Should my business have a ChatGPT policy?

According to a recent Time magazine article, “Almost 30% of the nearly 4,500 professionals surveyed…said that they’ve already used OpenAI’s ChatGPT or another artificial intelligence program in their work.”¹ Additionally, Business Insider reports, “Of these [30% of] people, 68% hadn’t told their bosses that they were using these tools for work.”²

These percentages will only grow as more and more businesses embrace generative artificial intelligence (“AI“) models like ChatGPT. These models have the ability to generate human-like responses, which aid in customer support, content creation, and other communication-oriented tasks.

However, as the saying goes, “With great power comes great responsibility.” It is crucial for businesses to establish an Acceptable Use Policy (“AUP”) for AI use in the workplace to ensure its responsible and ethical use within their operations.

Guidelines for crafting an effective AI Acceptable Use Policy

Here are some points that your AUP should address:

• Ethical use and professional responsibility. AI can generate “hallucinations,” which are confident responses that have no basis in fact. For example, consider the recent New York lawyer who used ChatGPT to write his court documents. Unfortunately, the AI invented and used the factitious case law in the lawyer’s court documents.³ Therefore, your AUP should make clear that AI and their generated content must not replace human decision-making.
• Privacy and confidentiality concerns. It’s important for your AUP to educate users on how generative AI systems work in order to protect your confidential and personal information. Generally, these AI systems learn by analyzing text and detecting its patterns and structures. To the extent confidential or personal information is included in such text, the AI system may use and retain that information to learn, and the AI system could disclose that information in its generated content to others. Such “use”, “retention,” and “disclosure” may not have been consented to by your customers, partners, or other relevant stakeholders.
• Intellectual property rights. As you may know, copyright protection does not extend to AI generated content unless a human exercises sufficient skill and judgment over the end result (FYI, changing the prompts and asking the AI to modify its generated content is likely not sufficient). Therefore, your AUP should instruct users to only use the AI as a preliminary tool in crafting their first draft.
• Transparency and explainability. If your business is using AI to make decisions about others, depending on the circumstances, your AUP should address the potential need of informing the affected individuals of the use of the AI system. Your business should also be prepared to explain the underlying logic, assumptions, and limitations of the AI system to those affected individuals.

In summary, businesses employing tools like ChatGPT or other generative artificial intelligence should have an AUP. If you have questions about or need any assistance drafting an AUP, please reach out to any member of Siskinds’ Technology Law Group for advice and direction. You can also reach out to the author, Savvas Daginis, a Canadian and American Business Lawyer focusing on technology transactions and privacy, at [email protected].

¹ https://time.com/6248707/survey-chatgpt-ai-use-at-work/

² https://www.businessinsider.com/70-of-people-using-chatgpt-at-work-havent-told-bosses-2023-3

³ https://news.bloomberglaw.com/business-and-practice/lawyers-ai-blunder-shows-perils-of-chatgpt-in-early-days