Is This the Reason Behind Apple’s Siri Problems?
Credit: keBu.Medien / Adobe Stock
Toggle Dark Mode
We’ve written quite a bit recently about Apple’s struggles with Siri not living up to the company’s promises and falling short of the exciting upgrades expected with Apple Intelligence. A little exploration of Siri’s history could shed some light on the heart of Siri’s shortcomings. It’s an interesting crossroads for both Apple and its users alike. The public’s disappointment with Siri could be the result of Apple maintaining strict adherence to one of its core values: user privacy.
Many of you may not know that Siri was not an original Apple application. It was developed by SRI International (Stanford Research Institute) and spun into a startup called Siri, Inc. Siri launched as an iPhone app before Apple acquired it in April 2010.
According to one of Siri’s founders, it was one of the last deals Apple’s late CEO, Steve Jobs, worked on. While there are no official details on the transaction, multiple publications believe Apple paid around $200 million for Siri.
With the release of Apple Intelligence, Apple promised a Siri that’s aware of your “personal context” and can help you in “ways that are unique to you…without compromising your privacy.” However, privacy could be the rub. Apple has repeatedly highlighted that Apple Intelligence runs on your device. However, there’s only so much that can happen on-device without obviously falling short of the capabilities of other chatbots like Alexa, Grok, and even ChatGPT.
The Siri Apple promised will run on-device and via Apple’s Private Cloud Compute servers. It’s unclear how far along Apple is with its server infrastructure or what its role will be in processing personal information stored on your device — a key aspect of Apple’s “personal context” improvements.
As of iOS 18.2, Siri added support for ChatGPT integration as an “extension.” This allows more complicated requests that require “world knowledge” to be sent to OpenAI’s servers, which are much better equipped to handle these since they’re capable of an estimated 600 times more parameters than the on-the-iPhone AI and have better access to the internet at large.
However, this is a separate process that’s merely brokered by Siri; iPhone users must grant Apple consent before these queries are sent to OpenAI’s servers, and ChatGPT’s responses are simply dictated back to you by Siri. The only place “Apple Intelligence” factors in is deciding whether to forward your request to ChatGPT or try to answer it on its own — and it hasn’t been doing a very good job.
If Apple is going to truly deliver on its promises of a new Siri backed by Apple Intelligence, will users ultimately need to compromise privacy and security?
The Siri issue also begs the question as to whether or not iPhone users care about Siri or not. There are plenty of third-party chatbots available. Yes, Apple may have overpromised, but will there be a meaningful fallout in Apple’s bottom line? There’s little doubt that Apple is capable of making Siri far better. Do they need to, and are users willing to give up their privacy in exchange? We’ll find out.
