Northwestern Chemistry welcomes Professor Carlos Baiz from UT Austin, hosted by Professor James Gaynor.
Ultrafast Dynamics of Biomolecules and Water in Biomolecular Condensates
Providing personalized education at scale has been a central challenge for decades. In 1984, Benjamin Bloom's "Two Sigma Problem" showed that students receiving one-on-one tutoring substantially outperform those in conventional classroom settings, yet individualized instruction is not feasible for most institutions. This gap motivated decades of development in educational technology, from early computer-assisted instruction platforms to the more sophisticated Intelligent Tutoring Systems of the 1980s and 90s, which incorporated adaptive learning paths. Each generation brought improvements, but also faced fundamental limitations: rigid rule-based architecture, narrow subject-matter coverage, and an inability to engage in the open-ended, responsive dialogue that characterizes effective human tutoring.
The emergence of large language models (LLMs) represents a significant shift in AI-assisted education that has the potential to address the gap in personalized education. Unlike their predecessors, AI’s engage in natural conversation, adapt explanations to a student's level in real time, interpret visual data such as graphs and spectra, and are available to every student. This talk presents a historical overview of computers in chemistry education, as well as findings from recent studies evaluating LLMs across a range of undergraduate chemistry problems. The presentation will offer concrete, practical guidance: how to design effective prompts, how to use AI to generate and evaluate student work, how to maintain academic integrity, and how to integrate these tools into your courses and research groups without compromising students critical thinking skills.
Kelly Levander
(847) 491-2967
Email