Apple's top software executives decided early last year that the company's virtual assistant Siri needed a brain transplant.
The decision came after executives Craig Federighi and John Giannandrea spent weeks testing OpenAI's new chatbot, ChatGPT. The product can write poetry, generate computer code and answer complex questions, two people familiar with the company's work said. was not allowed.
Introduced as the original virtual assistant in every iPhone in 2011, Siri has been limited to individual requests for years and has never been able to follow conversations. It often misunderstands questions. ChatGPT, on the other hand, knew that if someone asked for the weather in San Francisco and then said, “What about New York?” That user wanted another forecast.
The realization that new technology has led Siri to set in motion the tech giant's most significant restructuring in more than a decade. Determined to join the tech industry's AI race, Apple has made Generative AI a tentpole project — the company's special, internal label it used to organize employees around once-in-a-decade initiatives. Is.
Apple is expected to show off its AI work at its annual developers conference on June 10 when it releases an improved Siri that is more interactive and versatile, according to three people familiar with the company's work, who have Public speaking was not allowed. Siri's core technology will include a new generative AI system that will allow it to chat instead of answering questions one at a time.
Siri's update is at the forefront of Apple's broader efforts to embrace creative AI in business. The company is also expanding the memory in this year's iPhones to support its new Siri capabilities. And he has discussed licensing complementary AI models that power chatbots from a number of companies, including Google, Coheir and OpenAI.
An Apple spokesman declined to comment.
Apple executives fear that new AI technology threatens the company's dominance of the global smartphone market because of its potential to become the primary operating system, displacing the iPhone's iOS software. He didn't have it, said two people familiar with the leadership's thinking. Allowed to speak publicly The new technology could also create an ecosystem of AI apps, called agents, that could order Ubers or make calendar appointments, which would have undermined Apple's App Store. , generating approximately $24 billion in annual sales.
Apple also fears that the iPhone could become a “dumb brick” compared to other technologies if it fails to develop its own AI system. While it's unclear how many people regularly use Siri, the iPhone currently accounts for 85 percent of global smartphone profits and generates more than $200 billion in sales.
That sense of urgency contributed to Apple's decision to cancel its second big bet — a $10 billion project to develop a self-driving car — and reassign hundreds of engineers to work on AI.
Apple has also explored making servers powered by its iPhone and Mac processors, two of the people said. Doing so could help Apple save money and create consistency between the tools used to process the cloud and its devices.
Instead of directly competing with ChatGPT by releasing a chatbot that does things like write poetry, said three people familiar with its work, Apple has focused on making Siri better at those tasks. It already does, including setting timers, creating and adding calendar appointments. Items on a grocery list. It will also be able to summarize text messages.
Apple plans to bill Enhanced Siri as more private than rival AI services because it will process requests on iPhones rather than in data centers. The strategy will also save money. OpenAI charges about 12 cents per 1,000 words that ChatGPT generates due to cloud computing costs.
(The New York Times sued OpenAI and its partner Microsoft in December for copyright infringement of news content about AI systems.)
But Apple faces risks by relying on smaller AI systems housed on iPhones rather than larger systems stored in data centers. Research has found that smaller AI systems may be more prone to errors, known as illusions, than larger systems.
“It's always been the vision of 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. ” “Now that the technology has changed, it should be possible to do much better. As long as it's not a one-size-fits-all effort to respond to anything, they should be able to avoid the problem. “
Apple has several advantages in the AI race, including more than two billion devices in use worldwide where it can distribute AI products. It also has a leading semiconductor team building cutting-edge chips capable of powering AI tasks like facial recognition.
But for the past decade, Apple has struggled to develop a comprehensive AI strategy, and Siri hasn't improved much since its introduction. Assistant's struggles killed the appeal of the company's HomePod smart speaker because it couldn't consistently perform simple tasks like fulfilling a song request.
The Siri team has failed to garner the kind of attention and resources that have gone to other groups within Apple, said John Berkey, who previously founded Brighten.ai, a creative AI platform. Worked on Siri for two years. A company's divisions, such as software and hardware, operate independently of each other and share limited information. But AI needs to be threaded through products to be successful.
“It's not in Apple's DNA,” Mr Berkey said. “It's a blind spot.”
Apple has also struggled to recruit and retain top AI researchers. Over the years, it acquired AI companies that were leaders in the field, but they all left after a few years.
The reasons for their departures vary, but one factor is Apple's privacy. The company publishes fewer papers on its AI work than Google, Meta, and Microsoft, and it doesn't participate in conferences the way its rivals do.
“Research scientists say: 'What are my other options? Can I go back to academia? Can I go to a research institute, a place where I can do a little more work in the open?'” a said leading AI researcher Ruslan Salakhotdinov, who left Apple in 2020 to return to Carnegie Mellon University.
In recent months, Apple has increased the number of AI papers it publishes. But prominent AI researchers have questioned the value of the papers, saying they are more about creating the impression of meaningful work than providing examples of what Apple could bring to market.
Tsu-Jui Fu, an Apple intern and 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 with written commands instead of Photoshop tools. He said Apple supported the project by providing him with the GPUs necessary to train the system, but he had no contact with the AI team working on Apple products.
Although he said he interviewed for full-time jobs at Adobe and Nvidia, he plans to return to Apple after graduating because he thinks he can make a big difference there.
“The AI product and research at Apple is emerging, but most companies are very sane,” Mr. Fu said in an interview with The Times. “At Apple, I may have more room to lead a project than to be a member of a team doing something.”