Powered by recent advances in code-generating models, AI assistants like Github Copilot promise to change the face of programming forever. But what is this new face of programming? And how can we help programmers use these assistants more effectively?

I will start with a survey of studies on how programmers interact with state-of-the-art AI assistants, including a grounded theory study conducted in our group. I will then discuss recent work that proposes new interaction models for AI assistants, providing better support for tasks specific to AI-assisted programming, such as prompt crafting and suggestion validation. In particular, I will talk about LEAP, a python environment we developed, which uses live programming to help developers validate AI-generated code. Finally, I will share my outlook on the future of AI programming assistants more broadly, and discuss the most exciting opportunities our community has to make these assistants more effective, safe, and usable.

Nadia Polikarpova is an Assistant Professor of Computer Science and Engineering at the University of California, San Diego. She received her PhD from ETH Zurich (Switzerland), and spent a couple of years as a postdoc at MIT CSAIL. She is a 2020 Sloan Fellow and a recipient of 2020 Intel Rising Stars Award and the 2020 NSF Career Award. Her research interests are in program synthesis, program verification, and type systems.