Exploring GenAI with VAILL: Day 7 🛟
A 10-day journey into human-centered GenAI for legal professionals
🛟 Day 7: AI Safety & Ethics
Welcome back! Now that you're comfortable using GenAI tools for various aspects of your work, it's crucial to explore how to do so ethically and safely. Today, we'll focus on responsible GenAI usage that aligns with our professional obligations and protects client interests.
Today's Focus: Explore using GenAI tools ethically and safely while maintaining compliance with professional responsibilities and protecting client confidentiality.
🎯 Learning Objectives
Master essential safety protocols for GenAI usage in legal practice
Understand ethical considerations and professional obligations
Develop strategies for protecting client confidentiality when using GenAI tools
🔑 Key Concepts
As legal professionals, we have specific ethical obligations that must guide our use of AI tools. ABA Model Rule 1.6(c) requires lawyers to "make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client." This obligation extends to our use of AI tools.
Key considerations from ABA Model Rule 1.6 Comments:
Comment [18]: When using technology, lawyers must use "reasonable efforts" to prevent unauthorized access to client information
Comment [19]: Factors determining reasonableness include “the reasonableness of the lawyer's expectation of confidentiality include the sensitivity of the information and the extent to which the privacy of the communication is protected by law or by a confidentiality agreement”
Beyond confidentiality, we must also consider:
Competence (Rule 1.1, Comment [8]): Keeping abreast of changes in technology
Communication (Rule 1.4): Informing clients about use of AI tools when appropriate
Supervision (Rule 5.3): Ensuring AI tools are used in a manner compatible with professional obligations to supervise
💡 Pro Tips
Always remove identifying client information before using GenAI tools1
Maintain clear documentation of GenAI tool usage and verification processes
Regularly review and update GenAI safety protocols
Be transparent with clients about GenAI usage when appropriate
Verify all GenAI-generated content independently: check the AI’s work!
✍️ Today's Exercise
Let's practice implementing ethical GenAI usage through a structured approach:
Safety Protocol Development: Try this prompt: "Help me create a checklist for safely using AI tools in legal practice. Include:
Steps to protect client confidentiality
Data sanitization procedures
Verification requirements
Documentation practices
Please format this as a practical checklist I can use daily."
Client Communication Template: Follow up with: "Please help me draft a clear, professional paragraph I could include in client communications about AI tool usage. Focus on:
Transparency about AI use
Benefits to the client
Safety measures in place
Client consent considerations"
Risk Assessment: Finally, try: "Let's create a risk assessment framework for AI tool usage in legal practice. Please include:
Common risk scenarios
Prevention strategies
Response procedures
Documentation requirements"
Remember to spend some time in conversation with the AI about how it responds to your prompts. Ask questions, probe for greater detail, and give additional context to make the output relevant to your work and needs.
⚠️ Common Mistakes to Avoid
Sharing confidential (or potentially confidential) client information with GenAI tools
Failing to verify GenAI-generated content
Overlooking documentation requirements for GenAI usage
⚡️ Lightning Round (Optional)
Conduct a "privacy audit" on the AI tool you've been using throughout this course by:
Reading through its terms of service and privacy policy
Documenting key points about:
Data retention policies
Information sharing practices
User rights and controls
Data security measures
Creating a one-page summary that answers:
What can we safely share with this tool?
What should we never share?
What privacy controls should we enable?
What documentation should we maintain?
Testing the tool's built-in privacy features and sharing your findings
Compare what you discover to how you've been using the tool so far. Are there any practices you need to adjust based on what you learned? Share your insights with a colleague and discuss any implications for your professional use of GenAI.
🔄 Quick Recap
Professional obligations guide GenAI usage
Client protection is paramount
Documentation and verification are essential
Transparency around GenAI usage builds trust with clients
📔 Reflect
Before moving on, spend a few minutes in your learning journal noting:
Any challenges you encountered today and how you addressed them
Any guardrails you would need to see in place to feel fully confident using GenAI in your work
One specific way you'll apply today's learning tomorrow
Your reflections create valuable reference points for developing sustainable GenAI practices.
📚 Additional Resources
Article: ABA issues first ethics guidance on a lawyer’s use of AI tools (ABA website) - Introduction to ABA’s Advisory Opinion 512 on the use of GenAI in practice (link to the full opinion)
Audio/Interactive: The Legal Ethics of Generative AI - upload Suffolk Law School Dean Andrew Perlman’s article to GoogleLM; create an audio “podcast” of the article, then have a convo about it with GoogleLM, diving deeper into the content.
Interactive: Generative AI for Legal Professionals (Coursera) - Take a deeper dive into GenAI and prompting with VAILL’s founding co-director Mark Williams
👉 Coming Tomorrow
Get ready to advance your GenAI skills! Next, we'll explore more advanced prompting techniques to help you get even more from these tools while maintaining an ethical framework.
Keep experimenting safely, and share your discoveries with us at vaill@vanderbilt.edu!
While specialized GenAI tools purpose-built with security and safety protocols may support the safe inclusion of more sensitive data, when using off-the-shelf commercial tools like ChatGPT, Claude, etc., best practice is to omit any information that is even potentially confidential.