{"id":5226,"date":"2024-05-26T16:14:30","date_gmt":"2024-05-26T16:14:30","guid":{"rendered":"https:\/\/aitesonics.com\/yuck-slack-has-been-scanning-your-messages-to-train-its-ai-models-181918245\/"},"modified":"2024-05-26T16:14:30","modified_gmt":"2024-05-26T16:14:30","slug":"yuck-slack-has-been-scanning-your-messages-to-train-its-ai-models-181918245","status":"publish","type":"post","link":"https:\/\/aitesonics.com\/yuck-slack-has-been-scanning-your-messages-to-train-its-ai-models-181918245\/","title":{"rendered":"Slack has been using data from your chats to train its machine learning models"},"content":{"rendered":"
Slack trains machine-learning models on user messages, files and other content without explicit permission. The training is opt-out, meaning your private data will be leeched by default. Making matters worse, you\u2019ll have to ask your organization\u2019s Slack admin (human resources, IT, etc.) to email the company to ask it to stop. (You can\u2019t do it yourself.) Welcome to the dark side of the new AI training data gold rush<\/a>.<\/p>\n Corey Quinn<\/a>, an executive at DuckBill Group, spotted<\/a> the policy in a blurb in Slack\u2019s Privacy Principles<\/a> and posted about it on X (via<\/a> PCMag<\/em>). The section reads (emphasis ours), \u201cTo develop AI\/ML models<\/strong>, our systems analyze Customer Data<\/strong> (e.g. messages, content, and files<\/strong>) submitted to Slack as well as Other Information<\/strong> (including usage information) as defined in our Privacy Policy and in your customer agreement.\u201d<\/p>\n In response to concerns over the practice, Slack published a blog post<\/a> on Friday evening to clarify how its customers\u2019 data is used. According to the company, customer data is not used to train any of Slack\u2019s generative AI products \u2014 which it relies on third-party LLMs for \u2014 but is fed to its machine learning models for products \u201clike channel and emoji recommendations and search results.\u201d For those applications, the post says, \u201cSlack\u2019s traditional ML models use de-identified, aggregate data and do not access message content in DMs, private channels, or public channels.\u201d That data may include things like message timestamps and the number of interactions between users.<\/p>\n A Salesforce spokesperson reiterated this in a statement to Engadget, also saying that \u201cwe do not build or train these models in such a way that they could learn, memorize, or be able to reproduce customer data.\u201d<\/p>\n The opt-out process requires you to do all the work to protect your data. According to the privacy notice, \u201cTo 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 \u2018Slack Global model opt-out request.\u2019 We will process your request and respond once the opt out has been completed.\u201d<\/p>\n The company replied<\/a> to Quinn\u2019s message on X: \u201cTo clarify, Slack has platform-level machine-learning models for things like channel and emoji recommendations and search results. And yes, customers can exclude their data from helping train those (non-generative) ML models.\u201d<\/p>\n How long ago the Salesforce-owned company<\/a> snuck the tidbit into its terms is unclear. It\u2019s misleading, at best, to say customers can opt out when \u201ccustomers\u201d doesn\u2019t include employees working within an organization. They have to ask whoever handles Slack access at their business to do that \u2014 and I hope they will oblige.<\/p>\n Inconsistencies in Slack\u2019s privacy policies add to the confusion. One section states, \u201cWhen developing Al\/ML models or otherwise analyzing Customer Data, Slack can\u2019t access the underlying content. We have various technical measures preventing this from occurring.\u201d However, the machine-learning model training policy seemingly contradicts this statement, leaving plenty of room for confusion.<\/p>\n In addition, Slack\u2019s webpage marketing<\/a> its premium generative AI tools<\/a> reads, \u201cWork without worry. Your data is your data. We don\u2019t use it to train Slack AI. Everything runs on Slack\u2019s secure infrastructure, meeting the same compliance standards as Slack itself.\u201d<\/p>\n In this case, the company is speaking of its premium generative AI tools<\/a>, separate from the machine learning models it\u2019s training on without explicit permission. However, as PCMag<\/em> notes, implying that all of your data is safe from AI training is, at best, a highly misleading statement when the company apparently gets to pick and choose which AI models that statement covers.<\/p>\n Update, May 18 2024, 3:24 PM ET:<\/strong> This story has been updated to include new information from Slack, which published a blog post<\/a> explaining its practices in response to the community’s concerns.<\/p>\n Update, May 19 2024, 12:41 PM ET:<\/strong> This story and headline have been updated to reflect additional context provided by Slack about how it uses customer data.<\/p>\n","protected":false},"excerpt":{"rendered":" Slack trains machine-learning models on user messages, files and other content without explicit permission. The training is opt-out, meaning your private data will be leeched by default. Making matters worse, you\u2019ll have to ask your organization\u2019s Slack admin (human resources, IT, etc.) to email the company to ask it to stop. (You can\u2019t do it […]<\/p>\n","protected":false},"author":6,"featured_media":5226,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[138,3877,48,95,556,6438,1864],"tags":[143,3878,59,101,559,6439,1865],"yoast_head":"\n