← The issue Alumni DigestU of T

25UTSC / Research

Can a smaller AI model do more?

Scarborough researchers are working toward AI that is smaller, faster and more reliable without losing the point of the model.

Reported by U of T Scarborough Staff

U of T Scarborough researchers examine how to make AI smaller and more reliable.
UTSCU of T Scarborough researchers examine how to make AI smaller and more reliable.Photo: U of T Scarborough / official source image

Researchers at U of T Scarborough are exploring whether AI systems can become smaller, faster and more reliable. The engineering question is also a cultural one: what should a model be asked to carry, and what can be left out?

A smaller system is not automatically a better one. The promise lies in making performance, energy use and reliability visible enough that people can choose the right tool for the job.

Sources 1
  1. Original reporting · utsc.utoronto.ca
Next · p.26The Great Lakes’ fish problem is measured in billionsUTSC researchers are looking for ways to stop invasive round goby from moving farther inland, even when eradication is unrealistic.

Chapter 03 · Science & TechnologyRead it in the issue →

More from Science & Technology

p.14 · Science & TechnologyA small molecular edit with a wide horizonp.15 · Science & TechnologyThe chemistry of making more with lessp.16 · Science & TechnologySix researchers arrive with six new questions

A summer issue about the people, places and questions that keep the University of Toronto moving when the timetable loosens.

This story is from the August 2026 issue of Alumni Digest · U of T. It's free, monthly and independent.

Subscribe freeRead the whole issue

Alumni Digest is independent and is not an official University of Toronto publication. Stories summarize the public reporting linked under each piece; photographs belong to the credited sources and remain under rights review.