The popular videoconferencing app Zoom has recently been caught up in controversy over its attempts to use customer calls to train artificial intelligence (AI) software. Following an initial statement suggesting that the recordings would not be used for the purpose of AI training, Zoom Video Communications, Inc. has backtracked its stance on the matter, assuring customers that the recordings will no longer be used for such purposes.
The issue first arose last week when Zoom announced in a blog post that it had begun “using the transcripts of some customer calls to train the automatic speech recognition (ASR) system that will eventually help make Zoom an even better product.” As the company stated, “ASR can help make Zoom an even better product by improving its speech recognition capabilities, which can result in improved features such as transcriptions and the ability to quickly understand customer conversations.”
This announcement was met with an immediate and strong response from customers and the press. Many expressed concerns that their private conversations could become part of a training dataset and be used to train algorithms for unknown purposes.
In response to the outcry, Zoom backtracked its position and issued a statement insisting that its only intent was to “not use customer interactions for training AI algorithms.” The statement goes on to explain that, “In a blog post published on June 16, we described the potential value of using recordings of customer interactions to train ASR, but we made no commitment to do so.”
The company also went on to assure customers that any recordings that were done for the purpose of training AI will no longer be used. The statement reads, “We have decided that we will not use customer interactions for training AI algorithms, and any recordings that may have been made previously for this purpose will not be used.”
This backpedaling from Zoom is certainly welcome news from customers, given the recent string of privacy-related controversies surrounding many tech companies. With this latest assurance, users can rest assured that their private conversations will not be used to train AI without their expressed consent.
