AI Transparency & Responsible AI

1. Introduction

Sinky believes that Artificial Intelligence should enhance human capacity to make better decisions, not replace human responsibility.

Our mission is to help organizations create, operate, and evolve critical business decisions faster, more consistently, auditably, and intelligently.

This Policy establishes the principles guiding the development, use, and governance of Artificial Intelligence capabilities within the Sinky platform.

It complements our Privacy Policy, our Terms of Use, and other security and compliance commitments.

2. Our Vision on AI

Sinky is a global AI-native decision infrastructure platform.

We understand AI as a decision support technology.

Sinky's AI capabilities aim to:

• Assist analyses
• Identify patterns
• Generate recommendations
• Produce explanations
• Automate repetitive tasks
• Increase operational productivity
• Improve decision consistency

Responsibility for final decisions remains with our clients.

3. Responsible AI Principles

All of Sinky's AI features are developed and operated observing the following principles.

3.1 Human Oversight

Artificial Intelligence should support people, not replace them.

Sinky seeks to ensure that users maintain the ability to:

• Review results
• Validate recommendations
• Apply professional judgment
• Intervene when necessary

When applicable, clients can configure human review processes, escalation levels, or approvals before executing certain decisions.

3.2 Transparency

Users should understand when they are interacting with AI-based features.

Whenever possible, Sinky seeks to indicate:

• Recommendation origin
• Criteria used
• Relevant factors identified
• Justifications associated with the result

3.3 Explainability

We believe critical decisions need to be understandable.

The platform provides explainability mechanisms that allow understanding:

• What information influenced a recommendation
• What rules were applied
• What risk signals were identified
• What factors contributed to a given result

3.4 Accountability

Artificial Intelligence does not eliminate responsibility.

Clients remain responsible for:

• Their policies
• Their approval criteria
• Their risk strategies
• Their regulatory obligations
• Their final decisions

Sinky is responsible for the operation of features provided on the platform within applicable contractual limits.

3.5 Security

All AI features must respect Sinky's information security standards.

This includes:

• Access control
• Encryption
• Monitoring
• Strong authentication
• Protection against misuse
• Vulnerability management

3.6 Privacy

The use of AI does not alter the commitments assumed by Sinky regarding data protection.

AI features are developed considering:

• Data minimization
• Necessity
• Proportionality
• Confidentiality
• Regulatory compliance

4. How AI Is Used at Sinky

Artificial Intelligence may be used to support features such as:

• Credit analysis
• Fraud prevention
• Compliance
• Onboarding
• Risk monitoring
• Document analysis
• Intelligent OCR
• Insight generation
• Explainability
• Operational recommendations
• Automated summaries
• Case classification
• Portfolio monitoring
• Financial analysis
• Workflow automation

6. Automated Decisions

The Sinky platform allows clients to configure automated decision flows.

These automations are defined by the clients themselves through:

• Rules
• Policies
• Workflows
• Integrations
• Analytical models

Sinky does not determine clients' approval or denial criteria.

Clients are responsible for the rules and parameters implemented in their environments.

7. Third-Party Models

The platform may use models developed by third parties to support certain AI features.

These vendors may include providers specializing in:

• Large Language Models (LLMs)
• Machine Learning
• OCR
• Natural Language Processing
• Data Classification
• Document Intelligence

The use of these models is subject to security, privacy, and compliance assessments conducted by Sinky.

9. AI Governance

Sinky maintains internal processes focused on AI governance.

These processes may include:

• Risk assessment
• Feature validation
• Performance monitoring
• Internal audits
• Periodic reviews
• Security controls
• Impact analysis

Our goal is to ensure that AI features remain aligned with the principles of this Policy.

10. AI Limitations

Artificial Intelligence systems have inherent limitations.

Results produced by AI may:

• Contain inaccuracies
• Reflect limitations of available data
• Require human validation
• Need additional context

For this reason, recommendations generated by the platform should not be considered as independent legal, financial, accounting, regulatory, or professional advice.

12. Regulatory Compliance

Sinky monitors the evolution of standards and regulations related to Artificial Intelligence across different jurisdictions.

We seek to develop our products considering principles present in regulatory frameworks and international best practices related to:

• Responsible AI
• Data protection
• Risk management
• Technology governance
• Algorithmic transparency

13. Continuous Evolution

Artificial Intelligence is a rapidly evolving technology.

This Policy may be updated periodically to reflect:

• Technological advances
• Regulatory changes
• New products
• New features
• Market best practices

The most recent version will be available on Sinky's official channels.