Assist Mi Legal Responsible AI & AI Governance Statement

# Assist Mi Legal Responsible AI & AI Governance Statement Version 1.0 Effective Date: August 1, 2026 ## Our Philosophy Assist Mi Legal was built on a simple principle: Artificial intelligence should assist legal professionals, not replace them. Our goal is to reduce administrative burden, improve organization, and help attorneys spend more time practicing law. We believe that legal judgment, client advocacy, professional responsibility, and ethical decision-making remain human responsibilities. Artificial intelligence can help attorneys work more efficiently, but it should not replace professional judgment. ## Our AI Principles Assist Mi Legal operates according to the following principles. ## 1. Human Oversight AI should support human decision-making. Important legal decisions should remain under the control of legal professionals. Assist Mi Legal is designed so that attorneys and authorized users remain responsible for: Legal advice Client communications Court filings Legal analysis Litigation strategy Ethical obligations AI-generated content should be reviewed before reliance or distribution. ## 2. Attorney Responsibility The use of AI does not transfer professional responsibility. Attorneys remain responsible for: Compliance with professional rules Client representation Confidentiality obligations Professional judgment Supervising work product Assist Mi Legal is a technology platform and not a law firm. The platform does not provide legal advice. ## 3. Transparency We believe users should understand when AI is involved. Assist Mi Legal strives to make clear when: Content is AI-generated Recommendations are AI-generated Drafts are AI-generated Automated processing has occurred Users should not be surprised by AI activity occurring within the platform. ## 4. Human Review AI-generated content should be treated as draft material. Users should independently review: Draft communications Summaries Recommendations Legal authorities Research results Workflow outcomes AI output may contain errors and should be validated before use. ## 5. Responsible Automation Automation should increase productivity while preserving human control. Assist Mi Legal may automate administrative activities such as: Matter organization Task creation Email categorization Document organization Calendar support Workflow assistance Organizations remain responsible for determining which automation features are appropriate for their operations. ## 6. Privacy and Confidentiality Protecting customer information is a core design objective. Assist Mi Legal is designed to support legal professionals working with confidential information. We strive to: Limit access to authorized users Segregate tenant data Apply security controls Support confidentiality obligations Customers remain responsible for evaluating the suitability of the platform for their legal, ethical, regulatory, and contractual obligations. ## 7. AI Provider Governance Assist Mi Legal may utilize third-party AI providers to deliver AIpowered functionality. When evaluating providers, we consider factors such as: Reliability Security Privacy practices Performance Operational maturity Provider relationships may evolve as technology advances. ## 8. Data Minimization We seek to process only the information necessary to provide requested functionality. Examples include: Draft generation Summarization Categorization Workflow execution We strive to avoid unnecessary processing whenever reasonably possible. ## 9. Security AI functionality operates within our broader security program. Security measures may include: Authentication controls Access controls Encryption Monitoring Audit logging Tenant isolation Security practices continue to evolve as threats and technologies change. ## 10. Accuracy and Hallucinations AI systems can produce incorrect information. This may include: Incorrect facts Incorrect citations Incomplete summaries Misunderstood instructions Fabricated information Assist Mi Legal does not guarantee the accuracy of AI-generated output. Independent verification remains essential. ## 11. No Legal Advice Assist Mi Legal does not provide legal advice. Ava is not a lawyer. AI-generated content should not be interpreted as legal advice. Users remain responsible for determining whether information is accurate, appropriate, and legally sufficient. ## 12. Attorney-Client Privilege Questions involving privilege, confidentiality, waiver, and professional responsibility are legal determinations. Assist Mi Legal does not determine: Whether privilege exists Whether privilege has been waived Whether disclosure is appropriate Whether a communication is protected Users remain responsible for making these determinations. ## 13. Continuous Improvement We continuously evaluate: AI performance Safety mechanisms User feedback Security controls Workflow outcomes Feedback from customers plays an important role in improving the platform. ## 14. Customer Control Customers control: Which features are enabled Which integrations are connected Which users receive access Which workflows are automated Which permissions are granted Organizations remain in control of their deployment. ## 15. Responsible Innovation We believe innovation and responsibility should advance together. As Assist Mi Legal evolves, we intend to continue developing AI capabilities that: Respect professional obligations Support human oversight Improve productivity Preserve user control Promote transparency ## Our Commitment Assist Mi Legal was created to help legal professionals manage work more effectively, not to replace lawyers. We believe the future of legal technology is not attorney versus AI. It is attorneys empowered by well-designed AI tools operating under human supervision, professional judgment, and ethical responsibility. That principle guides how we build Assist Mi Legal today and how we intend to build it in the future.