While Apple has been criticized for the way that Siri has lagged behind, a new report says that this is because the company is aiming to be far more ambitious than existing generative AI systems …
Although the company hasn’t made any announcements about its generative AI work, a report back in the summer said that the company has already created its own Large Language Model (LLM) system, which has been dubbed AppleGPT.
The tool as it stands is only being used internally, but the report then said that “people familiar with the work believe Apple is aiming to make a significant AI-related announcement next year.”
While that report caused enough excitement to create a spike in AAPL stock, we cautioned at the time that customers shouldn’t expect to be using a powerful new Siri service anytime soon.
Apple’s goals are far more ambitious than ChatGPT
While Apple could launch its own LLM tomorrow, the company is aiming to leapfrog systems like ChatGPT in two ways:
- On-device processing
- Going beyond language-processing, to include AI audio and video
AI systems like ChatGPT do all their processing on remote servers, but research papers written by Apple suggest that the company has a far more ambitious objective: using on-device processing.
This would have the advantage of speed and privacy, but enabling such services to run in the limited memory of mobile devices poses a huge challenge – and that is believed to be a key reason for the time the company is taking before launch.
We have already seen the company take one small step toward this goal, with offline Siri launching in the latest Apple Watch models back in September.
AI audio and video, as well as language
ChatGPT is known as an LLM, because it’s all about language processing. But Apple’s ambitions go beyond this, seeking to incorporate AI processing of both video and audio.
We’ve again seen some early signs of this, with visual and audio AI processing seen in Door Detection and Live Captions.
Next generation Siri expected to launch in iOS 18
The Financial Times reports on the expected launch of a far more powerful version of Siri.
Apple’s Worldwide Developers Conference, usually held in June, is widely expected to be the event where the company reveals its latest operating system, iOS 18. Morgan Stanley analysts expect the mobile software will be geared towards enabling generative AI and could include its voice assistant Siri being powered by an LLM […]
“As far as the chips in their devices, they are definitely being more and more geared towards AI going forward from a design and architecture standpoint,” said Dylan Patel, an analyst at semiconductor consulting firm SemiAnalysis.
This echoes a report from an Apple leaker earlier this month, and Morgan Stanley points to Apple job listings as further evidence.
Almost half of Apple’s AI job postings now include the term “Deep Learning”, which relates to the algorithms powering generative AI — models that can spew out humanlike text, audio and code in seconds.
While there’s still a lot of speculation involved, the idea of a major Siri relaunch this year does seem increasingly plausible.
LLM models like ChatGPT have completely transformed expectations of what we now expect from AI systems. While Apple has been able to point to privacy as its reason for keeping Siri’s capabilities more limited than rivals which process lots of personal data, we have now reached a point where Siri looks unacceptably dumb. That has to change, and it’s hard to see how Apple could wait any longer than this year.
Using generative AI to power Siri will dramatically boost the capabilities of Apple’s intelligent assistant, while on-device processing will enable the company to continue to tout its privacy credentials. Incorporating video and audio processing gives the company a real shot at persuading everyone that it was worth the wait.
FTC: We use income earning auto affiliate links. More.
The post Next generation Siri likely to be launched in iOS 18, ChatGPT style first appeared on 9to5mac.com