9th August, New Jersey: In response to mounting concerns and backlash, Zoom has taken swift action to update its terms of service, vehemently denying any unauthorized training of its artificial intelligence (AI) on customer calls. The company’s move comes after users raised alarms over potential AI training using audio, video, and chat data without explicit consent.
The controversy emerged earlier this year when astute users noticed alterations to Zoom’s terms of service, sparking fears that the company was leveraging customer interactions to enhance its AI models. In a bid to quell the growing unease, Zoom has now issued a blog post assuring users that their data will not be utilized for AI purposes without their explicit permission.
Zoom, a prominent video-calling application, has unveiled innovative AI-powered features in recent months, including the ability to condense meeting summaries without necessitating full session recordings. Although initially introduced as part of a complimentary trial, these features have come under scrutiny due to concerns that the original wording of the terms of service could potentially grant Zoom broader access to user data, encompassing sensitive customer calls.
Robert Bateman, a data protection specialist, emphasized the significance of transparency in contractual agreements involving user data. He warned against broad provisions that grant service providers substantial latitude in utilizing user-generated data for various objectives.
Late on Monday, Zoom revised its terms of service to explicitly state, “Zoom will not use audio, video or chat customer content to train our artificial intelligence models without your consent.” This amendment aims to provide clarity and reassure users about the company’s commitment to safeguarding their privacy.
Artificial intelligence applications, powered by complex algorithms and extensive data, simulate human-like behavior to execute specific tasks. Nevertheless, the practice of extracting massive amounts of online data for AI training has sparked concerns about potential inclusion of personal or copyrighted material in the datasets.
Zoom responded by clarifying that customers retain control over enabling generative AI features and consenting to share customer content for product improvement. The company’s chief product officer, Smita Hashim, emphasized that account owners and administrators maintain the autonomy to activate these features, ensuring a transparent consent process for AI model training using customer content.
As Zoom endeavors to assuage privacy concerns and strike a balance between AI advancement and user protection, the spotlight remains on the evolving landscape of data usage and artificial intelligence ethics.