At the Intersection of AI, Governments, and Google - Tim Hwang

TL;DR

Tim Hwang, a global public policy lead on AI and machine learning at Google, describes his role at the intersection of technical capabilities and societal impact. He discusses challenges like bias in ML systems, adversarial examples, data privacy trade-offs, and how governments are still in an exploratory phase regarding AI regulation.

Key insights

• Fairness in ML involves a trade-off: collecting more diverse data to debias systems can raise privacy issues, especially for minorities.
• Adversarial examples reveal that machines perceive images fundamentally differently from humans; tiny pixel edits can cause a panda to be classified as a giraffe.
• Automation is not inevitable even if technically possible—security concerns (e.g., adversarial attacks on security cameras) can block deployment.