The complicated partnership between Apple and OpenAI
By Samantha Murphy Kelly, CNN
(CNN) — When OpenAI CEO Sam Altman attended Apple’s annual developer conference this week, he walked the campus, mingling with current and former executives, including Apple co-founder Steve Wozniak. Nearly an hour later, the iPhone maker announced a much-rumored partnership with OpenAI to bring its ChatGPT technology to devices later this year.
But Altman, who has emerged as the poster child for generative AI in the 18 months since the launch of ChatGPT, was not featured in Apple’s formal presentation, neither in person or via livestream. Nor did he join Apple CEO Tim Cook and other executives in a private press event about privacy and security and the partnership between the two companies.
“I was not surprised Sam Altman did not appear on stage,” said Ben Wood, an analyst at market research firm CCS Insight in an interview with CNN. “Apple had to manage the message carefully. OpenAI is merely the vehicle to address broader AI-powered inquiries that are not core to the Apple experience. Having him in the livestream would have only created an unnecessary level of confusion.”
Earlier this week, Apple showed off a handful of AI-powered features coming to the iPhone, iPad and Mac in the fall – the majority of which are fueled by the company’s own proprietary technology, called Apple Intelligence.
The company will offer OpenAI’s viral ChatGPT tool in a limited capacity, usually only when Siri is activated and needs more assistance answering an inquiry.
The move to invite Altman to the announcement but not have him appear before the public also represents in some ways how Apple is cautiously moving forward with the partnership. OpenAI, along with other AI companies, continues to face concerns from researchers, industry experts and government officials around misinformation, biases, copyright, privacy and security, and more. The deal also comes at a time when the industry is moving quickly, and government regulators, companies and consumers are still figuring out how to engage with the technology responsibly.
Apple hopes that a big push into AI could spur growth for iPhone sales as the device has languished without a major upgrade for years and users are now waiting longer to upgrade their devices as a result. An uncertain economic environment is also weighing on consumers, particularly in China.
The company also faces regulatory scrutiny in Washington and was recently passed by chip maker Nvidia as the second-largest public company in the US. In the 60 hours after Apple’s Monday event, however, the company’s stock price (AAPL) rose by as much as 10%, boosting Apple’s market capitalization by more than $300 billion, rocketing past Nvidia and putting Apple back in contention with Microsoft for largest market value.
The timing is further noteworthy: Apple is not always first to adopt and integrate emerging technologies — it typically researches, develops and aims to perfect new tech for years before including it in new products — but the speed at which the world is adopting generative AI is perhaps expediting the company’s need to have a smartphone with the latest buzzy technology.
“Apple needed to deliver an AI story and Apple Intelligence should help placate nervous investors and reassure them that Apple is keeping pace with its rivals,” Wood added. “The partnership with ChatGPT is a major development which beefs up Apple’s AI offerings, and new features like a significantly improved Siri will be welcome to users.”
But the partnership could also open Apple up to some vulnerability, as it has no control over OpenAI’s models or what it does with user inputs. Apple alignment with a company and a technology that have yet to win public trust could also cause some challenges down the road.
A limited partnership
Although Apple has been working on its own AI program for years, a partnership with OpenAI marks a way for it to fill in the competitive gaps.
When a user has a question beyond Siri’s scope, ChatGPT can step in. In the demo following the keynote, Apple showed CNN how someone could upload a picture of vegetables at a farmer’s market and ask what they could make for dinner. Siri may suggest that the question is better suited for ChatGPT, and provide the user with a prompt to give consent to use the service.
Apple using ChatGPT as a complimentary service could potentially cut down on risks associated with the company. It’s also very possible Apple could partner with other AI companies in the future too, such as Google’s Gemini or more niche providers with certain expertise, such as in healthcare.
“I think Apple will take a pragmatic approach to the OpenAI partnership,” Wood said. “If Apple finds that the relationship with OpenAI starts to affect the overall user experience or worse, create challenges around security and data integrity, it might look to add further guardrails, or find other ways to deliver AI-powered content.”
A focus on privacy and security
Apple has been vocal about how it intends to keep user data private and secure when using its own proprietary technology, noting most AI functions will be done on the phone and inputs will be kept away from a far-off server cloud.
“As we look to build in these incredible new capabilities, we want to ensure that the outcome reflects the principles at the core of our products, it has to be powerful enough to help with the things that matter most to you,” Cook said during the keynote. “It has to be intuitive and easy to use. It has to be deeply integrated into your product experiences.”
“And, of course, it has to be built with privacy, from the ground up,” he added.
Apple said it will not share any personal user information with OpenAI, so inquiries made through ChatGPT won’t be linked to an Apple user’s account. And the repeated consent prompt to require users to opt into using ChatGPT with Siri is noteworthy too; each time Siri wants to kick over a question to ChatGPT, it’ll ask for permission first.
Wood believes the consent prompts and other guardrails that Apple is implementing “reflect its nervousness.”
Meanwhile, Reece Hayden, a principal analyst at ABI Research, told CNN that Apple’s approach is smart because it provides customers with a choice of how to deal with their data.
“By providing a phased approach that blends ChatGPT and native capabilities, users will worry less about the partnership,” he said. “Apple can also continue to highlight their own AI capabilities and mitigate some of the risks of being associated with OpenAI, which remains in a state of flux.”
Industry concerns
Companies like OpenAI have acknowledged the serious risks posed by AI — from manipulation to a loss of control that could potentially result in human extinction — but many experts, researchers and AI employees believe they should be doing more to educate the public about risks and protective measures. Just last week, a group of OpenAI insiders demanded that artificial intelligence companies be far more transparent about the concerns over the technology they’re building.
So it came as no surprise when some industry watchers, including Elon Musk, were quick to react to Apple’s partnership with OpenAI.
In a post to X Monday, Musk said he would ban Apple devices at his companies — which include Tesla, SpaceX and X, among others — if the iPhone maker went ahead with its AI plans. If Apple “integrates OpenAI at the (operating system) level,” Musk said that would constitute “an unacceptable security violation.”
Although concerns over how employees are using AI models is a topic of worldwide discussion at the moment across many industries, Gartner analyst Annette Zimmermann said Musk’s reaction is a bit misdirected and the topic is not specific to iPhones.
“Any employee with a smartphone should follow company policies and not enter any private information into the open domain of ChatGPT,” she said. “This is not specific to the iPhone … or Tesla.”
Andrew Cornwall, senior analyst at Forrester, told CNN he thinks it’s unlikely that Apple users will become loyal to ChatGPT, as many people won’t use the service unless the iPhone maker can’t provide an appropriate response.
“When users do query ChatGPT, Apple will track the prompts and gather metrics to improve its own models,” he said. “Apple may switch providers or even use more than one third party — until it has mastered its own model. At that point, Apple will shut the garden gate.”
The-CNN-Wire
™ & © 2024 Cable News Network, Inc., a Warner Bros. Discovery Company. All rights reserved.
