A group of GOP Harvard alumni have written a letter to President Gay of Harvard University, expressing their dissatisfaction with the response to the PSC statement. In the letter, they describe the rollout of Harvard University Information Technology's artificial intelligence "sandbox" tool, which began on Sept. 4. The tool aims to provide a secure environment for Harvard affiliates to access large language models. According to a press release from HUIT, the AI sandbox ensures that data entered into the interface is only visible to the user and is not shared with LLM vendors for training data. The tool was developed in collaboration with various departments across the University. The pilot phase of the project is being rolled out gradually, with instructors within the Faculty of Arts and Sciences gaining access first.

However, affiliates still need to request permission for each specific use case. Harvard Business School professor Mitchell B. Weiss '99 has praised the AI sandbox, describing it as a valuable educational resource for his course on public entrepreneurship. He believes that integrating generative AI models into the curriculum can help solve public problems. Feedback from the pilot phase will inform Harvard's strategy on AI, and negotiations with vendors for expanding the variety of consumer AI tools are ongoing. The Office of Undergraduate Education has also provided guidelines for faculty on the use of generative AI in FAS courses, with varying levels of encouragement. Weiss's students have reacted positively to the pilot version of the AI sandbox, recognizing the transformative potential of these tools.