Evidently, Slack (my favorite corporate comms app) has been scanning user messages to train its AI models. The issue was highlighted in a recent tweet from Corey Quinn. You can read Slack’s strangely worded AI policy here, but it basically says that all Slack users are automatically opted in to its AI training unless the corporate system administrator opts out for your whole company. In other words: individual users cannot opt out by themselves.
Before we go into how messed up this is, here’s what Salesforce (Slack’s parent company) says to do: “To opt out, please have your Org or Workspace Owners or Primary Owner contact our Customer Experience team at feedback@slack.com with your Workspace/Org URL and the subject line ‘Slack Global model opt-out request.’ We will process your request and respond once the opt out has been completed.”
I don’t know how long Slack has been doing this, and I don’t know how much data they have taken from users. The privacy policy is confusing and poorly written, so your guess is as good as mine. Hopefully, someone at Slack will clarify this situation sooner than later. (See update below.)
In practice, AI model builders need vast amounts of data to train on. Proprietary data is far more valuable than public data since most available public data has already been scraped. What can you do to protect yourself and your company? For starters, read the fine print. It’s time to read the privacy policy for every piece of software you use. Search the terms: start with AI, artificial intelligence, and training. Then, read carefully. You may be surprised at what you find.
Update – We received this information from Sean Mills of Slack PR shortly after this article was posted. Here is what he said:
To be clear, Slack is not scanning message content to train AI models. We published a blog post to clarify our practices and policies regarding customer data. (Note: Slack’s policies & practices did not change; we’ve updated the language in our Privacy Principles to clarify Slack does not use customer data to train LLMs.)
The difference between Slack’s generative AI and other machine learning features:
- For Slack AI, which leverages third-party LLMs for generative AI in Slack, no customer data is used to train those third-party LLMs and Slack does not train any sort of generative AI on customer data.
- Separately, Slack has used machine learning for other intelligent features (like search result relevance, ranking, etc.) since 2017, which is powered by de-identified, aggregate user behavior data. These practices are industry standard, and those machine learning models do not access original message content in DMs, private channels, or public channels to make these suggestions.
Simply put: Slack’s other intelligent features (not generative AI via Slack AI) analyze user behavior data, but their models do not access message content. Some examples of user behavior data might be:
- A timestamp of the last message sent in a channel can help Slack recommend channels to archive
- The # of interactions between two users is incorporated into the user recommendation list when a user goes to start a new conversation
- The # of words overlapping between a channel name and other channels can inform its relevance to that user
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.