AI Transparency Notice

Artificial Intelligence can identify patterns, organize information, and assist with reflection.

It cannot fully understand the richness of a human life.

AI-generated observations are based on probabilities and available information—not certainty.

Accordingly, Arclly's AI may occasionally:

• misunderstand chronology
• miss sarcasm, irony, or humor
• misunderstand cultural or personal context
• overemphasize recent events
• underrepresent older experiences
• identify patterns that later disappear
• overlook meaningful life events that were never recorded
• generate observations that do not fully resonate with the user

Users are encouraged to view AI-generated insights as conversation starters rather than definitive conclusions.

Human lives are rarely simple.

Accordingly, Arclly's AI is designed to communicate uncertainty whenever appropriate.

Rather than presenting observations as absolute truth, Arclly may use language such as:

• "You may be noticing..."
• "Recent reflections suggest..."
• "It appears that..."
• "Based on your recent entries..."
• "This pattern may indicate..."
• "You might consider..."

This approach reflects our belief that thoughtful uncertainty is more trustworthy than false confidence.

One of Arclly's distinguishing capabilities is its commitment to explainability.

Whenever practical, Arclly attempts to show why an insight was generated by referencing the user's own reflections.

For example, an observation about resilience may be supported by several journal entries spanning multiple months rather than a single event.

Evidence Retrieval helps users understand:

• which reflections contributed to an insight
• how patterns developed over time
• why a summary was generated
• what memories appear connected

Our goal is to make AI feel understandable—not mysterious.

Arclly assists people.

It does not replace them.

Users remain responsible for:

• accepting or rejecting AI insights
• editing summaries
• revising MyArc narratives
• publishing memoirs
• inviting collaborators
• selecting Legacy Trustees
• gifting MyArc experiences
• making important personal decisions

AI proposes.

People decide.

Arclly has intentionally established boundaries for how AI may behave.

Our AI is designed not to:

• manipulate emotions
• pressure users into decisions
• exploit psychological vulnerabilities
• encourage unhealthy dependency
• fabricate memories or conversations
• invent supporting evidence
• impersonate family members or loved ones
• replace licensed medical, legal, financial, or mental health professionals
• claim certainty about a person's identity, future, or intentions
• make decisions on behalf of users

These boundaries are central to Arclly's commitment to trustworthy AI.

Many users choose to write during life's most difficult moments.

AI may encounter reflections relating to:

• grief
• trauma
• anxiety
• loneliness
• relationship conflict
• financial hardship
• illness
• identity
• spirituality
• personal loss

Arclly is designed to respond with empathy, clarity, and respect.

It is intentionally designed to avoid:

• judgment
• shame
• sensationalism
• manipulation
• emotionally exploitative language

Some users may record thoughts relating to:

• self-harm
• suicide
• hopelessness
• severe emotional distress

Where AI encounters this type of content while processing a requested feature, Arclly may:

• acknowledge the user's feelings with compassionate language
• encourage the user to seek support from someone they trust
• encourage contact with local emergency services or appropriate crisis resources if the user believes they are in immediate danger
• provide links to publicly available support resources where appropriate

Arclly does not:

• diagnose mental health conditions
• provide therapy or crisis counseling
• continuously monitor journal entries
• notify emergency contacts
• notify family members
• contact emergency services on a user's behalf
• guarantee that every expression of distress will be detected

If you believe you or someone else is in immediate danger, contact your local emergency services or an appropriate crisis support service immediately.

Voice allows people to capture thoughts naturally and preserve memories in their own words.

Where enabled, AI may assist by:

• transcribing spoken reflections
• improving readability
• generating summaries
• organizing narrated memories
• preparing audio versions of MyArc
• supporting future Voice Continuity experiences

Voice recordings remain under the user's control.

Persistent voice identity features, voice modeling, or Voice Continuity require separate, explicit user consent.

Users may disable voice features at any time through available account settings.

MyArc represents one of Arclly's most advanced AI experiences.

AI may assist by:

• organizing memories into chapters
• identifying recurring themes
• preserving chronology
• drafting narrative structure
• suggesting chapter titles
• identifying meaningful transitions
• supporting memoir creation

AI never determines what your life story should become.

Only you decide:

• what is included
• what is edited
• what is removed
• what is shared
• what becomes part of your legacy

Privacy is fundamental to how Arclly designs AI.

Unless users separately opt in where such an option is provided, Arclly does not use users' private journal entries, reflections, MyArc content, or voice recordings to train general-purpose AI models.

Content classified as safety-critical or protected is excluded from aggregated or de-identified use as well as model training.

Where third-party AI providers are used to deliver requested features, Arclly seeks contractual protections that prohibit the use of user content for unrelated commercial model training.

Additional information about how personal information is processed is available in the Arclly Privacy Policy and Consent & Sensitive Data Policy.

Arclly continuously evaluates its AI systems to improve:

• factual consistency
• emotional appropriateness
• evidence quality
• hallucination reduction
• fairness
• safety
• usefulness
• latency
• reliability
• overall user experience

Updates are introduced carefully and are intended to enhance—not fundamentally alter—the trusted experience users have come to expect.

Artificial Intelligence at Arclly is governed by policies, technical safeguards, and human oversight designed to ensure that our AI systems remain aligned with our values.

AI development is guided by the following objectives:

• protect user privacy
• improve understanding—not dependency
• minimize harmful outputs
• promote fairness
• maintain transparency
• continuously improve quality
• preserve human control over important decisions

AI features are reviewed throughout development and after release to evaluate quality, safety, and alignment with Arclly's principles.

Artificial Intelligence does not replace human responsibility.

Arclly's team remains accountable for:

• product design
• AI safety decisions
• model selection
• prompt engineering
• quality evaluation
• privacy protection
• security practices
• policy enforcement

Where meaningful product decisions affect users, responsibility ultimately rests with people—not algorithms.

AI systems can unintentionally reflect bias present in training data or user inputs.

Arclly actively works to reduce unfair or unintended outcomes by:

• evaluating AI outputs across diverse scenarios
• monitoring for inappropriate language
• testing for cultural sensitivity
• improving prompt design
• reviewing edge cases
• encouraging user feedback

Despite these efforts, no AI system can guarantee completely bias-free outputs.

Users are encouraged to critically evaluate AI-generated insights and provide feedback whenever something feels inaccurate or inappropriate.

Artificial Intelligence is continually evolving.

As Arclly grows, AI models, prompts, and product features may improve to provide:

• more accurate summaries
• better emotional understanding
• stronger evidence retrieval
• improved narrative quality
• reduced hallucinations
• faster responses
• safer interactions
• more meaningful long-term insights

Whenever significant AI capabilities change, this AI Transparency Notice may also be updated.

Arclly's long-term vision includes responsibly introducing new AI experiences, including:

• richer contextual understanding
• wearable integrations
• life rhythm intelligence
• ambient memory assistance
• advanced MyArc storytelling
• family continuity experiences
• improved Voice Continuity
• multilingual memoir generation

Any future capability involving materially different processing of personal information will be introduced in accordance with applicable privacy laws and, where required, user consent.

Arclly is designed to support your thinking—not replace it.

You remain the author of your life story.

We encourage users to:

• reflect thoughtfully
• verify important information
• edit AI-generated narratives
• share only what feels appropriate
• seek qualified professionals for medical, legal, financial, or mental health advice

The best outcomes come from combining human wisdom with thoughtful AI assistance.

We believe trustworthy AI improves through listening.

If you believe an AI response is:

• inaccurate
• confusing
• emotionally inappropriate
• unfair
• misleading
• offensive
• inconsistent with Arclly's values

we encourage you to provide feedback through the application or by contacting our support team.

User feedback helps improve future versions of the platform.

As Arclly evolves, we may update this AI Transparency Notice to reflect:

• new AI capabilities
• improved safeguards
• changes in applicable law
• updates to our governance practices
• enhancements to user controls

Where required by law, users will be notified of material changes before they become effective.

The latest version of this Notice will always be available through Arclly.

If you have questions regarding Arclly's use of Artificial Intelligence, please contact us.

Arclly, LLC

Email: support@arclly.com

Website: www.arclly.com

We welcome thoughtful questions about how our AI works and how it can continue to improve.