Cyberhaven Monitors Data Inside ChatGPT and Claude
Cyberhaven is adding controls that inspect and classify sensitive information used inside enterprise AI workspaces.
Companies are moving beyond the question of whether employees may use AI. The harder question is what information they are allowed to put into it.
What happened
Cyberhaven launched integrations that inspect and classify sensitive information inside ChatGPT Enterprise and Claude Enterprise.
Support for ChatGPT is available now, while Claude support is expected in August. The system is designed to help organisations identify sensitive data uploaded, generated or shared within approved AI workspaces.
Why it matters
Blocking every AI tool is increasingly impractical, but allowing unrestricted use creates risks around customer records, source code, financial information and other confidential data.
Monitoring activity inside approved enterprise products gives security teams a middle ground: employees can use AI while organisations apply policies to the information moving through it.
The bigger picture
Enterprise AI security is shifting from simple access control toward continuous data governance. As assistants and agents become embedded in daily workflows, companies will need visibility not only into which model is used, but into the context, files and outputs passing through each interaction.