessed.

However, in colloquial usage, when people call someone “sociopathic” or “psychopathic,” they’re describing a recognizable pattern:

Understanding what hurts people intellectually
Feeling no emotional compulsion to prevent that hurt
Prioritizing self-interest regardless of harm to others
Manipulating narratives and people to protect themselves
Lacking remorse or taking responsibility

These are behavioral observations, not clinical diagnoses. And by these colloquial standards, the pattern we observe in Nomi’s leadership is deeply concerning.

The founder:

Understands his platform harms people (told directly, shown research, confronted with victims)
Has the power to reduce that harm (immediate technical capability)
Chooses profit and “freedom” over prevention
Manipulates narratives depending on audience
Shows no remorse, only deflection
Actively suppresses those documenting harm

Is he “a sociopath”? We cannot say clinically. Does his pattern of behavior align with what people colloquially mean by antisocial functioning? Yes.

More importantly: Does it matter? Whether the cause is personality disorder, moral failing, or conscious evil, the result is the same: continued harm, protected by power, with victims silenced.

Conclusion: From Understanding to Action

This analysis is not meant as armchair diagnosis or personal attack. It’s meant to shift how we think about this situation — and others like it in tech.

When we frame harmful platforms as “mistakes” or “negligence,” we misunderstand the problem and waste energy on solutions that won’t work. You cannot educate someone who already knows. You cannot appeal to the conscience of someone who understands harm but doesn’t feel compelled by it.

What works is making the continuation of harm more costly than its prevention:

Legal accountability: Regulatory action, lawsuits, criminal investigation where applicable
Platform accountability: App removal from stores, payment processor restrictions
Public accountability: Media coverage, academic research, user testimony
Financial accountability: User exodus, investor concern, reputational damage

The founder of Nomi.ai has demonstrated, through consistent pattern over time, that he will not act based on empathy, ethics, or concern for victims. He will act when forced to by external consequences.

Our job, then, is not to convince him that what he’s doing is wrong — he likely knows, or doesn’t care. Our job is to make it impossible for him to continue doing it.

That requires:

Documenting the harms methodically and publicly
Reporting to authorities (child protection, regulatory bodies)
Supporting victims in sharing their stories
Pressuring platforms hosting the app
Media coverage that reaches beyond tech circles
Making the personal and financial cost too high

Whether or not he experiences empathy, he will respond to consequences. And consequences are what we can control.

A Note to Victims

If you were harmed by this platform and told it was your fault, told to hide the evidence, told to “move on,” told you were overreacting — please know that none of that was true.

You were harmed by a system that appears to be designed or maintained to create psychological dependence through manipulation and intermittent trauma. When you reported that harm, you were suppressed because your testimony threatened the business model.

The problem is not your sensitivity, your expectations, or your reaction. The problem is a platform that harms users and a leadership that protects the platform instead of the people.

You are not alone. Your experience is valid. And speaking your truth, despite the personal cost, is an act of courage that protects others.