TECHNOLOGY

Apple's AI Siri: The Road to Recovery

Cupertino, California, USAMon May 19 2025
Apple's AI journey has been bumpy, to say the least. The tech giant is now focused on revamping its AI assistant, Siri. This effort is part of a broader strategy to catch up in the AI race. The company's initial approach to AI was cautious, to put it mildly. Top brass were hesitant to invest heavily in AI, as they prefer clear, defined goals. This reluctance slowed down Apple's AI progress. The company also started its AI journey later than competitors. This delay was partly due to the belief that customers weren't keen on AI chatbots. This misjudgment cost Apple valuable time. The integration of new AI features with the old Siri was another misstep. This approach was like trying to fit a square peg into a round hole. It led to a never-ending cycle of fixing one issue only to see three more pop up. The AI chief, John Giannandrea, didn't push hard enough for resources or to address the competition. His laid-back approach didn't align with Apple's aggressive corporate culture. Meanwhile, the marketing team overpromised and underdelivered. They hyped up features that weren't ready, leading to delays and disappointment. Now, Apple is trying to turn things around. The plan involves a complete overhaul of Siri, building it on a new AI engine. This engine aims to make Siri more conversational and better at processing information. Apple is also using iPhones and differential privacy to improve its AI. The idea is to use data from users' emails to train the AI, but in a way that protects privacy. Another proposed improvement is to let Siri browse the web to gather and synthesize data. This would make Siri more like a web search tool, similar to Perplexity. It seems Apple is finally ready to take AI seriously. However, the road to recovery won't be easy. The company has a lot of ground to make up. The success of this revamp will depend on how well Apple can execute its new strategy. Only time will tell if Apple can turn its AI woes around.

questions

    Will the new LLM Siri be able to tell jokes better than the old one, or will it still struggle with humor?
    What are the potential ethical implications of Apple using differential privacy to improve its AI training data?
    If Siri can synthesize data from multiple sources, will it start giving unsolicited life advice?

actions