I'm confused about the data collection settings in PyCharm, particularly regarding JetBrains' AI services. There's a setting under Appearance & Behavior > System Settings > Data Sharing that says, 'Allow detailed data collection by JetBrains AI.' It mentions that this collection might include the full text of inputs, including source code snippets. I can't tell if this means that JetBrains can read my code or if there's any way to stop this from happening. Even with this setting turned off, it seems that some essential data will still be sent to different AI providers. This really worries me, especially when dealing with sensitive projects. Is there any way to ensure my code stays private? What's considered 'essential' data exactly?
2 Answers
You can indeed disable the AI Assistant plugin altogether if you’re worried about data sharing. Initially, there was a big push for local predictions, but you’re right – things seem to have changed. My workplace relies on Copilot for our coding needs, so I had to automatically disable other plugins. It's always better to be cautious, especially with sensitive info!
My company uses similar approaches, and it helps to enforce safe coding practices.
That’s how LLMs work, unfortunately. To keep your code safe, you can disable the AI Assistant plugin, turn off detailed data collection, and set your firewall to block outgoing requests to public LLM servers. I’ve seen it work for organizations looking to tighten security around sensitive projects.
Great tips! Keeping egress blocked and using an on-prem model for sensitive work can really help.
Totally agree. It’s all about policies and ensuring that specific projects adhere to those privacy standards.
That’s a solid plan! Disabling the AI Assistant seems like a good move to ensure privacy.