Apple’s own research sheds light on Siri’s AI laggardness | Digital Trends

With the introduction of the new iPad Mini, Apple made it clear that a software experience brimming with AI is the way forward. And if that meant making the same kind of internal upgrades to a tablet that costs nearly half as much as its flagship phone, the company would still march forward.

However, its ambitions with Apple Intelligence lack competitive vigor, and even by Apple’s own standards, the experience hasn’t managed to wow users. On top of that, the staggered rollout of the most ambitious AI features — many of which are still in the future — has left enthusiasts with a bad taste.

Now, it appears that the reason behind the delays has something to do with quality and performance, as per Apple’s own testing. “The research found that OpenAI’s ChatGPT was 25% more accurate than Apple’s Siri, and able to answer 30% more questions,” says a Bloomberg report.

To recall, Apple’s position with Siri is quite unique. For example, Siri is getting enhanced natural language understanding and deeper integration with apps as well as local files. However, there are tasks it can’t quite accomplish, and for such situations, the queries will be seamlessly offloaded to ChatGPT.

That’s part of a deal Apple inked with OpenAI. Now, it would make sense that Siri can’t quite pull the same kind of internet-connected tasks as ChatGPT, primarily because Siri and ChatGPT are two entirely different products. However, Apple is deploying OpenAI’s tech stack in more places than just Siri.

According to OpenAI, the ChatGPT will also lend a hand to users with “image and document understanding.” The Writing Tools – which have already arrived in tools like Notes and Safari — are also tapping into the ChatGPT kitty. Moreover, image generation will also be handled by OpenAI’s tech.

With such deep reliance on ChatGPT, one might think that’s because Apple isn’t quite there on the leaderboard with its own AI tech stack, something that could rival the likes of Google’s Gemini or Meta. That assumption won’t be entirely implausible, and even Apple’s team seems to agree with the status quo.

“In fact, some at Apple believe that its generative AI technology — at least, so far — is more than two years behind the industry leaders,” adds the Bloomberg report. Yet, it’s not merely about advancements, but also the pace of rollout.

Take a look at Galaxy AI, Samsung’s take on an AI ecosystem that has already appeared on a wide array of its phones and computing machines, with some help from Google’s Gemini stack. Chinese smartphone makers have already been offering generative AI features like image generation and a next-gen assistant for a while now.

At this point in time, it seems almost certain that Apple’s strategy with Apple Intelligence was hurried, apparently in a bid to quell investor concerns that the company was lagging in the AI race. So far, whatever little we have seen from Apple’s “AI revolution” has been far from revolutionary.

The best implementation of Apple Intelligence so far has been notification summaries and prioritization, but those are more utilitarian features than something that would reimagine the software experience for users. It would be interesting to see how Apple injects fresh energy into its AI approach next year.

But so far, the company hasn’t made any such announcements, and even the promises it made at its developers conference earlier this year are yet to materialize.