Apple's software chief decided early last year that Siri, the company's virtual assistant, needed a brain transplant.
The decision came after executives Craig Federighi and John Gianandrea spent several weeks testing OpenAI's new chatbot, ChatGPT. The product uses generative artificial intelligence that can write poetry, create computer code and answer complex questions, making Siri look outdated, said the company, which is not authorized to speak publicly. Two people familiar with the company's work said this.
Siri, introduced in 2011 as the original virtual assistant on every iPhone, was for years limited to individual requests and couldn't follow conversations. This is a question that is often misunderstood. ChatGPT, on the other hand, knew if someone asked you what the weather was like in San Francisco and you answered, “How's it in New York?” The user wanted a different prediction.
The realization that the new technology has leapfrogged Siri has set in motion the tech giant's most significant reorganization in more than a decade. Determined to catch up to his AI race in the technology industry, Apple has made generative AI a tent pole project. This is the company's special internal label that it uses to organize its employees around its once-in-a-decade initiatives.
Apple unveiled an improved version of Siri that is more conversational and versatile at its annual developer conference on June 10, according to three people familiar with the company's efforts who were not authorized to speak publicly. The company plans to make an announcement and demonstrate its AI efforts. The technology underlying Siri includes a new generative AI system that allows you to chat with questions rather than having him answer them one by one.
The Siri update is at the forefront of a broader effort to bring generative AI across Apple's business. The company is also adding more memory to his iPhone this year to support new Siri features. The company is also discussing licensing complementary AI models to power chatbots from several companies, including Google, Cohere, and OpenAI.
An Apple spokesperson declined to comment.
Apple executives are concerned that new AI technology could replace the iPhone's iOS software as the primary operating system, threatening the company's dominance in the global smartphone market. This was said by two people familiar with the thinking of Apple's leadership. Permission to speak in public. The new technology could create an ecosystem of AI apps called agents that can order Uber rides or make calendar reservations, undermining Apple's App Store, which generates about $24 billion in annual revenue.
Apple also worries that if it fails to develop its own AI system, the iPhone could become a “dumb brick” compared to other technologies. It's unclear how many people use Siri on a regular basis, but the iPhone currently accounts for 85% of global smartphone profits, generating more than $200 billion in sales.
That sense of crisis led to Apple's decision to cancel another big bet, a $10 billion project to develop self-driving cars, and redeploy hundreds of engineers to AI development.
Apple is also considering developing servers powered by iPhone and Mac processors, two of the people said. By doing so, Apple could potentially save costs and ensure consistency between the tools used for processes in the cloud and on devices.
Rather than competing directly with ChatGPT by releasing a chatbot with features such as writing poetry, Apple is aiming to set timers, create and add calendar events, according to three people familiar with ChatGPT's work. The company is focusing on making Siri better at handling tasks it already does. Add items to your grocery list. You may also be able to summarize text messages.
Apple plans to promote the improved Siri as more private than competing AI services because it processes requests on the iPhone rather than remotely in a data center. This strategy also saves you money. Because of the cost of cloud computing, OpenAI spends about 12 cents for each word his ChatGPT produces.
(The New York Times sued OpenAI and its partner Microsoft in December for copyright infringement of news content related to AI systems.)
But Apple faces risks because it relies on a small AI system inside the iPhone, rather than a large AI system stored in data centers. Research has shown that smaller AI systems are more likely to make errors known as hallucinations than larger AI systems.
“It's always been the vision for Siri to have a conversational interface that understands language and context, but that's a difficult problem,” said Siri co-founder Tom Gruber, who worked at Apple until 2018. he said. You can do better. As long as you don't come up with a uniform answer, you should be able to avoid trouble.''
Apple has several advantages in the AI race, including more than 2 billion devices in use around the world to which it can distribute its AI products. It also has a strong semiconductor team that has produced sophisticated chips that can power AI tasks such as facial recognition.
But over the past decade, Apple has struggled to develop a comprehensive AI strategy, and Siri hasn't seen much improvement since its introduction. The Assistant's struggles have dampened the appeal of the company's HomePod smart speaker. This was due to his inability to consistently perform simple tasks such as responding to song requests.
John Berkey, who worked on Siri for two years before founding the generative AI platform Brighten.ai, said the Siri team hasn't gotten the attention or resources that other groups within Apple have. Ta. Each division of the company, such as software and hardware, operates independently of each other and only shares limited information. But to be successful, you need to incorporate AI into your products.
“It's not in Apple's DNA,” Berkey said. “That's a blind spot.”
Apple has also struggled to recruit and retain top AI researchers. Over the years, the company has acquired AI companies led by leaders in the field, each of which exited after a few years.
The reasons for their departure vary, but one factor is Apple's secrecy. The company publishes fewer papers on AI-related research than Google, Meta, and Microsoft, and doesn't attend conferences like its rivals.
“Researchers are saying, 'What other options do we have?' Can I go back to academia? Can I go to a research institute, somewhere where I can work a little more outdoors?” 2020 said Ruslan Sarakudinov, a leading AI researcher who left Apple and returned to Carnegie Mellon University in 2017.
In recent months, Apple has been publishing an increasing number of AI papers. But prominent AI researchers have questioned the value of the papers, saying they create the impression of meaningful research rather than providing an example of what Apple could bring to market. It states that there is.
Tsu-Jui Fu, an intern at Apple and an AI doctoral student at the University of California, Santa Barbara, wrote one of Apple's recent AI papers. He spent last summer developing a system for editing photos using written commands rather than Photoshop tools. He said that Apple supported the project by providing his GPU needed to train the system, but that he had no interaction with his AI team working on Apple products.
He said he has interviewed for full-time jobs at Adobe and Nvidia, but plans to return to Apple after graduation. Because he thinks he can make a bigger difference there.
“AI products and research are emerging at Apple, but most companies are very mature,” Fu said in an interview with the Times. “At Apple, you have more room to lead projects rather than just being part of a team doing something.”
