Slack Clarifies AI and ML Policy Regarding Customer Data Usage

Slack Clarifies AI and ML Policy, Addresses Customer Data Concerns

Last week, Slack faced backlash after unveiling terms that suggested the company was using customer data to train AI and ML models. The language in question stated that the company’s systems analyze customer data, including messages, content, and files, to develop AI/ML models. Additionally, Workspace Owners were required to manually email the company to opt out, rather than having a simple option in the app or on the website.

In response to the concerns raised, Slack has clarified its stance on the use of customer data for developing or training large language models (LLMs). A spokesperson from Salesforce, the parent company of Slack, provided information to WPN stating that Slack uses industry-standard platform-level machine learning models to enhance the product experience for customers. These models do not access original message content in direct messages, private channels, or public channels, and they are not built or trained in a way that could reproduce customer data.

Furthermore, Slack clarified that it does not develop LLMs or other generative models using customer data. The company leverages third-party LLMs in its Slack AI product offering, ensuring that no customer data is used to train these models. Additionally, Slack hosts these models on its own infrastructure, preventing customer data from leaving the company’s trust boundary.

In a blog post accompanying the clarification, Slack emphasized its industry-standard privacy measures and stated that it does not leak data across workspaces. The company uses numerical features, such as message timestamps and user interactions, to make relevant suggestions without directly accessing message content.

While customers can opt out of training non-generative ML models by emailing Slack, the company highlights that mass opt-outs could impact the overall performance of these models. Slack assures customers that opting out will not disable any product features but may slightly affect the performance of globally trained ML models.

Despite the valid reasons provided by Slack for the email opt-out process, some critics argue that the company should offer a more user-friendly way for customers to opt out. Requiring users to send an email creates an unnecessary barrier and may deter users from taking advantage of the option.

Overall, Slack’s clarification on its AI and ML policy aims to address customer concerns and ensure transparency in how customer data is used to enhance the product experience. The company’s commitment to privacy measures and industry-standard practices underscores its dedication to protecting customer data while leveraging AI and ML technologies.