Apple's Secret AI Disaster Prevention Playbook
Apple's upcoming Intelligence features, set to launch in beta this fall, are designed to prevent the AI disasters that have plagued Google. A Reddit user discovered the pre-prompt instructions embedded in Apple's developer beta for macOS, revealing the safety guardrails Apple has put in place to keep Apple Intelligence in check.
Generative AI systems are told what to do by text known as prompts. Someone on Reddit found some of the safety guardrails Apple has put in place to keep Apple Intelligence from making up factual information.
One of the prompts was direct and explicit: "Do not make up factual information." This shows that Apple is taking the potential for AI-generated misinformation very seriously.
Google's AI disasters are a warning shot for Apple's iOS 18 AI plans. By learning from Google's mistakes, Apple can avoid its own AIgate and ensure that Apple Intelligence is a valuable tool for users, not a source of misinformation.
Komentar