Written by Kellie Telis and reviewed by Aaron Rocke.

Artificial intelligence has rapidly become part of day-to-day business operations. Employees are using AI tools to draft emails, summarize documents, analyze data, and prepare reports. Employers and HR professionals are using them to make significant employment decisions, ask for legal advice, and resolve workplace conflicts. For many businesses, these tools offer clear advantages. They can improve efficiency, reduce unwanted administrative workloads, and expand a team’s scope. On the flip side, the disadvantages are fewer but carry risks that call for careful consideration. 

Courts, regulators, and legal organizations are starting to pay more attention to how AI is used in professional settings. Recent guidance from the American Bar Association, resources published by the Federal Judicial Center, and a growing number of court decisions all point to the same reality: businesses need to be intentional about how AI is used, particularly when confidential information, employment decisions, or legal issues are involved. 

For Washington employers, the issue is no longer whether AI will be used in the workplace. It already is. The more important question is whether they understand the risks at hand and how to put safeguards in place.

The Risks of Using AI in the Workplace 

AI platforms are increasingly being used to complete tasks that people lack the time or desire to do themselves. Employers often use these tools to create job descriptions, draft handbook language, organize internal communications, summarize meetings, or assist with recruiting and administrative work. While it does save time and decrease mental fatigue, AI-generated content is not always accurate. 

Fabricated citations, incomplete summaries, misleading information, and biased results are frequently produced. Courts across the country have already sanctioned attorneys who submitted legal documents containing fake AI-generated case citations that were never independently verified. Recently, Reuters reported on the ABA’s formal guidance, emphasizing that lawyers using generative AI still have duties regarding competence, confidentiality, and the independent review of its work product.  

The same concerns affect employers. If AI-generated information is used during hiring, disciplinary decisions, workplace investigations, or compliance-related matters, inaccurate results could expose a business to costly legal issues. It can be a useful support tool, but not a replacement for human judgment and independent fact-checking. 

Your AI Prompts Can Be Used Against You in Court  

One of the biggest legal concerns involving AI is confidentiality.  

Many public tools collect and store user inputs to improve their systems. Users may not realize that entering information into these systems could expose sensitive company data to a third party.   

This becomes especially risky when businesses are dealing with:  

• employee complaints,   

• internal investigations,   

• trade secrets,   

• customer information,   

• financial records,   

• contract disputes, or   

• pending litigation.   

All the information being given to the tool, including sensitive or confidential internal data such as what is listed above, may be analyzed and redistributed by the LLM. 

Some courts have found that what you write to and receive from AI is fair game to be discovered in litigation. Many would not consider this before starting a new chat where they provide sensitive information that may have legal ramifications. For example, consider the following prompt:   

An employee I terminated claims I owe them 45 missed meal breaks. I did not know I  needed to give them meal breaks. Do I really need to pay them? 

If your opponent’s counsel requests that you provide the conversation(s) with the tool you use, even the ones you had while in contract with a lawyer, you may need to turn them over because AI is not your lawyer; a court may find there is no attorney-client privilege.  

A recent decision in United States v. Bradley Heppner, discussed by Harvard Law Review, took this stance. In that case, Heppner had chatted with AI Claude, inputting vital information from his attorney, asking the AI tool to brainstorm legal defense strategies. He did this without instruction from his attorney. The court later determined that these communications with Claude were not privileged because the information had been shared with a third party.  

This is an easy mistake to make. Owners, HR managers, supervisors, or anyone involved in the dispute may consider using AI to summarize a workplace complaint, prepare investigation materials, draft communications, or discuss the legal issues at hand. Businesses should start strategizing now about how AI-generated content is created, reviewed, stored, and documented.  

A Proactive Approach for Employers 

A few warnings. First, understand that AI can give you unreliable answers. Sometimes, it hallucinates. Sometimes, it gives you what it thinks you want to hear, and sometimes it gives you the right answer under the wrong governing law. Additionally, you may be educating the model with your input. Unless you have a paid version and have vetted the terms of use, anything you type or upload may be educating the large language model with new facts. The AI may incorporate what you upload as facts that it relies upon to answer other users, including your adversary.    

Moreover, it isn’t legal advice, so it is likely discoverable. While the law is new and undeveloped on this point, here are some steps you can take to shrink the likelihood that a court will order you to turn over the information.   

1. Use a paid version and read the terms of service. Only use AI that will keep your information private. Your information should not educate the model in a way that others can discover.  

2. Refer to privilege, possibly in your prompt. The attorney-client privilege includes communications within the organization in order to get legal advice. If AI is a tool to obtain legal advice, such as what questions to ask a lawyer or explain in plain language what your lawyer said, it is less likely to be discoverable.   

3. Consult with your attorney, as this may be the most important factor in determining protection.  

We predict these issues will be decided on a case-by-case basis as the law in this area develops, so do not use AI unless you appreciate the risks that it is discoverable.

AI Disclosure in Legal Matters 

Over the past year, courts and regulators have started taking a much closer look at AI-generated content, particularly in litigation and legal proceedings.  

Federal courts are beginning to address whether parties should disclose when AI was used to generate or alter evidence. A proposed federal rule discussed by Meyers Nave would require disclosure when evidence, including documents, audio, or images, was created or materially altered using AI tools. 

AI Bias and Employment Decisions 

Another growing concern is bias in AI-assisted employment decisions. 

AI systems are trained on large amounts of existing data, and that data can reflect historical biases. If employers rely too heavily on AI-generated recommendations in recruiting, hiring, promotions, or discipline, they risk unintentionally creating discriminatory outcomes. 

Federal and state regulators are increasingly paying attention to this issue. Although AI-specific employment regulations are still evolving, existing anti-discrimination laws already apply to workplace decisions made with the assistance of AI tools. 

Employers should remember that using AI does not eliminate responsibility for employment decisions. Businesses remain responsible for ensuring that their practices comply with wage and hour laws, anti-discrimination laws, disability accommodation requirements, and privacy obligations. Human oversight remains essential. 

What Employers Should Be Doing Now 

Businesses do not need to avoid AI altogether. In many workplaces, it can be a valuable productivity tool when used responsibly. However, employers should implement clear guidelines before employees begin relying on AI systems in ways that affect decision-making or legal compliance. 

Create an AI Use Policy 

Employers should establish written policies explaining how employees may and may not use AI tools in the workplace. Policies should address confidentiality, data privacy, acceptable uses, and required human review. 

Limit the Use of Public AI Platforms for Confidential Information 

Employees should not input confidential business information, HR records, customer data, trade secrets, or legal strategy into public AI systems unless approved and properly secured. 

Require Human Review 

AI-generated content should always be reviewed for accuracy. Employers should avoid relying exclusively on AI-generated information for employment-related decisions or compliance matters. 

Train Managers and HR on AI Risk 

Supervisors and HR professionals should understand both the benefits and risks of AI tools, particularly when handling any situation that may carry legal risk. 

Consult Legal Counsel When Necessary 

As AI laws and regulations continue to evolve, employers should work with legal counsel when implementing AI-related policies or using AI tools in sensitive employment contexts. 

Final Thoughts 

AI Usage: Risks for Employers outlines how AI is changing the workplace rapidly, and many businesses are already using these tools without formal policies or safeguards in place. While it can improve efficiency and streamline operations, it also creates significant legal and confidentiality risks if used carelessly. 

Employers do not need to fear AI, rather, they should approach it thoughtfully. Businesses that establish clear expectations, maintain human oversight, and prioritize confidentiality and compliance will be in a far stronger position as AI continues to reshape the modern workplace. 

If your business has questions about policies, employee management, confidentiality concerns, or compliance issues related to AI use, the employment attorneys at Rocke Law Group can help.