Key takeaways
- On Apple’s earnings call, John Ternus said AI agents are “promising tools,” but Apple first wants users to know what they give access to and to protect security and privacy. Siri AI stays “primarily request-based today,” with no date for more.
- The next day, Apple released the iOS 27.3 beta with new Siri “app skills”: Siri can find something in one app, recognize your partner in Contacts and send it to them through another app. Multi-step actions across apps.
- My view: security is not the reason, it’s a convenient excuse for laziness and a technology stall. And I would never give an AI agent access to my money or my email.
On November 2, on Apple’s fiscal fourth-quarter earnings call, two analysts asked CEO John Ternus about AI agents: assistants that don’t just answer, but do tasks for you across apps, such as sending, booking or buying. Morgan Stanley’s Erik Woodring asked about the shift “from chatbots to consumer agents,” and TD Cowen’s Krish Sankar pointed to growing competition from app-based agents. Ternus answered that agentic capabilities are “promising tools for users,” but that Apple wants “to make sure that users are aware of what they are giving access to” and that “this is done in a way that protects security and privacy.” He called Siri AI “primarily request-based today,” built on an architecture with “room to expand capabilities over time,” and gave no timeline.
On November 3, Apple released the first developer beta of iOS 27.3. It adds Siri “app skills,” which let any app teach Siri to coordinate multi-step actions. Apple’s own example: “Hey Siri, send my favorite sneakers to my partner” in a messaging app, where Siri finds the sneakers, works out who your partner is in Contacts and sends them. The beta also lets Siri create playlists, calendar events with invitees, schedule messages, and manage and delete files. Siri AI runs on models built with Google’s Gemini on Apple’s Private Cloud Compute servers, and Apple has said server-side features will get usage limits, with paid expanded access later.
| What Apple said | What Apple shipped the next day |
|---|---|
| Agents are “promising,” but users must know what they give access to | Siri “app skills” for multi-step actions across apps (iOS 27.3 beta) |
| “This is done in a way that protects security and privacy” | Siri finds content, picks a contact and sends it through another app |
| Siri is “primarily request-based today” | Siri can delete files, schedule and unsend messages, create events with invitees |
| No timeline for agents | iOS 27.3 public release expected later this year |
In this article: A slap in the face · Privacy as a cover story · A security lecture in the worst possible week · Not ready? Then what is this in the iOS 27.3 beta? · And some Apple sites applaud · Security is not the reason. It’s the excuse · What I would do with an AI agent · FAQ
A slap in the face
What is this supposed to mean? It’s a mockery, right in everyone’s face, and their shameless arrogance makes my blood boil. What kind of farce are they putting on in Cupertino? This is really over the line, and their endless corporate babble is starting to give me a headache.
Privacy as a cover story
For years, they fed us polished fairy tales about how privacy comes first for them. But the truth is somewhere else. Anyone with a bit of common sense can see that it’s just an embarrassing alibi for their own incompetence.
Others are releasing one agent after another and pushing the technology forward with giant steps, while Apple calmly tells us we have to wait, because Apple does it “safely.” What kind of nonsense is that? It’s a complete disgrace.
A security lecture in the worst possible week
And do you know what drives me crazy the most? Their boundless arrogance. In the same week that more than 400 dirty deepfake apps got through their “perfect” App Store, and in the same week they warned people in 99 countries about dangerous spyware, their famous Ternus reads us a sermon about security from a piece of paper.
Do they think we all fell off a tree? It’s infuriating, it’s pathetic, and honestly, this hypocritical theater is starting to make me physically sick.
Not ready? Then what is this in the iOS 27.3 beta?
And did you notice the outrage in their own iOS 27.3 beta? On one side, they say on the record that agents are not ready yet. On the other side, the very next day, Apple publicly releases a beta where Siri searches your apps by itself, recognizes your partner and sends things to her.
Apple would say this is still “request-based,” because you have to ask. Fine. But Siri now chains steps across apps, picks the person and sends the content. So which is it? Are they lying straight to our faces, or does the right hand not know what the left is doing? Either way, in my opinion, it’s a mess on wheels.
And some Apple sites applaud
And as if that weren’t enough, some Apple sites still applaud. AppleInsider even called the criticism of Ternus reading from a script during the earnings call “pathetic.” To me, something else is pathetic: a CEO who reads phrases about security while his company ships the exact thing he says isn’t ready.
Security is not the reason. It’s the excuse
Frankly, I’ve had enough of their sweet little speeches and corporate excuses. You buy an expensive phone, and instead of proper modern technology, you get a pile of talk about why something can’t be done.
Security is not the reason. In Apple’s hands, security is just a convenient excuse for laziness and a technology stall. And if they think we will keep swallowing this amateurism forever, they are badly mistaken. People are fed up with this dirty comedy.
What I would do with an AI agent
I would certainly never give an AI agent access to my money or my emails. For me, that’s clear. Whether it’s Siri, ChatGPT, Gemini or anything else.
And I’m 100% sure that when Apple’s agents come, they will be paid. Apple is already crying that servers are expensive: it told investors that memory costs and server-side Siri AI will weigh on its margins this quarter, and that Siri AI will get usage limits with paid expanded access. Ternus himself called “intelligence” another opportunity for its services business.
As for Siri running on Google’s brain, in my opinion people have known that for a long time, it’s been trumpeted everywhere. Apple just doesn’t say it anywhere on the iPhone itself.
Also on Evolt News: Siri with limits and a fee · Siri AI in six languages · HomePod mini 2 and the Siri AI fine print
Featured image: appshunter.io / Unsplash.
This article is opinion.
Siri AI agents: FAQ
Will Siri AI get agentic features?
Apple says it wants to. John Ternus called agentic capabilities “promising tools for users,” but said Apple wants users to understand what they give access to and to protect security and privacy first. He gave no timeline.
What are Siri app skills in iOS 27.3?
App skills let third-party apps teach Siri AI to handle complex, multi-step requests across apps. Apple’s example is Siri finding favorite sneakers in one app, identifying your partner in Contacts and sending them in a messaging app. It is in the iOS 27.3 developer beta released on November 3, 2026.
What is an AI agent?
An AI agent is an assistant that performs tasks for you, such as navigating apps or websites, filling in forms, sending messages or making purchases, instead of only answering questions.
Is Siri AI powered by Google Gemini?
Yes. Siri AI uses models built with Google’s Gemini, running on Apple’s Private Cloud Compute servers, under a partnership announced in January 2026.
Will Siri AI cost money?
Apple said server-side Siri AI features will have usage limits and that users will be able to buy expanded access in the future. It has not announced prices.
Is it safe to give an AI agent access to email and payments?
Agents can make mistakes and can be manipulated by malicious text on websites or in emails, an attack called prompt injection. A safe approach is to never give an agent payment and email access at once and to confirm every send or purchase yourself.
Sources: Apple (Q4 2026 earnings call, developer documentation), 9to5Mac, AppleInsider, Fortune, FourWeekMBA.
Our Take
John Ternus says AI agents must wait for security and privacy. The next day, Apple ships a beta where Siri chains actions across apps and sends things to your contacts. In my view, security is not the reason, it’s a convenient cover for a company that missed the moment.
In the same week as 400 deepfake apps and spyware warnings in 99 countries, a security sermon from a script is hard to take seriously. I would never give an AI agent my money or my email, and I’m sure Apple’s agents will cost extra.
