Just five months after the company quietly pushed an update to its terms of service allowing it to train its AI on customer data, Zoom has spoken out about customer data and privacy.
CPO Smita Hashim confirmed in a blog post that the video conferencing maker, “will not use audio, video, or chat customer content to train [its] artificial intelligence models without [customer] consent.”
Hashim said that this would be a move that reflects the company’s “commitment to transparency and user control,” though many are still concerned that opting out is not as easy as it should be.
Zoom doubles back on AI training terms
Like most other companies trying to win over enterprise customers, Zoom has released its own AI tools to save workers time.The pair of recently announced Zoom IQ features, Meeting Summary and Team Chat Compose, are currently available on a free trial basis.
Zoom says that “account owners and administrators control whether to enable these AI features for their accounts.”
Enabling the Zoom IQ features will present a “transparent consent process for training… AI models using your customer content,” which Zoom says is not used for third-party models and is only for product improvement.
The blog post serves to highlight the changes made to the terms of service in section 10, entitled ‘Customer Content,’ and particularly, the addition of a line detailing that audio, video or chat Customer Content is not used without consent.
While the ability to enable or disable data sharing is a clear option for admins, participants looking to join a call can either accept and join or decline and be denied entry.
Zoom isn’t the only company that has found itself in hot water over its AI terms – Google has also been in the spotlight recently about its terms.