Apple has cracked down on some workers’ use of ChatGPT and similar artificial intelligence tools because it’s wary of leaks as the iPhone maker produces its own AI-powered tech.
According to documents leaked to the Wall Street Journal, Apple is concerned that workers who use ChatGPT and Microsoft’s GitHub Copilot — which uses AI to generate software code — would reveal confidential information about Apple’s own high-tech projects.
Microsoft also backs OpenAI’s ChatGPT followings its $10 million investment in the chatbot, which uses AI to answer questions and write essays as a human would.
When users type prompts into ChatGPT and GitHub Copilot, the models automatically send data back to their developers, allowing the AI tech to get smarter over time.
When bugs surface, like the one ChatGPT experienced in March, users were able to view someone else’s chat history, including “payment-related information,” according to OpenAI.
Such an event presents an opportunity for an Apple employee to accidentally share sensitive information, the Journal reported.
In late April, ChatGPT added a feature that allows users to turn off their chat history. Conversations with the chat history function disabled won’t be used to train and improve the AI models, and won’t appear in the app’s “history” sidebar.
OpenAI boss Sam Altman warned Congress on Tuesday of the “significant harm to the world” AI-powered platforms can have without proper regulation.
Altman said at a Capitol Hill hearing to the Senate Committee on Privacy, Technology and the Law: “If this technology goes wrong, it can go quite wrong and we want to be vocal about that. We want to work with the government to prevent that from happening.”
He made the statement before Congress as lawmakers have increasingly scrutinized AI tech like Altman’s ultra-popular ChatGPT, and as other employers — including JPMorgan Chase and Verizon — have banned the chatbot’s use.
Amazon already has its own internal AI tool, which it reportedly encourages its engineers to use instead of rival-backed platforms.
Meanwhile, Apple is building out its own language model anyway, a spokeswoman told the Journal, and has seemingly been wary of ChatGPT.
Apple blocked BlueMail from updating in March when it began incorporating ChatGPT into email writing. The tech titan then required BlueMail to raise its age restriction for potential new users to 17 from 4 years old, or implement content filtering, to prevent children from seeing inappropriate content.
However, Apple welcomed a ChatGPT app for iPhones and iPads into the App Store on Thursday. According to the app’s listing, users must be at least 12 years old.
Apple also has its sights set on mixed-reality products, with rumblings of a ski goggle-like headset that allows users to simultaneously experience virtual reality and the physical world set to hit the market later this year for $3,000 a pop.
The high-tech specs will be unveiled by Apple CEO Tim Cook at Apple’s Worldwide Developers Conference on June 5.
It’s designed to be worn all day, though it’s a far cry from Cook’s initial vision of sleek eyeglasses designed to be worn all day, similar to prescription specs.
The high-priced product is a headset instead, and it’ll be powered by a separate, iPhone-size battery pack attached to the goggles by a power cord.
nypost.com