In the midst of ongoing debates about data privacy and AI, Slack, the Salesforce-owned chat platform, is facing significant backlash from its users. The controversy centers on Slack’s approach to using user data to train its AI services, highlighting the growing tension between big tech companies and their users over data appropriation and transparency.
Slack, like many other companies, utilizes its user data to enhance and train its AI services. However, the company’s method of implementing this practice has drawn widespread criticism. Users are automatically opted into this data usage, and those who wish to opt-out must do so by emailing Slack directly. This policy was buried in Slack’s privacy principles, which many users found to be outdated and confusing.
The issue gained attention when a user posted about it on Hacker News, a popular community site for developers. The post linked directly to Slack’s privacy principles without additional commentary, igniting a lengthy discussion. Many users were surprised to learn that Slack defaults them into AI training and requires an email to opt-out. The thread quickly went viral, sparking further conversations and questions across other platforms.
Confusion Over Privacy Principles
The Hacker News thread revealed a significant confusion among users about how Slack applies its AI privacy principles. Users were puzzled by the absence of any mention of “Slack AI” — a product that allows users to search for answers and summarize conversation threads — on the privacy principles page. Additionally, there were questions about Slack’s reference to both “global models” and “AI models.”
Slack’s privacy principles have been in effect since at least September 2023, according to pages on the Internet Archive. However, the recent uproar suggests that many users were unaware of these terms until now. Slack’s terms indicate that customer data is used to train “global models” for features like channel and emoji recommendations and search results. A company spokesperson clarified that these models do not learn, memorize, or reproduce customer data. Nonetheless, the policy does not address the broader scope and future plans for AI model training.
I'm sorry Slack, you're doing fucking WHAT with user DMs, messages, files, etc? I'm positive I'm not reading this correctly. pic.twitter.com/6ORZNS2RxC
— Corey Quinn (@QuinnyPig) May 16, 2024
Clarifications from Slack
In response to the backlash, Slack provided further clarifications. The company stated that while customers who opt-out of data training would still benefit from “globally trained AI/ML models,” it remains unclear why customer data is used for features like emoji recommendations in the first place.
Moreover, Slack emphasized that it does not use customer data to train Slack AI. Slack AI is a separately purchased add-on that employs large language models (LLMs) hosted within Slack’s AWS infrastructure, ensuring that customer data remains in-house and is not shared with any LLM provider. This setup ensures that customer data stays within the organization’s control and is used exclusively for that organization’s purposes.
Aaron Maurer, a Slack engineer, acknowledged the need to update the privacy principles page to clarify the relationship between these principles and Slack AI. He noted that the current terms were written before Slack AI was developed and primarily reflect the company’s work on search and recommendations.
The controversy surrounding Slack’s AI data usage policies serves as a stark reminder that in the rapidly evolving field of AI, user privacy cannot be an afterthought. Companies must ensure that their terms of service clearly and transparently outline how user data is utilized. As AI continues to integrate more deeply into everyday technology, maintaining user trust through clear communication and robust privacy protections will be paramount.
For Slack, the current backlash represents a critical moment to reassess its policies and ensure that they align with user expectations and industry best practices. By addressing these concerns proactively, Slack can better navigate the complex landscape of AI and data privacy, ultimately fostering a more trusting and engaged user community.