However, the move has raised questions about the necessity of AI in Slack and the privacy implications of the new features. Earlier this year, Slack faced criticism for training a recommender system on customer data by default, although it was later revealed that the data was used for emoji recommendations, not for training large language models. Despite these concerns, Dresser maintains that enterprise AI solutions need to be trustworthy and safe, and insists that no large language models are trained on Slack data.
Key takeaways:
- Slack is shifting its business chat platform into a “work operating system,” making it a hub for AI applications from Salesforce, Adobe, and Anthropic.
- Slack announced several new features for a pricier tier of the messaging platform, Slack AI, which includes AI-generated Huddle summaries and the ability to chat with Salesforce’s AI agents.
- Slack's CEO, Denise Dresser, believes that AI is a natural fit for Slack as it allows for conversational interaction and action within the flow of work.
- Despite privacy concerns raised earlier this year, Slack maintains that it does not use customer data to train its AI and is focusing on ensuring its AI solutions are trustworthy and safe.