Which parts of my code where uploaded to the cloud?

[KenoAlbrecht]

Select Topic Area

Question

Body

Hello there,

I am an active user of the Github Copilot and other LLM-Tools as ChatGPT and Microsoft Copilot. My experience using these tools where mixed as they sometimes really help creating new ideas and sometimes really don't help when they produce professional sounding bullshit featuring completely false facts. Which is pretty dangerous in my opinion.

But anyway, my question is: Where can I look up, which parts of my code (or is everything automatically uploaded?) got uploaded to the cloud for further analyzation? I already tried to search on Google for that but got no official answer. It only said that "snippets" where uploaded (on medium.com from someone who is at least not official working at Microsoft/Github), what ever that means. But that was no official document/website. And even if it was, it is still quite unclear in my opinion.

I also tried asking the Github Copilot directly via its chat feature but that only said this: "No, your workspace was not uploaded to the cloud. I can only interact with the code you provide or ask about within this session. If you have any specific questions or need assistance with a particular part of your code, please let me know!"

Which is confusing as before that I just asked it something regarding my workspace (where did I define function xyz and which arguments does it need to work properly) using the "@workspace" operator.
Which could or would indicate that the whole workspace containing all my code was uploaded to answer my question. Or does it do some processing offline? Which would be great and awesome but I can't really see any additional CPU, GPU and memory usage on my system and it also would not have enough RAM or VRAM for a larger language model.

I really like the idea and concept of Github Copilot and other AI-assistents but it is alway quite unclear for me which data they use and store and for how long and where and so on. Which is not really usable for people which are working on projects with security concerns, for ex. in research and so on I think. Or is my conclusion here wrong as I missed or misinterpreted important information?

An explanatory answer would be much appreciated and would increase my trust rg. these tools massively!

Kind regards
Keno

[iufor]

Hello
When using tools like GitHub Copilot or Microsoft Copilot, parts of your code may be uploaded to the cloud, but it typically depends on the tool's privacy policy and data handling procedures. Some services mention that snippets of code are used to improve their models, but they don’t usually specify what is shared or for how long. You can review the official terms and privacy policies to understand what data is collected. Tools like Copilot are designed to provide assistance without storing full projects, but it's crucial to confirm these details through official documentation.

[finewiki]

Hello Keno,

I understand some of your concerns about how GitHub Copilot and other AI-based coding tools work. These tools generally analyze the code you write and provide suggestions, but it is important to understand what happens behind the scenes. Below, you'll find a general explanation of how these tools use data and some information on how to approach security concerns.

1 - Code Not Being Uploaded or Stored in the Cloud
Tools like GitHub Copilot do not directly upload the code you write to the "cloud." Instead, these tools process the code locally on your machine and then interact with large language models used to generate suggestions. The language models that GitHub Copilot works with do not use the user's code or questions as "training data," but the model's suggestions are derived from more general and broadly trained datasets.

2 - How to Learn Where Your Code Is Used and How Long It Is Stored
GitHub is a platform with a privacy policy that governs how user data is handled. You can review this policy on GitHub's official site. GitHub uses only temporary data when generating suggestions, and this data is mostly processed temporarily just to provide suggestions. Data provided by users does not contain personally identifiable information and is generally anonymized. However, depending on the dataset used and the type of model, there may be occasional tracking of the data. GitHub Copilot and similar tools have very limited data storage other than temporarily analyzing code in your local software development environment to provide suggestions. In other words, most user data is only used temporarily for suggestion creation.

You can review GitHub's privacy policy and learn more about Copilot's data usage through the following links:

GitHub Privacy Policy: https://docs.github.com/en/github/site-policy/github-privacy-statement
GitHub Copilot Data Usage Explanation: https://docs.github.com/en/github/copilot/copilot-privacy-statement

[KenoAlbrecht]

Honest question: how do you know? In your Github profile you state you are working at Google, not at Microsoft.
Also your second link does not work for me (Error 404). Did you just copy an answer generated with Microsoft Copilot or another LLM tool? That would explain the broken link and would also fit the overall writing style from my experience using these tools.

Maybe you could clarify this!

[github-actions[bot]]

🕒 Discussion Activity Reminder 🕒

This Discussion has been labeled as dormant by an automated system for having no activity in the last 60 days. Please consider one the following actions:

1️⃣ Close as Out of Date: If the topic is no longer relevant, close the Discussion as out of date at the bottom of the page.

2️⃣ Provide More Information: Share additional details or context — or let the community know if you've found a solution on your own.

3️⃣ Mark a Reply as Answer: If your question has been answered by a reply, mark the most helpful reply as the solution.

Note: This dormant notification will only apply to Discussions with the Question label. To learn more, see our recent announcement.

Thank you for helping bring this Discussion to a resolution! 💬

[KenoAlbrecht]

Are there any news from an official Github/Microsoft employee? Or is this a valuable information that the users should not know? If so, why? This implies for me that everything I type into this thing and all the code this technology (Github Copilot) gets uploaded to Github/Microsoft servers in order to either train their AI models or just to steal/use/store it for the sake of it. Even though it was stored privately on a local machine and was not meant to be shared, only for local development.

Or that they don't know which parts are used and uploaded which would be even worse and absolutely crazy.
I really do like this technology but why is it so hard, to get an honest and easy to understand, official answer? This does really derogate my trust in you as a corporation and is not a desirable goal in my opinion, especially for a technology as powerful as this!

Please give me an answer! This is not acceptable!

[KenoAlbrecht]

How do I remove the label "inactive"? This should be actively discussed or answered from someone official, that really knows whats going on there internally, respectively!