AI in the Workplace: Legal Risks and Practical Realities

Link to The Current Landscape The Current Landscape

Generative artificial intelligence (“AI”) tools are now widely used across the workforce, and the use is dual track, meaning employees are using AI to perform work, while employers are deploying AI-enabled tools to manage and make decisions about their workforce. This creates a recurring tension between efficiency gains and the need for control, accuracy, and fairness, particularly because law and organizational governance often evolve more slowly than technology itself. A practical challenge is and will be that many employers do not fully understand how (or how often) employees are using AI tools in day-to-day work.

Link to Employees Using AI: Productivity Upside, Real Legal Risk Employees Using AI: Productivity Upside, Real Legal Risk

Employees are using AI for drafting, coding, analytics, and problem-solving, often with meaningful productivity benefits. At the same time, the most common and consequential risks tend to arise from unintentional misuse rather than bad intent.

Key employee-side risk areas include:

1. Confidentiality and proprietary data exposure (for example, pasting client or company data into a public AI tool);
2. Inaccurate or fabricated outputs (hallucinations) that can undermine work quality and decision-making;
3. Intellectual property (“IP”) ownership and authorship concerns, particularly where AI is used to generate work product;
4. Shadow AI use, i.e., employees adopting tools without approval, oversight, or security review; and
5. Employee discipline or termination risk if AI use violates policy, confidentiality obligations, or performance expectations.

Link to Employer Use of AI: Employment Decisions Employer Use of AI: Employment Decisions

On the employer side, AI is increasingly embedded in core employment functions, including: (1) hiring (resume screening and candidate ranking); (2) productivity monitoring and work assignment; (3) performance evaluation; and (4) discipline and termination decisions.

Because these tools can make or inform decisions that affect applicants and employees, employers should assume that AI-enabled workflows will be scrutinized like any other selection or decision process, especially when the organization cannot clearly explain how the tool works, what data it uses, and how outputs are reviewed.

Link to Key Legal Risk Areas Key Legal Risk Areas

AI presents several recurring legal risk categories, including:

1. Discrimination and bias risk under major employment statutes (including Title VII, the Americans with Disabilities Act, the Family and Medical Leave Act (“FMLA”), and the Age Discrimination in Employment Act.
2. Lack of transparency between employer and employee in how decisions are made or reviewed.
3. Failure to accommodate disabilities or approval of FMLA leave in automated systems or AI-mediated processes.
4. Continued development of state and local AI regulation trends, including notice and audit concepts, which may be applicable to employers.
5. From the employee perspective, common concerns include lack of due process, limited ability to challenge automated decisions, and privacy and surveillance issues tied to monitoring tools.

Link to AI-Driven Litigation and Enforcement AI-Driven Litigation and Enforcement

AI presents several recurring and emerging claim types and Enforcement themes, including:

1. Disparate impact allegations tied to AI hiring tools.
2. Failure-to-accommodate theories where automated systems create barriers for individuals with disabilities.
3. Wage and hour issues are connected to AI-based tracking and monitoring.
4. Wrongful termination disputes where decisions are based on flawed AI outputs or overreliance on automated scoring.

U.S. state and federal regulators are also paying attention. Several agencies are developing and implementing focused areas of AI regulation, including the Equal Employment Opportunity Commission on AI bias, the Federal Trade Commission on deceptive AI practices, and the Department of Labor on worker monitoring and classification issues. Employers should expect these agencies to issue regulations on the use of AI in the workplace. In litigation, the trend is toward scrutiny of automated decisions, discovery disputes over algorithms and decision logic, and claims that target both the employer and the vendors/customers who use AI.

Link to AI Governance: What Employers Should Be Doing Now AI Governance: What Employers Should Be Doing Now

A consistent theme is that AI use is already happening—whether or not an organization has formally approved it, so governance of AI by the employer should be proactive rather than reactive.

Practical steps and recommendations for employers:

1. Conduct AI risk assessments across both internal use cases and third-party tools.
2. Implement human oversight for AI-informed employment decisions.
3. Audit tools regularly to identify and address risk patterns early.
4. Train human resources professionals and managers on appropriate use, limitations, and escalation paths.
5. Policy and contracting updates are also central as employers should adopt formal AI usage policies, restrict unauthorized tools and shadow AI, implement disclosure requirements where appropriate, watch for and include AI language in vendor or other contracts.
6. Update employment agreements and expand confidentiality provisions to address AI-related data handling, and also add AI-specific IP assignment language and negotiate indemnification protections where feasible.

Link to Key Takeaways Key Takeaways

1. AI use in the workplace is already widespread, and organizations should assume it is occurring even where it is not formally sanctioned.
2. Employee AI use creates meaningful confidentiality, data security, and IP risks that are often unintentional.
3. Employer AI usage may create bias and liability risk, particularly when AI tools influence hiring, performance, discipline, or termination decisions.
4. Regulators are actively focused on workplace AI, and litigation trends show increasing scrutiny of automated decision-making and customer/vendor involvement.
5. Governance, policies, training, human oversight, and vendor/customer contracting are critical to managing risk while preserving the productivity benefits of AI.