Emerging Technology and Innovation Program (ETIP)

Emerging Technology and Innovation Program (ETIP)

Introduction

The Emerging Technology and Innovation Program (ETIP) was developed to foster a culture of curiosity and experimentation by identifying, testing, and assessing emerging and innovative technologies at Harvard. ETIP is an inclusive approach to innovation, soliciting ideas directly from members of the HUIT community.

Recent pilots

Talk with HOLLIS

  • Implement a generative AI chat interface in the HOLLIS catalog, allowing users to search and locate books using natural language queries. 
  • Enhance user experience with efficient searches and potentially set a new standard across the University. 

AI for Financial Aid Matching

  • Automate student-to-fund matching using AI to convert a month-long manual process into an overnight task, eliminating bottlenecks in the FAO's business cycle. 
  • Save significant time and resources for FAO during peak periods, reducing the need for additional staff and setting a precedent for AI automation in other systems. 

PWRngr

  • Implement a custom-developed device to optimize personnel dispatch and improve response times for power-related events in Harvard's network utility rooms and data centers. 
  • Improve service reliability and cost reduction through real-time monitoring and instant alerts, with potential scalability across the University's critical infrastructures. 

Tuning LLMs with Course Content

  • Explore methods for tuning AI models by building a TutorBot and using course materials from the ‘Physics 143A: Quantum Mechanics I’ and ‘Expos 20: To What Problem Is ChatGPT the Solution?’ courses as test cases. 
  • Enable instructors to use this tool with minimal technical involvement and allow them to upload course content and review interactions between users and the Bot. 

Standardizing documentation using AI

  • Utilize Generative AI to format ServiceNow articles according to established Knowledge Base style guide and accessibility standards. 
  • Potential expansion of this formatting approach to other style guides and knowledge repositories following a successful pilot.