The real risk with AI isn’t sci-fi. It’s how we adopt it.

In this final episode of The STEM Career Coach, we’re joined by Vered Shwartz (Assistant Professor of Computer Science at the University of British Columbia) and author of Lost in Automatic Translation: Navigating Life in English in the Age of Language Technologies. We unpack where language AI is genuinely useful — and where using it to automate decisions can go wrong, especially in high-stakes contexts like healthcare and law.

What you’ll learn in this conversation:

- Why automating “critical error” processes is where the risk lives

- How bias can show up through data and deployment, not “evil AI”

- Why language models can sound confident while still being unreliable

- How large language models are trained (in plain English)

- Why culture and context are hard for AI to get right

If you’re adopting AI at work, this is the mindset shift that keeps you safe and effective.

Resources:

- Vered’s book: https://lostinautomatictranslation.com/

- Vered at UBC: https://www.cs.ubc.ca/~vshwartz/

Subscribe for more STEM career stories and practical, real-world conversations on technology and work.