Simplelaw Legal Case Management Software | Library

Overview of ABA Formal Opinion on Generative Artificial Intelligence

Written by SimpleLaw | 8/1/24 3:15 PM

The American Bar Association's (ABA) Formal Opinion 512, issued on July 29, 2024, provides crucial guidance for lawyers and law firms using Generative AI (GAI) tools. As legal professionals increasingly rely on technology, understanding and complying with these ethical obligations is critical. Here’s what you need to know to ensure compliance and maintain professional standards.

Maintaining Competence with Evolving Generative AI

Lawyers have a duty to stay current with technological advancements, including GAI tools. This means understanding how these tools function and their limitations. Regular training and education are essential to maintain this competence. Taking the time to understand the process, data sources, and any related inherit bias is critical. 

Lawyers should understand or consult experts about the benefits and risks of GAI tools. They must stay informed as these tools evolve rapidly. While GAI tools can improve efficiency and service quality, they also pose risks, such as creating inaccurate or biased content due to issues in the data used to create the content. Lawyers should critically assess GAI outputs, as reliance on flawed AI-generated content can lead to bad legal advice or misleading courts. It's critical to watch out for hallucinations. Hallucinations are false or misleading information created by GAI. It's essential to verify and contextualize AI-generated information.

There are many online options available to understand GAI, what it is, how it works, and how it can be most safely used. Depending on how law firms use GAI, providers like Google's Gemini, Microsoft's CoPilot, OpenAI's ChatGPT, and many others offer on demand overviews to help non-tech people understand their products. If you are looking for a training course, Coursera has many options available. 

Confidentiality and Data Review

Confidentiality is a cornerstone of legal ethics. When using GAI tools, lawyers must ensure that client information is secure and protected from unauthorized access. This includes vetting GAI service providers for robust security measures and compliance with data protection standards. Firms should establish clear policies for handling sensitive information within GAI platforms.

When using Generative AI (GAI) tools, attorneys must independently verify their accuracy, especially for critical tasks. The level of review depends on the tool's use and prior experience with it. Attorneys cannot rely solely on GAI for professional judgment tasks, such as legal advice. Additionally, they must protect client confidentiality under Model Rule 1.6, ensuring sensitive information is not improperly disclosed or accessed. This includes understanding the terms and policies of GAI tools and obtaining informed client consent if necessary.

Transparent Communication with Clients

Model Rule 1.4 requires lawyers to communicate with clients about the methods used in their representation, including the use of GAI tools. This transparency allows clients to make informed decisions. Lawyers should explain how these tools are used, the benefits, and any potential risks involved. Documenting this communication, preferably in writing, helps prevent misunderstandings. Even when not explicitly required by Rule 1.6 or 1.4, informing clients about GAI use can improve transparency and trust. It's essential to consider each case's specifics, the client's expectations, and the sensitivity of the information involved.

Text Generation Accuracy 

Under Rules 3.1, 3.3, and 8.4(c), lawyers must ensure that all information presented to courts is accurate and truthful. This includes verifying the output generated by GAI tools. Misleading or false information, even if generated unintentionally, can lead to serious ethical violations. Lawyers should diligently review and cross-check all AI-assisted outputs before submission.

Supervisory Responsibilities

For firms using GAI tools, there is an obligation to supervise both lawyers and non-lawyers involved in the process. This includes ensuring proper training and adherence to ethical standards. Supervisory lawyers must also ensure that third-party GAI service providers comply with professional obligations, such as maintaining confidentiality and providing accurate information.

Supervisory lawyers must establish policies and train staff on GAI tools, ensuring adherence to ethical standards. In addition, lawyers must evaluate the security and confidentiality measures of GAI tools, especially when outsourcing. Proper oversight and adherence to professional standards are crucial to avoid ethical breaches.

Reasonable Fees and Billing Practices

ABA Formal Opinion 512 also addresses fee-related ethics. Lawyers must charge reasonable fees for services, including those involving GAI tools. It is essential to bill clients only for the actual time spent and the reasonable cost of using these tools. Clear communication regarding these costs, preferably at the outset of the engagement, helps avoid disputes and maintains transparency.

When using Generative AI (GAI) tools, lawyers must ensure that fees and expenses charged to clients are reasonable, as per Model Rule 1.5. Lawyers should explain the basis for charges, ideally in writing, before or soon after starting representation. Time billed must reflect actual time spent, including for GAI use. Overhead costs like a fee to access a GAI tool should not be billed to clients. Lawyers cannot charge for learning how to use GAI tools, but they may charge if a client requests the use of a specific GAI tool.

Generative Artificial Intelligence Summary

Compliance with ABA Formal Opinion 512 is not just a legal obligation but a commitment to ethical practice. By understanding and adhering to these guidelines, lawyers and law firms can leverage GAI tools effectively while upholding the highest professional standards. Stay informed, communicate transparently, and prioritize client protection to navigate the evolving landscape of legal technology. And remember, the devil is in the details. The importance of review and verification of generated information is key. From our perspective, GAI models are not yet reliable enough to include in case management software. However, as with all things tech, we actively monitor ongoing developments and when the reliability increases, we will introduce it.