Source: Hard AI

Apple, which has been leading the trend of smart phones since the iPhone 4 era, has now fallen into a downturn in the AI arms race.

Based on the principle of joining if you can't beat it, Apple has reached an agreement with OpenAI to incorporate OpenAI's generative AI technology into the underlying iOS 18 technology to fully upgrade Siri's conversation experience.

According to news from well-known tech journalist Mark Gurman, Apple will release an improved Siri at the WWDC conference on June 10. The new Siri is more conversational and more versatile, and can evolve Siri into a chat assistant instead of the current one; it can only answer one question at a time.

In the future, Apple expects Siri to be an “autonomous intelligent” assistant to help users handle all aspects of their daily lives.

Gurman also added that Apple's AI strategy will begin with chatbots and spread all the way to search engines and software and hardware updates.

Apple's transformation to AI was “forced”

And the breakthrough point of Apple's AI strategy this time is probably Apple's voice assistant Siri, which has been molested for many years.

Remember when Apple unexpectedly cancelled Apple Cars at the beginning of this year?

It's about preparing ahead of time for Siri to evolve. After not building cars, Apple transferred most of its engineers to the AI team to concentrate on major tasks. According to reports, Apple burns millions of dollars every day to upgrade Siri, and is about to turn the “old urchin” Siri, which has been around for 10 years, into its own AI fist product.

Furthermore, Apple is also in negotiations with Google, and may also introduce Gemini to further strengthen Siri, but the agreement has yet to be finalized.

However, Apple's recent frequent moves to transform AI are all due to too much pressure from rivals.

In particular, OpenAI and Google, which are now in the first tier of AI, recently updated their strongest AI creations. GPT-4O, which was launched by OpenAI last week, has evolved to the point where it can cosplay human conversations, prepare job interviews for users, and even sarcasm and make fun of users. I have to say it's a bit aggressive.

Meanwhile, at the Google I/O developer conference last week, Google also upgraded the AI model Gemini, and also deeply integrated AI into the search engine, tinkering with AI functions such as generating summaries, advance planning, and video scanning and searching.

In contrast, Apple can be said to be unable to do anything in the AI field. With the exception of the iPad Pro (M4 chip equipped with an NPU neural network engine) launched in early May, it can be said that progress is slow.

This also put a lot of pressure on Apple CEO Cook. If Apple wants to catch up with its rivals in this wave of AI competitions, it must change its strategy.

In addition to chatbots, Apple's AI strategy also has to balance the speed of search engines and software and hardware updates

However, cooperation with rivals on AI is only a mitigating measure for Apple.

Gurman pointed out that if it wants to go further in the AI wave, Apple will eventually have to get rid of cooperative methods and focus on building its own AI chatbot, then deeply integrate it into its own products, then increase the speed of technology iteration, and finally have a chance to overtake corners.

Although it is not possible to quickly build your own chatbot with Apple's current technology, Apple executives are also afraid that chatbots built in a hurry will have all kinds of bugs, just like the rollover incident where Google first launched Gemini. This kind of accident will damage Apple's reputation.

The good news is that for Apple, which has a well-developed family and has the advantage of talent and an ecological platform, it has a lot of momentum.

Gurman said that next, Apple's AI strategy will fully focus on Siri, online search, system update speed, and hardware updates.

Over the years, Apple has always wanted to build its own search engine, a search engine with user privacy protection functions, not like Google or Bing Search, but this would be very expensive, and it would directly target Google, the big guy that already focuses on AI search.

Moreover, starting in 2022, Google has paid Apple billions of dollars every year, making the Google search engine the default search engine for Apple's Safari browser. Apple can easily earn this license fee, which in turn reduces the motivation to build an independent search engine.

Today, the development of AI has forced Apple to make more changes, because in the long run, if Apple has its own AI search engine, it will have more potential for development, depending on Apple's next choices.

Finally, there is the issue of the speed of system updates. Apple basically makes a major update to the iOS system every year, but currently AI is developing too fast, and the annual update rate seems too slow.

Gurman believes that if Apple needs to catch up with the first tier of AI, it needs to speed up the update speed of iOS systems that integrate generative AI technology, but currently Apple's focus does not seem to be on speeding up system updates.

Fortunately, Apple is quite keen on hardware upgrades. At the beginning of this month, Apple launched a new iPad Pro equipped with an M4 chip. One of the highlights of the M4 is that it has an NPU (Neural Network Engine). Apple expects that by 2025, the M4 chip will enter every Mac product to fully upgrade the product line. In September of this year, the new iPhone 16 Pro will also make another leap forward in chips.

Apple still has a long way to go before transforming AI.

Just as Jobs withstood all kinds of pressure back then to launch the iPhone 4, but now the pressure is on Cook, can he seize the AI era belonging to Apple?

edit/ruby