The rapid advancement of artificial intelligence systems has resulted in some peculiar outcomes, such as user queries resulting in putting glue in pizza recipes and AI affirming cats on the moon. Apple’s CEO admits that even their proprietary AI system isn’t entirely free from such ‘hallucinations’.
In an interview with The Washington Post, Apple’s CEO Tim Cook confirmed that it’d be impossible to assure with absolute certainty that Apple’s AI system won’t produce false or misleading information. “I believe we’ve done everything necessary, including deep deliberation on the readiness of this technology in various fields, so I’m quite confident about its high quality. However, frankly, this isn’t 100% certainty. I would never claim this assurance to be 100%,” Cook explained.
Discussing the selection of OpenAI as the inaugural partner for integrating a chatbot with Apple’s voice assistant Siri, Cook stated that OpenAI is a ‘pioneer’ in protecting user privacy and currently has ‘the best model’ at their disposal. Both Cook and Apple’s Senior Vice President Craig Federighi hinted at ongoing talks with other AI system developers for potential integration. Notably, Federighi mentioned Google Gemini in this context, albeit as an example.
