Creating a workplace policy on artificial intelligence requires human intelligence. With or without employer guidance, workers are rapidly incorporating AI tools (such as ChatGPT, Gemini, Claude, and CoPilot) into their work. Organizations should create AI policies based on their unique business needs and considerations for how employees may use AI in the workplace. This is an area in which a one-size-fits-all approach will not serve your business well.
Organizations should begin by considering how AI tools may be used in the workplace and which risks need to be managed. Some familiarity with AI tools by those crafting the policy will help ensure a realistic approach. Some questions to consider include:
- Will AI notetakers be allowed in virtual meetings? Some virtual meeting platforms have begun to include virtual notetakers by default. Some employees actively seek out AI notetaker programs to create meeting summaries. Consider whether this is acceptable for your team’s culture and whether it will be allowed for internal meetings and/or external meetings. Generally, if an AI notetaker will be used, employers should require that its “presence” be disclosed. Employers must consider how they will handle meetings in which some participants want an AI notetaker and others do not. For meetings involving sensitive matters, such as legal strategy, AI notetakers should be avoided, as they may create records disclosable to other parties in litigation or, in the case of public employers, under the Freedom of Information Act.
- Will AI be used in the hiring process? While it may seem that reducing human involvement will reduce human bias, studies have found that AI reproduces human biases, particularly when existing company demographics are used to train the AI model. Consider allowing HR personnel to use only pre-approved AI technology, carefully selected to avoid discriminatory outcomes.
- How will the organization develop and protect its intellectual property and confidential information? Courts have held that works created by AI without significant human involvement cannot be copyrighted. Additionally, providing trade secrets or other confidential information to AI platforms may result in a loss of protection. Organizations with a high value on confidentiality, particularly healthcare employers, should consider limiting employee use of AI to proprietary platforms or platforms designed to comply with HIPAA or other applicable privacy laws.
- How will the organization avoid disclosing sensitive legal information? Individuals and organizations must realize that generative AI platforms are not attorneys. While many have turned to such platforms for legal advice, they are often inaccurate or lacking in key context. Importantly, there is generally no attorney-client privilege when asking AI for legal advice or to prepare documents that may ultimately be given to an attorney, but that have not been drafted at the attorney’s request. This means that such inquiries may create records disclosable to other parties in litigation or, in the case of public employers, under the Freedom of Information Act. Attorneys are beginning to seek such records in litigation, and these communications may be subject to record retention requirements.
- How will the organization ensure that the use of AI tools does not lead to poor quality work product? AI tools can be extremely helpful for creative work or for summarizing information. For other types of work, AI may lead to unacceptably poor work product, such as incorrect calculations, “hallucinated” facts, or even client communications with an off-puttingly unnatural tone. Employees must ultimately be held accountable for their work product, whether or not AI tools are allowed to assist.
As you identify your organization’s goals with respect to AI, it is important to clearly communicate expectations to your employees. The labor and employment attorneys at Berchem Moses PC can assist with creating an effective AI policy for your workplace.
Share this post
