SAN FRANCISCO — Nearly two years after OpenAI ignited a race to add generative artificial intelligence into products, Apple jumped into the competition Monday, as it revealed plans to bring the technology to more than 1 billion iPhone users around the world.
During a two-hour presentation from its futuristic Silicon Valley campus, Apple said it would be using generative AI to power what it is calling Apple Intelligence. The system will prioritize messages and notifications and will offer writing tools that are capable of proofreading and suggesting what users have written in emails, notes, or text. It also will result in a major upgrade for Siri, Apple’s virtual assistant.
While introducing its new AI, Apple emphasized how it planned to integrate the technology into its products with privacy in mind. The new AI features could help calm concerns that Apple had slipped behind its biggest rivals in the tech industry’s embrace of AI.
The company said that the technology, which can answer questions, create images, and write software code, would perform sensitive tasks. It showed how the system would be able to automatically determine if a rescheduled meeting time would complicate plans to attend a child’s theater performance.
It said the computer processing would be done on an iPhone rather than in data centers, where personal information has a greater risk of being compromised. For complex requests that require more computing power, it has created a cloud network with Apple semiconductors that, it said, is more private because it’s not stored or accessible, even by Apple.
Apple struck a deal with OpenAI, the maker of ChatGPT, to support some of its AI capabilities. Requests that its system can’t field will be directed to ChatGPT. For example, a user could say that they have salmon, lemon, and tomatoes and want help planning dinner with those ingredients. Users would have to choose to direct those requests to ChatGPT, ensuring that they know that the chatbot — not Apple — is responsible if the answers are unsatisfying. OpenAI CEO Sam Altman attended the Apple event.
“As we look to build in these incredible new capabilities, we want to be sure that the outcome reflects the principles at the core of our products,” said Apple CEO Tim Cook. “It has to be powerful enough to help the things that matter the most to you. It has to be intuitive and easy to do.”
Apple also said that it would be making improvements to its iPhone software system. This fall, messaging will add the ability to schedule messages and respond to messages by tapping back with more emojis. Apple will also release a redesigned Photos app to make it easier to browse images by topics like pets and travel. And it said that iPhone users would be able to send high-resolution images to Android smartphones.
Apple brings several strengths to the AI race. Its semiconductor development team is among the most talented in the industry and has been making chips that power complex AI functions for years. The company has also marketed itself as being better at protecting people’s personal information than its rivals because it makes money by selling devices, not from advertising.
But Apple has several weaknesses that could slow its AI development. The secretive company has struggled to recruit and retain leading AI researchers because it limits how much research it publishes. It has also sought to license published material and objected to collecting it without permission, as other generative AI companies have done to build and train their technology.
Though Siri has been around for more than a decade, Apple has let the voice assistant languish. The assistant has frustrated users with its failure to recognize various requests, and its ability to actually converse is limited because it is programmed to follow each individual command.
Generative AI could improve Siri because it has been trained on spoken conversations pulled from podcasts and videos. The result is a system that can mimic the way people talk.
Apple said Siri will be able to remember the context of something that a user has raised with it. For example, if someone asks for the weather in Muir Woods National Monument and later asks to schedule a hike there, Siri will know the hike it is scheduling is in Muir Woods.
Siri will also understand more of the things that people want to get done in iPhone apps. For instance, users can ask it to show a photo of a friend and it will find and surface those images in the Photos app. It also will be able to perform tasks for people such as finding an image of a user’s driver’s license and filling it out into a form.
Other generative AI capabilities that Apple demonstrated included automatically summarizing audio recordings, allowing customers to create movies from photos by writing a description, and cleaning up photos by removing distracting images in the background.
Apple’s push into generative AI was in part a response to Wall Street investors. The technology has lifted the values of Microsoft, a big player in generative AI, and Nvidia, which sells AI chips. This year, Microsoft dethroned Apple as the most valuable technology company in the world.
“This is the biggest event for investors since the iPhone because they have to infuse AI into their products to thrive,” said Gene Munster, a managing partner at Deepwater Asset Management, which invests in emerging technology companies. “Today was about showing that they will make AI a core competency and that they can deliver an AI experience that consumers want.”
This article originally appeared in The New York Times.