Apple's Siri AI catches up as agentic assistant race moves ahead

9to5Mac said Siri AI's two defining characteristics are that it is finally useful and that it can be painfully slow to answer even simple queries. The report's author wrote that the first time Siri AI impressed them was when they asked how much they had paid for a hotel room for an upcoming trip to San Francisco and how far it was from their first appointment in the city. Siri got both answers exactly right. The author said they have increasingly relied on Siri AI for questions such as the contact information for a plumber who fixed a kitchen sink leak years earlier or the name of a PR person who made an introduction, information stored somewhere in iMessage or email. For the first time in 15 years, the author wrote, they can ask such questions without expecting Siri to fail halfway through.

The report said the problem is that Apple has only now caught up with what AI assistants could do in 2024. Over the past few months, the frontier AI market has shifted toward agentic assistants, with products from OpenAI, Anthropic, Meta and xAI moving beyond answering questions to carrying out tasks on users' behalf. The author subscribes to ChatGPT and uses Codex and ChatGPT Work for everyday tasks involving files and information scattered across a Mac and various cloud services. Those tools can open, create, edit and organize files and interact with websites and services the user is logged into. Each time ChatGPT completes such a task, the report said, the new Siri AI starts to feel like the old Siri again.

9to5Mac said Apple appears to be falling further behind as AI labs improve agentic capabilities, sometimes rolling out major updates weeks apart, while Apple remains largely tied to its annual software release cycle.

The report also identified a limitation in today's agentic assistants that Apple may be well positioned to address. Tools such as Codex, Dots and ChatGPT Work still depend heavily on the user to provide permissions and context, according to the report. They can access files, search emails and interact with websites, but they generally do not know what the user is doing, what the user has been discussing with other people or what the user might need next unless explicitly told. The report argued that the next major leap in usefulness will come not just from improving what these tools can do but from improving their proactive understanding of the user's context in real time.

Apple has recently shown it is willing to capitalize on the trust it has built around privacy, the report said, even if that means testing what users are comfortable sharing with an AI assistant. Last month Apple introduced two Apple Watch features: Siri Recap, which can automatically generate summaries of conversations throughout the day, and Live Rewind, which lets users retrieve the last 15 seconds of a conversation they may have missed. Both features rely on the Apple Watch microphone to capture information about what is happening around the user, potentially giving Siri more information about the user's surroundings.