Zoom caused a minor internet kerfuffle with some overly broad ToS. These terms made it sound like Zoom could train AI systems on customer logs and potentially calls, the latter with vaguely defined consent. Now it’s walking that back with specifics:

Zoom has updated its terms of service and reworded a blog post explaining recent terms of service changes referencing its generative AI tools. The company now explicitly states that “communications-like” customer data isn’t being used to train artificial intelligence models for Zoom or third parties. What is covered by communications-like? Basically, the content of your videoconferencing on Zoom.

In the arms race for datasets to train LLMs, it’s easy to see this as Zoom trying to open up a customer treasure trove on the sly. But enough years on Daily Tech News Show taught me never to ascribe to malice what can be explained by incompetence, or at least over zealous lawyers. We’ve seen this pattern over and over again, lawyers trying to give a company plenty of wiggle room, people see this and assume there is a larger objective, company revises terms.

This isn’t to say we should give these companies a pass for shoddily written ToS. We should highlight obvious overreach and potential for abuse. These terms are obscure by design and without (occasionally) hyperbolic coverage of them, we’d likely never see them. Still it’s always my instinct to withhold ire in these cases until the company responds.