Tim Cook is “not 100 percent” sure that Apple can stop AI from having delusions.

Tim Cook, the CEO of Apple, isn’t even sure that the company can fully stop AI from having dreams. Apple CEO Tim Cook told The Washington Post that he would “never claim” that the company’s new Apple Intelligence system wouldn’t give out false or incorrect information.

Cook says, “I think we’ve done everything we know how to do, including giving a lot of thought to how ready the technology is in the places we’re using it.” “That’s why I’m sure it will be very good.” But to be honest, I’d say that’s not totally 100%. That being said, I would never say “100%.”

Apple showed off its new Apple Intelligence system at its Worldwide Developers Conference on Monday. The iPhone, iPad, and Mac will all be able to use AI features. With these features, you can write email replies, make your emoji, recap text, and more.

The same thing can happen with this AI system as with any other AI system: dreams are possible. Recent examples of AI going wrong are Google’s Gemini-powered AI overviews telling us to use glue to put cheese on pizza last month, and a bug in ChatGPT that made it give us answers that made no sense.

Apple also said that it will work with OpenAI to add ChatGPT to Siri. When it gets a question that would be better answered by a robot, the voice helper will use ChatGPT. But it will ask you for permission first. For the feature that was shown at WWDC, there is a note at the end of the answer that says, “Check important info for mistakes.”

The CEO of Apple, Tim Cook, said that the company picked OpenAI because it is a “pioneer” in privacy and has “the best model” right now. Apple might not just work with OpenAI in the future either. Cook replied, “We’re also integrating with other people.” In a live discussion on Monday after Keynote, Craig Federighi, senior vice president of Apple, said that Google Gemini might one day be added to iOS as well.