Apple prohibits employees from using ChatGPT for fear of data leakage

Apple has banned employees from using AI tools such as OpenAI’s ChatGPT over concerns that confidential information fed into these systems could be leaked or collected.

according to report from The Wall Street JournalApple employees have also been warned against using GitHub’s AI programming assistant. bloomberg Reporter Mark Gorman chirp that ChatGPT had been on Apple’s list of banned software for “months”.

Apple has good reason to be concerned. By default, OpenAI caches all interactions between users and ChatGPT. These conversations are collected to train OpenAI systems and can be screened by moderators to break the company’s terms and services.

In April, OpenAI launched a feature that lets users turn off chat history (coincidentally, not long after several EU countries began investigating the tool for potential privacy violations), but even with this setting enabled, OpenAI still keeps chats for 30 minutes. days with the option to review them “for abuse” before permanently deleting them.

Given ChatGPT’s usefulness for tasks like optimizing code and brainstorming ideas, Apple may be right to be concerned about its employees entering information about confidential projects into the system. This information can then be viewed by an OpenAI moderator. Research shows it’s possible, too Extracting training data from some language models using its own chat interface, although there is no evidence that ChatGPT itself is vulnerable to such attacks.

However, Apple’s ban is noteworthy given that OpenAI launched an iOS app for ChatGPT this week. The app is free to use, supports voice input, and is available in the United States. OpenAI says it will release the app in other countries soon, along with the Android version.