With the OS 27.3 updates, developers will finally be able to make Siri accessible to their apps, so Siri can perform actions within the app and access your data, if desired. Along with on-screen awareness and personal context, it’s the latest pillar of the advanced capabilities Apple first announced two years ago.
But the latest trend in AI is agents. What started with hobbyist software like OpenClaw has evolved into consumer-ready products like Meta’s Muse, OpenAI’s Dots, and Anthropic’s Claude Duet. These AI agents run continuously and somewhat autonomously, taking care of almost anything a person can do online, on their phone or computer. There are some concerns about privacy and security, and rightly so, but they have already proven to be very popular.
That’s a big step beyond “telling Siri to perform actions within apps” and is where Ternus would like to do things. Gurman said Apple could make Siri act more like an autonomous agent “as early as next year,” noting that the company has no choice: Announcing such a major feature at WWDC in June would put Apple a year behind the major players in the AI space, at least.
Gurman’s report paints a picture of a CEO concerned not only with matching the characteristics of AI companies, but also with recognizing the growing role of AI in computer use in general. The general idea is that, over time, more of what we do on our computing devices will not be handled by typing, clicking, and swiping, but by a personal agent who can essentially use any software we want based on our instructions. If you’re a company like Apple whose strength comes from selling physical computing devices and creating operating systems that run on them, ceding that level of control to a rival company could be an existential threat.
