Using AI to make processes more efficient is a win, but using it haphazardly can cause legal exposure. 

By Vanessa Matsis-McCready 

The use of AI in the workplace presents both opportunity and risk, and generative AI tools can improve efficiency and support better decision-making. But without clear guardrails and human oversight, the same tools can expose organizations to new legal and workplace risks. 

AI to Doing the Job of the Employee 

AI is not a substitute for experience or professional judgment. Recent headlines have highlighted attorneys sanctioned for submitting court briefs drafted by AI that cited nonexistent case law. The issue was not the technology itself. The problem was the absence of a professional review. Without verifying the AI-generated content, the attorneys submitted citations that did not exist. This type of shortcut could lead to discipline, ethical violations, or malpractice exposure. It also illustrates how relying on AI without human oversight can expose an employer to liability.  

Leaking Trade Secrets 

Confidential information can also be exposed when employees input sensitive company data into open AI platforms. Once proprietary information enters a public AI system, employers may lose control of how that data is used. Organizations can reduce this risk by: 

• establishing clear AI use policies; 

• training employees on responsible AI practices; and 

• providing secure internal AI tools. 

Employees will use AI. Providing a secure, internal option is non-negotiable. 

When AI Joins the Meeting, Privilege Can Go Out the Window 

AI meeting assistants and transcription tools access and store sensitive information on external servers. These records may later be accessed through cybersecurity breaches, subpoenas, or litigation discovery requests. This creates several risks, including: 

• sensitive discussions stored outside the employer’s control; 

• transcripts and recordings that may become discoverable in litigation; and 

• potential complications involving attorney-client privilege. 

Multistate employers face additional exposure. Employees who activate transcription tools may unintentionally violate state recording or surveillance laws. While AI transcription can help with note-taking, organizations should weigh that convenience against the legal and cybersecurity risks of storing sensitive conversations outside their control.  

Recruiting and Hiring 

AI makes data analysis easier and can identify training gaps, but it cannot interpret the human dynamics that shape workplace performance. It is not a replacement for an HR professional.  

For example, AI can screen resumes faster, but it can also miss excellent candidates. Algorithms may filter out qualified candidates whose experience does not match expected patterns and may repeat bias in past hiring data. Consequently, states like New York, regulate its use. AI can help craft interview questions or summarize candidate information, but it should not replace a recruiter’s ability to evaluate experience, ask follow-up questions, and assess candidate potential.  

Performance Management 

Managers who rely on AI analysis when calculating productivity or reviewing quantifiable work product forget that this is only one piece of the puzzle. Dozens of factors impact performance and AI misses these. It cannot identify the existence of a reasonable accommodation or understanding that people work differently, but a human can. 

Performance conversations are difficult but outsourcing them to AI is even worse. Savvy managers use AI to test messages for clarity or tone. A prompt of “what action items result from this feedback” can assist in creating a performance improvement plan. If the manager cannot identify the improvement needed, then the tool is useless. AI should not be the messenger. Using AI-generated messages to notify employees of layoffs, for example, harms morale among remaining staff and damages future recruiting efforts. It sends the message that the employee is expendable and not worth the direct interaction.  

Using AI to Plan Layoffs or Decide Who to Terminate 

There is a temptation to use AI as a “neutral” decision-maker for position elimination, but AI cannot solve for organizational forces. HR professionals know certain employees are the glue of the workplace. They may appear to be average producers, yet their value may lie in bringing out the best in colleagues, their reliability, or their ability to mentor others. AI misses this human reality and may expose companies to discrimination claims tied to automated decision-making. 

Employees Can See Right Through AI 

Employees are increasingly adept at recognizing AI-generated communication. Failing to disclose its use contributes to poor morale and costly disengagement of employees who feel misled. 

Today’s workforce values authenticity. Managers who hide behind AI to deliver important information risk a loss of credibility. AI-generated language often sounds formal and mismatched with a manager’s voice, making the disconnect obvious. And the morale impact can be immediate. Employees may assume the manager did not take the time to consider the situation or lacked the knowledge to address it directly. Over time, that perception erodes loyalty and engagement. Companies that invest in and develop their workforce, rather than treating employees as another “point of delivery,” distinguish themselves as employers of choice. 

As organizations adopt AI tools, the goal should be to use them to support the work of people, not replace the human judgement that employees value most.   

Vanessa Matsis-McCready is associate general counsel and vice president of HR services for Engage PEO.