#  Guidance on AI meeting assistants; data privacy principles 

 





February 11, 2025

 

 

Dear Members of the Harvard Community,

Since the rapid emergence of generative artificial intelligence (AI) tools, members of our community have embraced experimentation in exciting ways, examples of which you can see on [Harvard’s generative AI website](https://www.harvard.edu/ai/). As these tools evolve, the University is continuously assessing their use to support responsible experimentation while protecting the privacy and security of Harvard’s data.

As a reminder, before using generative AI in your Harvard work, please consult the [University-wide guidelines](https://huit.harvard.edu/ai/guidelines), which are updated periodically based on community feedback and technological developments. Some Schools, departments, and courses have also issued [additional guidelines or policies](https://www.harvard.edu/ai/learning-resources/).

We write today to share additional guidance on the use of AI meeting assistants and to share best practices on data privacy more broadly.

### **New guidance on AI meeting assistants**

AI meeting assistants or “bots”—either built into existing collaboration tools such as Zoom or added by individual meeting participants using third-party software—can transcribe and summarize online meetings or enhance meeting accessibility. However, they can also pose substantial privacy, regulatory, and legal risks, and have the potential to stifle conversation and open inquiry.

Accordingly, the University has issued [new guidance on AI assistants](https://huit.harvard.edu/ai-assistant-guidelines). **AI meeting assistants should not be used in Harvard meetings, with the exception of approved tools with contractual protections as outlined in these guidelines.** We will continue to evaluate their use in our environment.

### **Data privacy principles**

Data privacy considerations —including the collection and use of personal information—apply across the breadth of Harvard’s activities and are not limited to the use of AI. To assist our community in understanding and managing data privacy, HUIT has collaborated with faculty and administrators to develop University-wide [**Privacy Principles**](https://privsec.harvard.edu/privacy-principles) and online training.

These principles are not new policy but rather aspirational values, such as data minimization, transparency, and stewardship. We invite you to familiarize yourself with these principles and [contact HUIT](https://huit.harvard.edu/get-help) or [your School’s Information Security and Data Privacy Officer](https://harvard.service-now.com/ithelp?id=kb_article&sys_id=00fbde4993690e904cf93f9a7bba10f9) with any questions.

We thank you for continuing to experiment responsibly with generative AI, and for keeping Harvard’s data private and secure.

Sincerely,

Klara Jelinkova

Vice President and University Chief Information Officer

Michael Tran Duff

University Chief Information Security and Data Privacy Officer