According to Bloomberg, Apple plans to launch a generative AI overhaul of Siri next year, utilizing a chatbot iPhone app currently in development for testing the upcoming Siri large language model.
This new software resembles a ChatGPT-like iPhone app but is not intended for consumer use. Instead, it serves to evaluate the enhanced capabilities of a reimagined, AI-powered Siri.
Veritas: Testing the AI Behind the New Siri
Bloomberg reports that the new iPhone app will test whether the AI-enhanced Siri can:
- Search personal data and content, including music and emails.
- Edit photos and perform other in-app actions.
To fully test the new Siri, Apple needed a version deployable in real-world environments, such as on an iPhone. The new tool, codenamed Veritas, makes this possible.
Microsoft's Recall feature faced strong backlash from privacy and security experts when introduced as part of Windows 11. Similarly, Google’s Gemini on Android offers features like Circle to Search and AI Overview.
Reportedly, Apple’s AI leverages both an internally developed LLM from its foundational model team and an external model.
The AI Race for More Helpful Agents
AI companies are competing to capture the market for intuitive agents that anticipate user needs. OpenAI has taken steps toward this goal with ChatGPT Pulse, a morning summary based on user schedules and chat history.
Apple’s delayed AI rollout can be seen either as a missed opportunity or as a strategic move to avoid the early pitfalls faced by Microsoft’s Copilot and other AI tools. Meanwhile, Google and Samsung have embraced generative AI more aggressively in smartphones, while Apple’s progress in this space has been relatively incremental.
Bloomberg also reports that the development of the new Siri has faced setbacks. Apple originally intended to unveil the updated Siri last spring, but engineering challenges caused features to fail in up to one-third of cases.
