Apple continued its gradual integration of artificial intelligence into its devices during its WWDC event, unveiling a range of new features and upgrades. The tech giant introduced the Foundation Models framework, enabling developers to tap into Apple’s AI capabilities.
Among the standout AI announcements was Live Translation, a real-time language translation feature for phone and FaceTime calls. Apple also showcased Workout Buddy, an AI-powered voice assistant designed to provide encouragement and updates during workouts. In a demo video, Workout Buddy told a jogging woman, “This is your second run this week. You’re crushing it.”
Other notable updates included enhancements to Visual Intelligence, a tool that interprets the world through a device’s camera. The new version can analyze screenshots to identify products or summarize webpages. Apple also improved Genmoji and Image Playground, tools that generate stylized images using AI.
While Apple’s AI announcements were seen as incremental, some analysts believe this measured approach is justified. “The jury is still out on whether users are gravitating towards a particular phone for AI-driven features,” said Paolo Pescatore, an analyst at PP Foresight. “Apple needs to strike the fine balance of bringing something fresh and not frustrating its loyal core base of users.”
Francisco Jeronimo, an analyst at IDC, noted that making Apple’s AI models accessible to developers is significant due to the company’s vast reach among coders. This move “brings Apple closer to the kind of AI tools that competitors such as OpenAI, Google, and Meta have been offering for some time,” Jeronimo said.
One advantage of Apple’s AI models is that they run on personal devices, eliminating the need for network connections and associated fees. The company also introduced Private Cloud Compute, a secure way for developers to use cloud models while protecting private data.
However, Apple may need to make more significant strides in AI in the future, as competitors explore how the technology could revolutionize personal computing. Both Google and OpenAI have demonstrated advanced AI helpers that can engage in real-time conversations and interpret their surroundings through device cameras.
Despite currently lagging behind in advanced AI development, Apple is actively publishing AI research. A recent paper highlighted significant limitations in today’s most advanced AI models, particularly when faced with complex problems. The researchers tested various models on increasingly complex versions of the Tower of Hanoi mathematical puzzle, finding that they succeeded up to a point before failing dramatically.
Subbarao Kambhampati, a professor at Arizona State University, believes Apple’s research reinforces the idea that simulated reasoning approaches in AI need improvement to tackle a wider range of problems. While the findings suggest a more cautious approach to AI is warranted, Kambhampati doesn’t think Apple is being complacent. “If you know what’s going on inside Apple, they’re still pretty gung-ho about LLMs,” he said.
