Finding ChatGPT: U of T grad’s startup helps detect AI’s fingerprints
Co-founded by Alex Cui last year, GPTZero has already raised US$3.5 million
With the ink barely dry on his master’s degree from the University of Toronto, Alex Cui is tackling a towering issue spawned by the recent revolution in artificial intelligence tools like ChatGPT: How to tell if written content is generated by person or machine?
Enter Cui’s startup GPTZero. It analyzes patterns in writing, using sentence length and syntax to identify the likelihood text was generated by AI.
“GPTZero has a lot of imagination and potential in a future where AI is everywhere and it’s rampant,” says Cui, who studied computer science as a master’s student in the Faculty of Arts & Science and graduated in June.
“We haven’t even really cracked the surface of what we could do.”
Cui co-founded GPTZero with Edward Tian last January – just a few weeks after ChatGPT went public and triggered a torrent of questions about academic integrity and authenticity. It has already drawn the attention of Silicon Valley investors, raising more than US$ 3.5 million to date.
Cui’s fascination with AI began at University of Toronto Schools – an independent secondary school affiliated with U of T that Tian also attended before going to Princeton University – and at U of T’s Faculty of Applied Science & Engineering’s renowned DEEP Summer Academy. His early affiliation with the university gave him regular access to lectures by U of T faculty, who, even a decade ago, urged Cui and his classmates to learn AI skills as soon as possible.
After high school, Cui earned his bachelor’s degree from the California Institute of Technology but kept his Toronto connections. As an undergrad, he interned with one of the world’s leading experts on self-driving cars, U of T’s Raquel Urtasun, professor in the department of computer science and co-founder of the Vector Institute for Artificial Intelligence.
When Cui returned to U of T, Urtasun supervised his master’s thesis, which included writing code for the Waabi self-driving truck.
“U of T is one of the world leaders in machine learning,” says Cui. “There’s no other ecosystem in the industry that combines AI innovation, commercialization and research partnerships.”
As GPTZero’s chief technology officer, Cui is overseeing fresh innovations, including plugins for Microsoft and Google products that help users detect AI text in any document or on any website.
There’s also Origin for Google Docs, and soon for Microsoft Word. The extension helps writers prove they created the work, giving users the ability to replay a document’s edit history and produce a report verifying human engagement.
Cui predicts this will be a valuable tool in higher education.
“We see a lot of great feedback from teachers because this takes away a lot of stress,” he says. “The teacher will know the student wrote it, and the student won’t be accused of AI plagiarism.”
Cui sees GPTZero offering immense value for other sectors, including publishing. He says the volume of author submissions is growing exponentially, especially with AI-driven content. Software like his could help editors determine what’s been written authentically, and how to take the writing to the next level.
Cui continues to be supported by U of T. GPTZero recently started consulting with the Creative Destruction Lab at the Rotman School of Management and is also a member of the Vector Institute’s FastLane program, which advances AI adoption for Canadian startups and scaleups.
“Alex is an exceptional and experienced researcher and builder. His accomplishments speak for themselves. He is humble, knowledgeable and great to work with,” says Vector’s Junid Ebadi, the FastLane program’s business analyst.
“GPTZero is a fast-growing, well-resourced and driven company. They have been great partners for us, and we are happy to support them in their journey.”
Cui hopes his story will ease the stress of aspiring startup founders.
“It’s not as scary as you think,” he says. “Especially in AI right now, a lot of people want to make bets on promising ideas, and I think if you are a talented individual solving a pressing issue, there’s a good chance you can get some investors to fund your ideas.”
Despite the attention from media and investors, Cui is staying humble and focused on remaining a leader in AI detection.
“We have people who are betting on us, but I wouldn’t say we’re proven yet. To build a sustainable business, you can’t just be a flash in the pan. We’re still very much in the mindset that we are priming ourselves to go up against some stiff competition. We’re not going to take any liberties.”