What Apple’s New Siri AI Is and Why It Matters
Apple’s new Siri AI is an upgraded version of its assistant on macOS that combines Apple Intelligence features with chatbot‑style conversations, aiming to answer general questions, work with your files, and control system settings while staying integrated across your devices. Available in a developer beta via waitlist on supported Macs, this version of Siri is designed to be more conversational, more responsive, and less error‑prone than the old one, while still acting as a focused productivity assistant rather than a chat companion. In practice, that means Siri behaves closer to tools like ChatGPT and Gemini for many text requests but remains tightly tied to macOS features, from Spotlight search to context menus. As Apple experiments with a framework for third-party AI models behind Siri, this upgrade is also a hint of a broader shift toward user choice in which model actually handles your requests.
A New Framework for Third-Party AI Models in Siri
Behind the scenes, Apple has built an internal framework so Siri can plug into third-party AI models, including ChatGPT, Gemini, and others, instead of only relying on Apple’s own model. This means that in future macOS releases, Siri could act as a front-end where you choose which AI engine answers a given request, without leaving Apple’s interface. While details remain limited, the reported framework points to a platform approach: Siri controls permissions, on-device context, and UI, while different models provide the language intelligence. According to iClarified, Apple’s framework is designed for “third-party AI models in Siri,” suggesting Apple expects multiple providers rather than a single default. For users, that hints at less lock-in and more flexibility: you might use Apple Intelligence for on-device tasks, a model like ChatGPT for creative writing, and Gemini for web-heavy research, all inside the same assistant flow.
Siri AI Performance on macOS: General Knowledge and Search
In macOS Siri testing based on the same prompts used for ChatGPT and Gemini, Siri AI shows clear progress but also gaps. For broad questions like “Why did the Roman Empire fall?”, Siri gives a concise spoken summary and bullet-point causes, then cites online sources with links. That behavior is similar in usefulness to other chatbots, though Siri’s tone is less chatty and more like a direct briefing. However, when asked “What laptop should I buy?” with specific preferences, Siri first responded with a list of links instead of a clear recommendation. A follow-up request asking for a summary and opinion produced a better, more advisory answer. Compared with ChatGPT and Gemini, the Siri AI performance feels more constrained: it can answer many knowledge questions well, but it often needs nudging to move from link aggregation to confident, synthesized advice.
macOS Siri Testing: Local Files, Photos, and App Control
Where Siri stands out against ChatGPT and Gemini is its tight link with macOS. You can trigger it by voice, via a dedicated Siri AI app, with a double-Command shortcut, through Spotlight, or by right-clicking to Ask Siri on an item. In theory, that makes Siri ideal for managing local content: calendars, settings, and files. In practice, testing shows mixed Siri AI performance. It can display appointments or respond to natural-language questions, but its file and photo search is unreliable. When asked to find all photos of the Abraham Lincoln statue in the Photos library, Siri surfaced three images even though six matched. Similar partial results appeared with other photo queries, suggesting inconsistent object and scene recognition. For tasks like identifying paintings from uploaded images, Siri sometimes mislabels both work and artist, while only occasionally matching popular pieces correctly.
ChatGPT vs Siri vs Gemini: Conversation, Accuracy, and the Road Ahead
Compared with ChatGPT and Gemini, the new Siri feels more like an assistant glued to macOS than a free-form conversational partner. It offers helpful advice in practical scenarios, such as suggesting food options for a picky cat and asking follow-up questions, but conversation flow can be awkward. Instead of keeping a dialogue open, Siri often closes the interaction after each answer, forcing you to restart context or repeat details that other chatbots maintain more smoothly. Accuracy is also uneven: news summaries and short explainers are solid, while photo recognition and some specific lookups still lag. Still, the combination of deep system hooks and an emerging framework for third-party AI models hints at a significant shift. Rather than walling users into one engine, Apple seems to be building Siri into a unified front-end where you can pick the model that best fits each task.
