Apple’s new AI-powered Siri may launch as beta feature with limited access: Report
Follow UsApple is internally treating the new Siri as a "beta" and "preview," meaning it will not be marketed as finished software
The new Siri is powered by Google's Gemini model and likely hosted partly on Google servers
Many of the new Siri features being announced on Monday were first promised in 2024
Apple’s biggest WWDC keynote in years takes place on Monday and the centrepiece is a ground-up overhaul of Siri. The assistant is being transformed from a voice command shortcut into a full AI companion capable of handling complex tasks across iOS, iPadOS and macOS. But ahead of the announcement, Bloomberg’s Mark Gurman has reported a detail that tempers expectations considerably. He claims Apple is labelling the new Siri as a “beta” and “preview” internally, which suggests it will not be presented as finished software when it ships with iOS 27 later this year. There is also a possibility of a waitlist for certain features, similar to how Apple Intelligence rolled out in 2024. The original Siri carried the same “beta” label for two years after its 2011 launch.
SurveyWhat the new Siri can do
The feature list for what new Siri can do is ambitious. Siri is becoming a full chatbot in the style of ChatGPT and Gemini, with a dedicated app that holds past conversations, syncs them across devices via iCloud and supports auto-deleting chats after 30 days, a year or indefinitely. The conversational interface looks like an iMessage thread and supports both text and voice input, as well as photo and file attachments.
The assistant will live inside the Dynamic Island for voice queries, opening with a new animation when activated. Swiping down from the centre of the display will bring up a ‘Search or Ask’ panel, which also incorporates Siri Suggestions. This panel lets users toggle between Siri and third-party chatbots including Claude, Gemini and ChatGPT. Apple is also launching its own in-house web search product as part of Siri, competing directly with Perplexity AI, presenting results with summaries, bullet points and images rather than redirecting users to Google.
A few more capabilities include handling multiple commands in a single prompt, composing emails using data pulled from a user’s calendar, contacts and past messages and a ‘Write with Siri’ option accessible from a dedicated keyboard button. Visual Intelligence is also moving from the Camera Control button to a dedicated Siri mode within the Camera app, making it easier to scan nutrition labels and read contact information from business cards.
Apple is going into WWDC knowing that what it is about to announce is not complete. Several features being unveiled on Monday were originally shown at WWDC 2024 and then repeatedly delayed. The personal context and onscreen awareness capabilities, which let Siri access a user’s emails, messages and files and see what is on screen, were promised two years ago, but they are arriving only now and that too in beta form.
Apple is also using a Gemini model from Google to power the new Siri and a significant portion of processing is taking place on Google’s servers. For a company that has built much of its brand identity around on-device processing and user privacy, it has not explained publicly how it intends to address the gap between its privacy commitments and the reality of routing Siri queries through a competitor’s infrastructure.
WWDC 2026 kicks off on Monday and that’s when we expect Apple to give us some answers.
