Following the introduction of ChatGPT, the first widely available free artificial intelligence (AI) interface, use of generative AI in and out of the workplace is booming. Given the wide-reaching legal implications of AI use in the workplace, now is the time for employers to consider implementing an AI usage policy.
What is Generative AI?
Broadly, AI is any machine-based system that can provide recommendations, make decisions, or give predictions, based on imitation of human intelligence and data collection. AI is commonly used in the workplace for services such as spam detection or translation services. Generative AI is technology capable of creating new content based on an input or prompt, including written, visual, or auditory materials. This includes ChatGPT, which is most widely used by the public today. Generative AI, such as ChatGPT, learns and pulls from information not just from its original data set, but also from information provided by users. Of note, not all AI services are free and available to the public.
85% of American workers surveyed indicated that they use AI to help complete tasks at work. With ChatGPT being the most well-known and easily accessible, unless an employer pays for an AI service from a third party, it is likely that employees are using a free source such as ChatGPT when in the workplace, which lacks security and privacy safeguards. Employees may use generative AI, perhaps without the knowledge of their supervisors, for a variety of activities, including as a search engine, to create images, and to write emails, articles, evaluations and more. Generative AI can also summarize patterns in data. However, 69% of employees using AI report feeling uncertain or fearful about telling their supervisor when they use AI.
Are There Potential Legal Risks if Employees use Generative AI at Work?
Currently, there are no major federal laws or regulations specifically addressing AI usage. However, existing laws and federal guidance, as well as emerging lawsuits and administrative challenges, illustrate how use of generative AI in the workplace can lead to risks for employers. For example, employers could face data privacy violation claims by inputting employee, job applicant, or customer personal data into an AI service. AI utilizes, shares, and stores large volumes of personal data, so if an employee inputs such data, they may violate data privacy laws or confidentiality agreements. Because free, publicly available, generative AI such as ChatGPT stores data that it receives from users, employers may be liable for employee input of proprietary information.
Using generative AI to form work product can also lead to potential copyright violations. Generative AI poses significant concerns if employees accidentally engage in plagiarism from content produced through ChatGPT. Currently, the Supreme Court and federal agencies find that AI-generated content cannot receive a patent, nor can it receive copyright protection. Many artists have filed lawsuits against AI art-generating services because AI, which is trained on artwork and images from others, returns content that matches the artists’ style or work without providing them credit or compensation. It is unclear who owns content created by generative AI, whether AI infringes on copyrights by using data from other artists or writers to form its own content, and who is liable when these issues arise.
Finally, using generative AI to aid in hiring decisions may subject employers to risk of discrimination claims, as AI, like humans, can make decisions based on implicit bias. If the AI service is trained with a set of data that leads to inherent biases, the AI itself will provide biased suggestions and decisions. For example, a tutoring company that used AI-powered hiring tools to make selection decisions recently paid $365,000 in a voluntary settlement with the EEOC for age and gender discrimination, because the AI software screened out male applicants over 60 and female applicants over 55. Any employer who uses AI for employment decisions is liable for resulting adverse employment actions. Companies may not avoid responsibility by blaming AI as the decisionmaker.
Before writing off generative AI in the workplace, employers should examine its potential advantages. Encouraging AI use within certain boundaries will likely benefit businesses, as on average, generative AI has been shown to improve employee productivity by 66%. While the risks may seem daunting, businesses can prepare themselves to leverage AI wisely and safely by implementing an AI usage policy for employees.
What Type of Policy do I Need to Inform Employees and Protect my Business?
Similar to an information technology policy or an anti-discrimination policy, employers can begin implementing AI usage policies to address the rise in generative AI use in the workplace. Employers may add this AI policy to their existing employee handbook. For new employees, employers should review the AI policy and discuss the benefits and risks of generative AI usage, which AI tools are acceptable to use, and for which job functions AI usage is permitted. Clear expectations about how generative AI fits into the workplace can aid the employee in making informed decisions. Statistical data shows that employees may be receptive to such guidance, with 85% of employees globally desiring training on how to implement AI into their job performance.
When crafting an AI policy, the employer must determine if it wants to implement an open use, limited use, or prohibited use policy. The thoroughness of the policy depends on the type of business, the employee’s responsibilities, and the level of confidentiality of potential data input involved. Depending on the field of business and the sensitivity of the data, it may make sense for one employer to implement a prohibited AI usage policy, while another organization might benefit from a limited use policy that outlines which AI services employees are permitted to use and which require supervisor approval prior to use.
Whether an employer views AI as an integral tool for future operations or wants to proactively limit employee use of ChatGPT and similar services, it is wise to implement a usage policy now. In the quickly changing AI landscape, addressing usage in the workplace is a tool business owners should embrace to mitigate potential risk.
For more information and assistance in drafting an AI policy to fit the needs of your business, please contact an attorney in our Business Planning & Transactions Practice Group here.