Terms of Service for Kimi K2

Effective Date: January 15, 2025
Last Updated: January 15, 2025
Version: 1.0

Agreement Overview
Service Description
Account Registration and Management
Acceptable Use Policy for Agentic AI
Intellectual Property Rights
Agentic AI Terms and Autonomous Operations
Open Source License and Contributions
API Access and Developer Terms
Data Privacy and AI Safety
Service Availability and Performance
User Responsibilities and AI Ethics
Limitation of Liability
Indemnification
Termination and Suspension
Dispute Resolution
Changes to Terms
Contact Information

Agreement Overview

1.1 Introduction and Acceptance

Welcome to Kimi K2, MoonshotAI's revolutionary 1 trillion parameter mixture-of-experts model that pioneers true agentic intelligence, transforming AI from mere response generation to autonomous problem-solving and action execution. These Terms of Service ("Terms," "Agreement") constitute a legally binding contract between you ("User," "you," "your") and MoonshotAI ("we," "us," "our," "Company") governing your access to and use of our agentic AI model and related services.

1.2 Binding Agreement for Agentic AI

By accessing, using, or interacting with Kimi K2 in any manner, you acknowledge that you have:

Read and understood these Terms regarding agentic AI capabilities
Agreed to be legally bound by all provisions for autonomous AI operations
Confirmed your legal capacity to enter into agreements involving AI systems
Accepted our Privacy Policy and AI safety guidelines
Understood the autonomous nature of agentic AI operations

1.3 Age and Capacity Requirements for AI Systems

To use Kimi K2's agentic capabilities, you must:

Be at least 18 years of age (due to autonomous AI nature)
Have the technical understanding of AI systems and their limitations
Possess legal authority to make decisions involving autonomous AI
Not be prohibited from using advanced AI systems under applicable law
Accept responsibility for autonomous AI-generated actions

1.4 Agreement Hierarchy and AI Governance

These Terms incorporate by reference:

Privacy Policy - Our data handling practices for agentic AI
AI Safety Guidelines - Safety protocols for autonomous operations
Acceptable Use Policy - Ethical guidelines for agentic AI usage
Open Source License - Terms for open source agentic AI components
Research Ethics - Guidelines for AI research and development

Service Description

2.1 Revolutionary Agentic AI Services

Kimi K2 provides unprecedented agentic intelligence capabilities, including:

Autonomous Intelligence Features

True Agentic Reasoning - Autonomous problem decomposition and solution planning
Multi-Step Execution - Independent completion of complex, multi-phase tasks
Contextual Understanding - Deep comprehension with 128K token context window
Expert Network Routing - Intelligent utilization of specialized expert networks
Adaptive Learning - Real-time adaptation to user preferences and contexts

Mixture-of-Experts Architecture

Trillion Parameter Scale - Unprecedented model capacity and intelligence
320 Billion Active Parameters - Efficient activation for optimal performance
Expert Specialization - Specialized networks for different domains and tasks
Dynamic Routing - Intelligent selection of optimal expert combinations
Scalable Performance - Efficient scaling across diverse use cases

Research and Development Platform

Open Source Core - Transparent and inspectable agentic AI technology
Academic Collaboration - Platform for advancing agentic AI research
Safety Research - Continuous improvement of AI safety and alignment
Community Development - Collaborative improvement of agentic capabilities

2.2 Service Tiers and Agentic Access Levels

Open Source Access (Free)

Full agentic capabilities with standard response times
Complete model access to both base and instruct variants
Unlimited usage for research and development purposes
Community support through forums and documentation
Open source transparency with full code accessibility

Professional API Access

Enhanced performance with optimized infrastructure
Priority processing for time-sensitive agentic operations
Advanced integrations with enterprise development tools
Professional support with dedicated technical assistance
Custom fine-tuning capabilities for specialized applications

Enterprise Solutions

On-premises deployment options for sensitive environments
Custom model training on organization-specific data
Advanced security features and compliance certifications
24/7 dedicated support with SLA guarantees
White-label solutions for product integration

2.3 Open Source and Research Commitment

Kimi K2 advances the state of agentic AI through:

Transparent development - Open source core for community inspection
Research collaboration - Partnerships with leading academic institutions
Safety research - Ongoing research into AI safety and alignment
Knowledge sharing - Publication of research findings and methodologies

Account Registration and Management

3.1 Account Creation for Agentic AI Access

To access enhanced agentic AI features, you may create an account by providing:

Required Information

Valid email address - Primary communication for AI safety and updates
Technical competency verification - Basic understanding of AI systems
Use case declaration - Intended applications for agentic AI capabilities
Responsibility acknowledgment - Understanding of autonomous AI implications

Professional and Research Information

Organization affiliation - Academic or professional institution details
Research purpose - Academic or commercial research objectives
Safety training - Completion of AI safety and ethics training modules
Collaboration interests - Areas for potential research collaboration

3.2 Account Security for AI Systems

Enhanced security requirements for agentic AI access:

Advanced Authentication

Multi-factor authentication - Required for all agentic AI access
Biometric verification - Optional enhanced security measures
Device registration - Approved devices for autonomous AI operations
Session monitoring - Continuous monitoring of AI interaction sessions

Responsibility and Compliance

Legal compliance - Adherence to AI governance and regulatory requirements
Ethical usage - Commitment to responsible AI development and deployment
Safety protocols - Implementation of appropriate safety measures
Incident reporting - Prompt reporting of safety incidents or concerns

Acceptable Use Policy for Agentic AI

4.1 Permitted Uses of Agentic Intelligence

You may use Kimi K2's agentic capabilities for legitimate purposes, including:

Research and Development

AI research - Advancing the field of agentic artificial intelligence
Academic studies - Educational and scholarly research applications
Product development - Building beneficial applications with agentic AI
Open source contributions - Contributing to the agentic AI ecosystem

Professional and Commercial Applications

Business automation - Legitimate business process automation
Creative assistance - Content creation and creative problem-solving
Technical consulting - Professional services enhanced by agentic AI
Educational tools - Teaching and learning applications

Safety and Beneficial Uses

Scientific discovery - Using agentic AI for scientific advancement
Social good - Applications that benefit society and humanity
Accessibility - Tools that improve accessibility for people with disabilities
Environmental solutions - Applications addressing environmental challenges

4.2 Prohibited Uses of Agentic AI

You may not use Kimi K2 for:

Harmful or Dangerous Activities

Autonomous weapons - Development of lethal autonomous weapon systems
Social manipulation - Large-scale manipulation of public opinion or behavior
Privacy invasion - Unauthorized surveillance or privacy violations
Discrimination - Perpetuating or amplifying harmful biases and discrimination

Illegal or Unethical Applications

Illegal activities - Any activity that violates applicable laws or regulations
Fraud or deception - Using agentic AI for fraudulent or deceptive purposes
Harassment - Automated harassment or abuse of individuals or groups
Misinformation - Deliberately spreading false or misleading information

Misuse of Autonomous Capabilities

Unsupervised critical decisions - Using agentic AI for life-critical decisions without oversight
Financial manipulation - Autonomous market manipulation or financial fraud
Infrastructure attacks - Using agentic capabilities to attack critical infrastructure
Model extraction - Attempting to extract or replicate our agentic AI models

4.3 Enforcement and AI Safety Measures

Violations may result in:

Immediate suspension - For safety-critical violations
Safety review - Comprehensive assessment of usage patterns
Mandatory training - Required completion of AI safety education
Permanent termination - For severe violations or repeated offenses
Legal action - Referral to authorities for illegal activities

Intellectual Property Rights

5.1 MoonshotAI Intellectual Property in Agentic AI

Kimi K2 Technology and Innovation

Agentic AI algorithms - Proprietary methods for autonomous intelligence
Mixture-of-experts architecture - Novel approaches to expert network design
Training methodologies - Advanced techniques for agentic AI development
Safety mechanisms - Proprietary safety and alignment technologies

Open Source Components

Core agentic framework - Open source foundation for transparency
Community contributions - Code and improvements from the community
Research publications - Open sharing of research findings and methodologies
Educational resources - Training materials and documentation

5.2 User Rights and Responsibilities

Your Content and Interactions

Input ownership - You retain ownership of your inputs to Kimi K2
Generated content rights - Rights to content created with Kimi K2's assistance
Research data - Ownership of research data and findings from your work
Collaborative contributions - Rights in collaborative research and development

License Grant for Improvement

By using Kimi K2, you grant us a limited license to:

Improve agentic capabilities - Using aggregated, anonymized interaction data
Enhance safety measures - Learning from usage patterns to improve AI safety
Research advancement - Contributing to open research in agentic AI
Community benefit - Sharing improvements with the open source community

5.3 Respect for Third-Party Rights in AI Context

Ethical AI Development

Training data rights - Respect for intellectual property in training data
Fair use principles - Adherence to fair use in AI training and deployment
Attribution requirements - Proper attribution of sources and influences
License compliance - Compliance with open source and proprietary licenses

Agentic AI Terms and Autonomous Operations

6.1 Understanding Agentic Intelligence

Nature of Agentic AI

Autonomous decision-making - Kimi K2 can make independent decisions within defined parameters
Multi-step reasoning - Capability to plan and execute complex, multi-phase tasks
Contextual adaptation - Real-time adjustment to changing circumstances and requirements
Goal-directed behavior - Oriented toward achieving specified objectives autonomously

Limitations and Boundaries

Domain constraints - Agentic capabilities are limited to defined operational domains
Safety boundaries - Built-in limitations to prevent harmful or dangerous actions
Human oversight - Mechanisms for human intervention and control when necessary
Uncertainty handling - Appropriate responses to ambiguous or uncertain situations

6.2 Autonomous Operation Guidelines

Responsible Agentic AI Use

Clear objective setting - Provide clear, ethical objectives for autonomous operations
Appropriate supervision - Maintain appropriate levels of human oversight
Safety verification - Verify safety measures before autonomous execution
Result validation - Review and validate autonomous AI outputs and actions

Risk Management for Autonomous AI

Impact assessment - Evaluate potential impacts of autonomous operations
Fallback procedures - Establish procedures for handling unexpected situations
Monitoring systems - Implement monitoring for autonomous AI behavior
Escalation protocols - Clear procedures for escalating concerning behavior

6.3 User Responsibility for Agentic Operations

Decision Authority and Accountability

Final responsibility - Users maintain ultimate responsibility for agentic AI decisions
Ethical oversight - Ensuring agentic operations align with ethical principles
Legal compliance - Ensuring autonomous actions comply with applicable laws
Harm prevention - Taking measures to prevent harm from autonomous operations

Professional Standards

Industry best practices - Adherence to emerging best practices in agentic AI
Continuous learning - Staying informed about agentic AI developments and risks
Peer collaboration - Engaging with the community on agentic AI safety and ethics
Transparency - Being transparent about agentic AI use when appropriate

Open Source License and Contributions

7.1 Open Source Agentic AI Framework

License Compliance for Agentic Components

Apache 2.0 License - Core agentic framework under permissive open source license
Derivative works - Freedom to create derivative works with proper attribution
Commercial use - Permission for commercial applications of open source components
Distribution requirements - Compliance with license requirements for distribution

Community-Driven Development

Collaborative improvement - Community participation in agentic AI advancement
Transparent development - Open development process for core capabilities
Democratic governance - Community input on major architectural decisions
Safety collaboration - Collective work on AI safety and alignment

7.2 Contributing to Agentic AI Development

Contribution Standards

Safety-first approach - All contributions must prioritize AI safety
Ethical review - Ethics review for contributions affecting agentic behavior
Quality assurance - Rigorous testing and validation of agentic AI components
Documentation - Comprehensive documentation for agentic AI features

Research Collaboration

Academic partnerships - Collaboration with leading AI research institutions
Open publication - Sharing research findings with the broader community
Safety research - Coordinated research on agentic AI safety and alignment
Standards development - Contributing to industry standards for agentic AI

7.3 Intellectual Property in Agentic AI Contributions

Contributor Recognition

Research attribution - Proper attribution for research contributions
Patent considerations - Clear policies on patents related to agentic AI
Publication rights - Rights to publish research based on contributions
Community standing - Recognition within the agentic AI research community

API Access and Developer Terms

8.1 Agentic AI API Usage

Permitted API Applications

Agentic integrations - Building applications that leverage autonomous AI capabilities
Research platforms - Developing tools for agentic AI research and experimentation
Educational tools - Creating educational applications for learning about agentic AI
Beneficial automation - Automating beneficial tasks with appropriate oversight

API Safety Requirements

Safety integration - Implementing appropriate safety measures in API applications
Human oversight - Maintaining human oversight for autonomous operations via API
Rate limiting - Respecting rate limits to ensure system stability and safety
Error handling - Robust error handling for autonomous AI operations

8.2 Developer Responsibilities for Agentic AI

Application Safety Standards

Safety by design - Building safety considerations into application architecture
User disclosure - Clear disclosure of agentic AI capabilities in applications
Consent mechanisms - Appropriate consent for autonomous AI operations
Monitoring systems - Implementing monitoring for agentic AI behavior in applications

Ethical Development Practices

Bias mitigation - Taking steps to mitigate bias in agentic AI applications
Transparency - Providing transparency about agentic AI capabilities and limitations
Accessibility - Ensuring applications are accessible to diverse user populations
Privacy protection - Implementing strong privacy protections for user data

8.3 API Evolution and Agentic Capabilities

Capability Updates

Safety-first updates - Prioritizing safety in capability enhancements
Breaking change policies - Clear policies for changes affecting agentic behavior
Migration support - Assistance with migrating to enhanced agentic capabilities
Community input - Involving the developer community in capability development

Data Privacy and AI Safety

9.1 Privacy in Agentic AI Operations

Data Protection for Autonomous Operations

Privacy by design - Built-in privacy protections for agentic AI operations
Minimal data principle - Using only necessary data for autonomous operations
Purpose limitation - Limiting data use to specified autonomous objectives
User control - Granular user control over data usage in agentic operations

Autonomous Data Processing

Consent mechanisms - Clear consent for autonomous data processing
Processing transparency - Transparency about how data is processed autonomously
Data minimization - Minimizing data collection for autonomous operations
Deletion rights - Rights to delete data used in agentic operations

9.2 AI Safety in Autonomous Operations

Safety Architecture

Constitutional AI - Built-in principles governing autonomous behavior
Value alignment - Training for beneficial and safe autonomous actions
Harm prevention - Multiple safeguards against potential harm
Human oversight - Mechanisms for human intervention in autonomous operations

Continuous Safety Monitoring

Behavioral monitoring - Continuous assessment of agentic AI behavior
Safety metrics - Quantitative measures of safety performance
Incident response - Rapid response protocols for safety incidents
Community reporting - Mechanisms for community safety reporting

9.3 Research Ethics and AI Governance

Ethical Research Practices

IRB review - Institutional Review Board approval for research involving human subjects
Consent protocols - Appropriate consent for research participation
Data governance - Ethical governance of research data
Publication ethics - Ethical standards for publishing agentic AI research

Regulatory Compliance

AI governance - Compliance with emerging AI governance frameworks
International standards - Adherence to international AI safety standards
Industry best practices - Following emerging best practices in agentic AI
Regulatory cooperation - Cooperation with regulatory authorities

Service Availability and Performance

10.1 Agentic AI Service Levels

Performance Standards for Autonomous Operations

Response times - Target response times for agentic AI operations
Reliability - High reliability standards for autonomous task execution
Scalability - Ability to scale autonomous operations as needed
Quality metrics - Continuous measurement of agentic AI performance quality

Infrastructure for Trillion-Parameter Models

Computational resources - Adequate resources for trillion-parameter inference
Geographic distribution - Distributed infrastructure for global access
Redundancy - Multiple systems ensuring service continuity
Performance optimization - Continuous optimization of model serving infrastructure

10.2 Service Limitations and Dependencies

Technical Constraints

Computational limits - Inherent limitations of current computational capabilities
Network dependencies - Reliance on internet connectivity for cloud services
Model constraints - Limitations inherent in current agentic AI technology
Safety boundaries - Intentional limitations for safety and ethical reasons

Planned Maintenance and Updates

Safety updates - Regular updates to improve AI safety and alignment
Capability enhancements - Planned improvements to agentic capabilities
Infrastructure maintenance - Regular maintenance of computational infrastructure
Security patches - Prompt deployment of security updates

User Responsibilities and AI Ethics

11.1 Ethical Use of Agentic AI

Fundamental Ethical Principles

Beneficence - Using agentic AI to benefit humanity and society
Non-maleficence - Avoiding harm through autonomous AI operations
Autonomy - Respecting human autonomy and decision-making authority
Justice - Ensuring fair and equitable access to agentic AI benefits

Responsible Innovation

Precautionary principle - Exercising caution with powerful autonomous capabilities
Stakeholder consideration - Considering impacts on all affected stakeholders
Long-term thinking - Considering long-term implications of agentic AI deployment
Social responsibility - Taking responsibility for societal impacts

11.2 Professional Standards for Agentic AI

Technical Competence

Understanding limitations - Clear understanding of agentic AI capabilities and limitations
Safety knowledge - Knowledge of AI safety principles and practices
Ethical awareness - Understanding of ethical implications of autonomous AI
Continuous learning - Commitment to staying informed about agentic AI developments

Community Engagement

Knowledge sharing - Sharing knowledge and best practices with the community
Safety collaboration - Participating in collective AI safety efforts
Ethical dialogue - Engaging in discussions about AI ethics and governance
Mentorship - Supporting others in responsible agentic AI development

11.3 Reporting and Community Safety

Safety Reporting Mechanisms

Incident reporting - Prompt reporting of safety incidents or concerns
Bias reporting - Reporting observed biases in agentic AI behavior
Misuse reporting - Reporting observed misuse of agentic AI capabilities
Research sharing - Sharing safety-relevant research findings

Community Standards Enforcement

Peer review - Community review of agentic AI applications and research
Standards development - Participating in development of community standards
Education and outreach - Educating others about responsible agentic AI use
Collective governance - Participating in community governance processes

Limitation of Liability

12.1 Agentic AI Service Disclaimers

"As Is" Provision for Autonomous AI

Kimi K2 agentic capabilities are provided "as is" without warranties regarding:

Autonomous decision quality - No guarantee of optimal autonomous decisions
Error-free operation - Potential for errors in autonomous reasoning and actions
Predictable behavior - Autonomous behavior may vary based on context and learning
Complete safety - No absolute guarantee against all potential risks

Inherent AI Limitations

Training data limitations - Constraints based on training data and methodologies
Contextual understanding - Potential misunderstanding of complex contexts
Novel situations - Limitations in handling unprecedented situations
Bias potential - Possible biases inherited from training data or processes

12.2 Liability Limitations for Autonomous Operations

Excluded Damages for Agentic AI

To the maximum extent permitted by law, we exclude liability for:

Autonomous AI decisions - Decisions made autonomously by the AI system
Indirect consequences - Indirect results of autonomous AI operations
Business impacts - Business losses from autonomous AI actions or inactions
Third-party effects - Impacts on third parties from autonomous operations

Liability Caps for Different User Categories

Research users - Limited liability for academic and research applications
Commercial users - Liability limited to fees paid for commercial services
Enterprise users - Liability terms specified in enterprise agreements
Open source users - No liability for open source usage

12.3 User Assumption of Risk for Agentic AI

Acknowledged Risks

By using Kimi K2's agentic capabilities, you acknowledge risks including:

Autonomous AI limitations - Inherent limitations in current autonomous AI technology
Unpredictable behavior - Potential for unexpected autonomous AI behavior
Decision responsibility - Your responsibility for consequences of autonomous decisions
Emerging technology risks - Risks associated with cutting-edge AI technology

Risk Mitigation Responsibilities

Users are responsible for:

Appropriate oversight - Implementing appropriate human oversight for autonomous operations
Safety measures - Implementing safety measures for autonomous AI applications
Impact assessment - Assessing potential impacts before deploying autonomous AI
Professional consultation - Seeking expert advice for critical applications

Indemnification

13.1 User Indemnification for Agentic AI

Scope of Indemnification

You agree to indemnify MoonshotAI for claims arising from:

Autonomous AI misuse - Misuse of agentic AI capabilities beyond intended purposes
Safety violations - Failure to implement appropriate safety measures
Ethical violations - Use of agentic AI in violation of ethical principles
Legal violations - Use that violates applicable laws or regulations

Specific Agentic AI Scenarios

Unauthorized autonomous actions - Actions taken by AI beyond authorized scope
Harm from autonomous decisions - Harm resulting from autonomous AI decisions
Privacy violations - Privacy violations through autonomous data processing
Third-party rights - Infringement of third-party rights through autonomous operations

13.2 Defense and Settlement for AI-Related Claims

Specialized AI Legal Support

AI law expertise - Access to legal experts specializing in AI law
Technical expert witnesses - Technical experts for AI-related legal proceedings
Industry standards - Reference to emerging industry standards for agentic AI
Academic research - Incorporation of relevant academic research in defense

Termination and Suspension

14.1 Termination Rights for Agentic AI Services

User Termination Rights

Immediate termination - Right to immediately terminate agentic AI access
Data export - Right to export data before termination
Gradual transition - Support for gradual transition away from agentic AI services
Research preservation - Provisions for preserving ongoing research

MoonshotAI Termination Rights

We may terminate access for:

Safety violations - Violations of AI safety principles or guidelines
Misuse of autonomous capabilities - Inappropriate use of agentic AI features
Ethical violations - Use contrary to ethical AI principles
Legal requirements - Termination required by law or regulation

14.2 Suspension Procedures for AI Safety

Emergency Suspension

Immediate safety concerns - Immediate suspension for critical safety issues
Autonomous behavior concerns - Suspension for concerning autonomous AI behavior
Harm prevention - Suspension to prevent potential harm
Investigation period - Reasonable time for safety investigation

Progressive Safety Enforcement

Warning systems - Initial warnings for minor safety concerns
Capability restrictions - Temporary restrictions on specific agentic capabilities
Enhanced oversight - Required enhanced oversight for continued access
Educational requirements - Mandatory safety education for reinstatement

Dispute Resolution

15.1 Agentic AI Specific Dispute Resolution

Technical Dispute Resolution

Expert panels - Technical expert panels for agentic AI disputes
Peer review - Community peer review for technical disagreements
Scientific method - Resolution based on scientific evidence and methodology
Safety prioritization - Safety considerations in all dispute resolution

AI Ethics Mediation

Ethics boards - Specialized ethics boards for ethical disputes
Multi-stakeholder process - Inclusion of diverse stakeholders in resolution
Transparency - Transparent process for ethics-related disputes
Community standards - Reference to community ethics standards

15.2 International Arbitration for AI

Specialized AI Arbitration

AI law specialists - Arbitrators with expertise in AI law and technology
International standards - Reference to international AI standards and best practices
Multi-jurisdictional - Consideration of multiple jurisdictions and legal frameworks
Safety expertise - Inclusion of AI safety experts in arbitration panels

Changes to Terms

16.1 Evolution of Agentic AI Terms

Technology-Driven Updates

Capability evolution - Updates reflecting advancing agentic AI capabilities
Safety improvements - Changes incorporating new safety measures and protocols
Research findings - Updates based on latest research in agentic AI
Community feedback - Changes based on community input and experience

Regulatory Adaptation

AI governance - Updates to reflect evolving AI governance frameworks
International standards - Alignment with developing international AI standards
Best practices - Incorporation of emerging best practices in agentic AI
Safety regulations - Compliance with new safety regulations

16.2 Community Involvement in Terms Evolution

Participatory Development

Community consultation - Community input on major terms changes
Expert review - Review by AI ethics and safety experts
Stakeholder feedback - Input from diverse stakeholders affected by changes
Democratic process - Democratic decision-making for significant changes

Contact Information

17.1 MoonshotAI Contact Information

Primary Contacts for Agentic AI

General Support: support@moonshot.cn
AI Safety Team: safety@moonshot.cn
Research Collaboration: research@moonshot.cn
Ethics Inquiries: ethics@moonshot.cn

Specialized Support

Technical Support: technical@moonshot.cn
Developer Relations: developers@moonshot.cn
Academic Relations: academic@moonshot.cn
Enterprise Solutions: enterprise@moonshot.cn

17.2 AI Safety and Ethics Contacts

Safety and Ethics Leadership

Chief AI Safety Officer: safety-officer@moonshot.cn
Ethics Review Board: ethics-board@moonshot.cn
Safety Research Team: safety-research@moonshot.cn
Community Safety: community-safety@moonshot.cn

International Collaboration

Global Research: global-research@moonshot.cn
International Standards: standards@moonshot.cn
Regulatory Affairs: regulatory@moonshot.cn
Multi-stakeholder Initiatives: stakeholders@moonshot.cn

17.3 Legal and Compliance

Legal Department

Legal Counsel: legal@moonshot.cn
Compliance Officer: compliance@moonshot.cn
Data Protection: privacy@moonshot.cn
Intellectual Property: ip@moonshot.cn

Mailing Address

MoonshotAI Legal Department
Kimi K2 Agentic AI Division
Beijing, China

Acknowledgment of Agentic AI Understanding

By using Kimi K2, you acknowledge that you understand:

Agentic AI Nature - The autonomous and goal-directed nature of agentic artificial intelligence
Responsibility Sharing - The shared responsibility between users and developers for safe AI deployment
Technological Limitations - Current limitations of agentic AI technology and ongoing research
Ethical Obligations - Your ethical obligations in deploying autonomous AI systems
Safety Commitment - The paramount importance of safety in agentic AI development and deployment

Our Commitment to Responsible Agentic AI

MoonshotAI is committed to:

Safety-first development - Prioritizing safety in all agentic AI advancement
Transparent research - Open sharing of research and safety findings
Community collaboration - Working with the global community on AI safety
Ethical leadership - Leading by example in ethical AI development
Beneficial deployment - Ensuring agentic AI benefits all of humanity

Thank you for joining us in the responsible development and deployment of agentic artificial intelligence. Together, we can ensure that autonomous AI systems are developed and used safely, ethically, and for the benefit of all humanity.

Last Updated: January 15, 2025
Effective Date: January 15, 2025
Version: 1.0

Kimi K2 represents a new frontier in artificial intelligence - true agentic intelligence that can autonomously reason, plan, and act. With this unprecedented capability comes unprecedented responsibility. These terms reflect our commitment to responsible innovation and the safe, beneficial deployment of autonomous AI for the betterment of humanity.