Slack AI Training with Customer Data (slack.com)

Article summary

Slack uses customer data to train AI models, but claims to have technical controls in place to prevent data leakage across workspaces. Customers can opt out of having their data used to train global models, but this opt-out is only available by email. Slack's AI models are used for features such as channel recommendations, search results, and autocomplete. The company states that it does not use customer data to train generative AI models without affirmative opt-in consent.

Main themes

• AI training data
• Customer data privacy
• Slack's data usage policies
• Opt-out mechanisms
• Data leakage prevention
• Generative AI models

What commenters say

• The lack of a clear and easily accessible opt-out mechanism for customers to prevent their data from being used for AI training is a significant concern.
• Some companies prioritize their own interests over customer privacy and will inevitably update their terms of service to use customer data for AI training.
• Having an open-source client can help verify a company's claims about data privacy and security.
• The statement that a company 'can't' access customer data is not necessarily meaningful, as it may be possible for employees to access the data despite technical measures in place.
• Legislation is necessary to bring about real change in how companies handle customer data and AI training.
• The use of customer data for AI training without explicit opt-in consent is a violation of customer trust.
• Some argue that if data is stored in a database that a company can access, it will eventually be used for AI training, regardless of initial promises to the contrary.
• The distinction between 'can't' and 'won't' access customer data is important, and companies should be transparent about their capabilities and intentions.