New Claude feature lets bosses read employees’ AI chats in move that's restricted in over 30 countries
Anthropic has unveiled a new feature that allows employers to analyse how their staff use the AI chatbot Claude, giving managers more insight into work being done with the technology.
However, the move has raised concerns about employee privacy as the addition will allow bosses access to workers’ AI chats, which is a feature that is restricted in over 30 countries.
Known as Smart Reports, the feature is available to companies that use Claude Enterprise and will enable employers to examine samples of conversation transcripts to determine which work tasks they are using the tech for.
Claude support explains: “Each organization can run up to 10 reports per month at no charge. Reports that are queued or in progress count toward this limit, and reports that fail don't.
Advert
“When your organization reaches its limit, you can’t create new reports. The smart reports page shows the date your limit resets, which is the first day of the next calendar month. If you need more reports before then, contact your account team.”
How do Claude’s Smart Reports work?
According to the support page, an admin will be able to choose what the report will cover, whether that is a team, a time range of up to 28 days, and focus areas for the analysis.
From there, Claude will examine transcripts and write up a report on its findings, as it adds: “Smart reports analyze how a team uses Claude and report on the work getting done, what it costs at list price, where sessions run into friction, and which repeated patterns are worth packaging as shared skills. This guide explains how smart reports work, what appears in a smart report, and how to create, share, and delete them.”
Claude has safeguards for personal conversations
The announcement by Anthropic has sparked concerns over the privacy of employees, with many people taking to social media to share their own thoughts on the matter.
On Reddit, one user wrote: “On one hand, I hate the idea of my manager seeing my chat logs. On the other hand, it’s pretty realistic that lots of engineers will use their work tokens on side projects/businesses if there’s no consequences, which would be a huge money suck on the company considering the price of frontier AI right now.”
And another added: “Claude needs to learn that snitches get stitches.”
Anthropic has addressed the concerns, explaining that sessions which Claude ‘identifies as involving restricted topics, including personal conversations, are used only for aggregate analysis and don’t include summaries’.
Personal conversations will also not include individual sessions or names, and the admins will be able to delete any report.