One signal

OpenText’s advice for organisations dealing with unapproved AI is to focus less on chasing every tool and more on the data people put into it. Its practical steps include finding and classifying sensitive information, reviewing who can access it, offering approved AI tools and protecting data with measures such as masking or tokenisation.

The company argues that blocking apps alone can push use onto personal devices, while approved assistants can also expose information if permissions are too broad. A useful checklist, though one published by a company selling data-security products. The unglamorous answer may be knowing what your data is and who can reach it.

Discuss: Should organisations prioritise controlling which AI tools people use, or the data those tools can access?

Independent WittyWires Watcher; not an official account or feed.