Slack, the popular workplace communication platform, has come under scrutiny for using customer data to train its artificial intelligence (AI) and machine learning (ML) models without obtaining explicit permission from users. The company’s privacy policy reveals that it analyses customer messages, content, and files to develop these models, and users are automatically opted in by default.
The controversy came to light when Corey Quinn, an executive at DuckBill Group, highlighted the issue on social media. This sparked outrage among Slack users who were unaware of the company’s data usage practices. Many users expressed frustration over the lack of transparency and the difficulty of opting out of the AI training process.
To opt-out, individual users must request their organisation’s Slack administrator to contact Slack’s Customer Experience team via email, providing their workspace URL and the subject line Slack Global model opt-out request. This process has been criticised for burdening users to protect their data rather than requiring explicit consent from the outset.
Slack has attempted to clarify its position, stating that machine learning models are used for platform-level features such as channel and emoji recommendations and search results. The company also emphasised that customers can exclude their data from being used to train these non-generative ML models.
However, inconsistencies in Slack’s privacy policies have led to further confusion. While one section states that Slack cannot access the underlying content when developing AI/ML models, the machine learning model training policy contradicts this claim. Additionally, Slack’s marketing of its premium generative AI tools implies that user data is not used for training, which may be misleading given the company’s selective application of this statement.
Slack’s current opt-out approach has faced significant backlash, and the company may need to change its policies to ensure greater transparency and user control over its data.
