Labor & Employment Law Update

Many employers have implemented AI tools without fully considering that the resulting chat histories may be discoverable in litigation. Like emails, text messages, Slack communications, and other electronically stored information (ESI), generative AI chat histories are increasingly relevant to litigation. Though often viewed by employees as informal brainstorming, employee prompts, AI-generated responses, uploaded documents, and usage logs may later be scrutinized in relevant employment disputes, commercial disputes, and investigations.

As employees increasingly rely on AI to assist with workplace decisions and business operations, organizations should evaluate whether their existing records management, preservation, and litigation hold practices adequately address AI chat histories.

AI Discovery Risks for Employers: Confidentiality and Privilege

For employers, the risk is not simply that AI chat histories may be produced in litigation. When employees treat a chatbot like a private sounding board, they may be sharing sensitive company data with a third-party vendor. Risks include loss of intellectual property, HIPAA violations, vendor training exploitation, data breach and regulatory fines, and breach of contract with clients.

From a litigation perspective, employers should also consider how AI-generated content may affect confidentiality and privilege. Public AI tools are not substitutes for legal counsel, and communications with commercial AI platforms generally do not receive the same protections as those between attorneys and clients. As a result, employee prompts, AI-generated responses, uploaded documents, and usage logs may become relevant evidence in litigation and could provide insight into the employer's decision-making process, the information considered, and the reasons underlying employment actions. These interactions may be used to challenge the employer's stated rationale, analyze the timing of events, or examine the employee's intent and thought process.

Consider a manager who asks an AI chatbot to document performance issues before terminating an employee or an HR professional who uses AI to draft a termination letter or summarize an internal investigation. If litigation follows, those prompts, outputs, and related records may become relevant discovery material. Opposing counsel may seek to use them to evaluate what information was considered, whether AI-generated content influenced the decision, and whether the employer's current explanation is consistent with contemporaneous records.

Whether privilege is waived depends on the circumstances, including the purpose of the communication, the platform’s terms, and applicable law. Similarly, contractual privacy protections, data retention practices, and restrictions on model training vary among AI providers. As a result, the scope of potentially discoverable AI-related information can differ significantly depending on the platform used and the nature of the data involved. Questions regarding the handling of proprietary business information, trade secrets, client data, and sensitive personnel records may further complicate discovery disputes and increase scrutiny of how AI tools were used in the workplace.

Steps Employers Should Take to Reduce AI-Related Legal Risks

Businesses should take proactive steps to manage AI-related information and reduce legal exposure:

1. Audit AI use. Identify which AI tools employees use, including enterprise-approved platforms and unauthorized consumer accounts. Evaluate each platform’s data-use terms, security features, access controls, and retention practices.
2. Adopt a clear AI policy. Specify which tools employees may use and what information they may enter. Prohibit the use of unapproved platforms for confidential business information, sensitive personnel matters, internal investigations, and legal issues. Establish training and approval procedures for higher-risk uses.
3. Update litigation holds. When litigation is reasonably anticipated, assess whether relevant AI prompts, outputs, uploaded files, and activity logs must be preserved. Include all applicable AI systems in litigation preservation instructions and address automatic deletion where necessary.
4. Coordinate with IT and legal. Work with technology teams and vendors to understand what AI data is retained, who can access it, how it can be collected, and whether it can be produced in litigation.