The news, 365 days behind — on purpose Delayed live · replaying 2025

One Year Ago.AI

Remember how fast this is.

06AUG2023replayed
one year on
policyZoom

Zoom terms of service now draw backlash over broad customer-content license and AI-training concerns

Users and privacy advocates raise alarms over a policy change that grants Zoom a perpetual, worldwide license to customer content, including for AI model training, without a method to decline.

Zoom has updated its terms of service, effective August 11, 2023, to explicitly grant itself a “perpetual, worldwide, non-exclusive, royalty-free, sublicensable, and transferable license” to customer content for what the company calls “Permitted Uses.” That list includes improving the service, which critics say could cover training AI models on user data, with no opt-out mechanism mentioned in the terms.

The change was flagged on Hacker News where it quickly rose to the front page with over 1,500 points and 500 comments. Many users expressed dismay that the company behind the pandemic-era video conferencing staple now claims rights to their private conversations for AI development.

“Thankfully nothing like this is in Jitsi Meet’s TOS,” wrote one commenter, though others quickly noted that Jitsi’s terms contain similar language, albeit less explicit about AI training. The debate highlighted the broader tension: every major video conferencing service requires some license to operate, but Zoom’s broad license language — without offering users a choice — has reignited discussions about data ownership in an era of increasingly hungry AI systems.

_
_ddzr@_ddzr

Suggested Jitsi Meet as an alternative, claiming its TOS does not include such provisions.

J
jonas21@jonas21

Pointed out that Jitsi's TOS grants similar rights with mushier language, and argued that explicit disclosure is generally a good thing.

G
gnicholas@gnicholas

A former corporate lawyer argued the 'for You' qualifier in Jitsi's TOS would likely prevent training AI for general use.

One year later — open only if you can handle spoilers

Zoom later backtracked, clarifying in a blog post that it would not use customer audio, video, or chat to train AI without consent. The episode became a case study in how poorly communicated TOS changes can trigger backlash, but the underlying permissive licensing language remained a concern for privacy advocates.

Replay thisPost on XRedditHNLinkedIn

The Weekly Replay · free by email

This week, one year ago — every Sunday.

One email each Sunday: the week's replayed AI news, with the one-year-later annotations included. Written like it's breaking — dated like it isn't.

Free · double opt-in · unsubscribe anytime · privacy