The AI Revolution: What Does the Zoom AI Training Update Mean for your Privacy?

The year 2023 has undoubtedly solidified its status as the year of the Artificial Intelligence (AI) revolution. From ChatGPT to our very own AI lawyer, Robin – everyone is looking for more data. Enter Zoom AI training, which has caused quite a (rightful) stir.

The adage “with great power comes great responsibility” rings particularly true, especially concerning the realm of privacy and data safeguarding.

What Zoom AI Training Is All About

Initially, Zoom released updated Terms of Service which gave the platform “perpetual, worldwide, non-exclusive, royalty-free, sublicensable and transferable license and all other rights” to use your video recordings, and other data for “machine learning, artificial intelligence, training, and testing”.

The platform later released an updated version, stating that consent is required to use your data to train AI. Zoom AI training was further clarified by the platform’s Chief Product Officer, Smita Hashim who stated that for Zoom AI training “we do not use audio, video or chat content for training our models without customer consent”.

However, currently there is no option to specifically opt out of Zoom AI training using your data. This leaves open the concern that users’ data is being used for machine learning and other training and testing purposes, which raises a big question for data in general when it comes to AI – can you really trust platforms with your data anymore?

UPDATE: Since the publication of this article, Zoom has changed its terms of service after it raised widespread privacy concerns.

It now states that the company will not use customer content whatsoever for Zoom AI training. However, the recent back and forth by Zoom does leave the question open as to what do platforms do with your data and is your privacy at risk.

Data Protection and Privacy Is Paramount

Platforms, like Zoom, harbor a lot of data, some of it extremely sensitive – think about having a call with your lawyer. You and your lawyer want the conversation to be completely private and privileged. But with the rise of AI, platforms are prone to harnessing your data to use to train AI.

Our approach with Robin using the Microsoft Azure OpenAI service ensures that your questions and your outputs are not available to the general OpenAI environment or Microsoft or any 3rd party products or services.

The Azure OpenAI service is fully controlled by Microsoft. Microsoft hosts the OpenAI models in Microsoft’s Azure environment and the Service does NOT interact with any services operated by OpenAI.

The data we are adding to the bot is continuously curated by the Legal & Tax team and includes updates from case law, policies, and tax updates.

That’s why it has become imperative for users of any platform to take the following into account:

Terms of Service: Know what the terms of service are with regard to their data
Privacy Rights: It is crucial that users understand their rights concerning privacy and data protection
Data Ownership: Who owns user data and what can be done with it by the platform that holds it
Consent Parameters: Whether user consent is implied or must be explicit in terms of data use/ownership
Data Exchange: Only exchange data with platforms you trust and only exchange the data that is absolutely required.

As a user, it is really important to understand the difference between a provider collecting network performance data versus the actual content of the meeting. Whilst both may be in the interest of advancing the technological aspirations of the vendor, the former is not so much of a problem whilst the latter could be very material.

What has become very clear is that videos are now very much in the data domain. AI is making it easy and reasonable to train one’s system based on this content.

With the new Zoom AI training update and the advancement of AI, users need to ensure that they understand completely what happens to their data given that if you are a doctor or a lawyer you may have obligations such as client confidentiality that existed long before these digital tools ever did.

If you would like to chat to us about AI feel free to reach out. You can do so by emailing us at in**@le***********.com, calling us on +27 11 719 2000 or scheduling time with us here