CS alumna Patricia Thaine is the CEO of Private AI, a fast-rising privacy firm she co-founded with privacy and machine learning experts from U of T. (Photo: Kemeisha McDonald)
The power and potential of AI was still largely unknown when Patricia Thaine came to U of T for graduate computer science studies in 2013.
“I just really liked the idea of playing around with language and finding its patterns,” says the Arts & Science alum, who earned her master’s degree in 2015.
Thaine is now CEO of Private AI, a fast-rising privacy firm she co-founded with privacy and machine learning experts from U of T.
She credits the school’s rich ecosystem of cutting-edge education and support for nurturing her natural interests in computers and languages while opening her eyes to the possibilities machine learning and AI created.
Thaine says the classes and research she was able to do in privacy and natural language processing were pivotal in helping her understand where the technology was and where the gaps were. Microsoft’s voice-activated digital assistant “Alexa” is a common example of a system based on natural language processing.
“I was working on acoustic forensics and it really made me understand why it was so important to have privacy constraints in place,” says Thaine.
“Before AI, people would use a combination of regular expressions, like exact pattern matching, but with some bad results. And you can't have bad results when you're dealing with personal information.”
Another complication is complying with various data protection regulations when sending data to a third party.
“The reason we use AI is because for data like text — unstructured text and free text — images, and documents, you really need to understand the context to determine what is personal information,” she says.
She met company co-founder Pieter Luitjens (MEng, 2015) during her studies and she eventually tested an early version of Private AI at the Rotman School of Management’s Creative Destruction Lab in 2017, a program for massively scalable science and technology-based startups, before scrapping the original idea and starting this second version with Luitjens. They returned to the Creative Destruction Lab to complete the program in 2021.
Thaine says they also found the U of T non-profit partner Vector Institute, which is geared to AI enterprise, was immensely helpful for networking.
“The general knowledge base I was able to get at U of T was very valuable.”
Private AI detects and removes personally identifiable information in a client’s data and keeps it safe so it doesn’t get shared.
The layer of privacy it provides also helps clients be compliant with regulations in other jurisdictions such as the European Union’s general data protection regulations, which set a global standard for collecting, storing and sharing data.
Although Private AI’s privacy layer for software and compliance has been finding traction in health care data security, Thaine says their mission goes far beyond that sector.
Founded in 2019, Thaine’s company is backed by Microsoft's M12 venture capital fund and ranked on several key industry lists of the most promising private AI startups.
“We're vertical agnostic. We also deal with a lot of customer service, pharma, insurance and banking. It's applicable to a lot of different kinds of industries with personal information in their data,” says Thaine.
“Our goal is to be the privacy layer for software, to be everywhere.”
— Original story by Peter Boisseau for A&S News