Apple is attempting to strike a difficult balance between innovation and privacy as it prepares to launch several artificial intelligence (AI) programs.
As Apple ramps up its AI offerings through updates to its own features and a partnership to integrate OpenAI’s ChatGPT into Apple devices, the company is touting new systems to keep users’ data private and secure.
The tech giant that built its brand around prioritizing user privacy and safety is aiming to keep up its reputation as it attempts to make waves in the ongoing AI arms race through guidelines that differentiate from typical ChatGPT norms and a new cloud system to process some AI requests.
But plans to put ChatGPT straight onto devices with an upcoming operating system update could risk changing that dynamic for Apple.
“There [are] regulators that are looking at every single tech company with sort of a microscope in terms of antitrust, in terms of privacy, and what they’re doing associated with consumer data that they’re utilizing associated with monetizing it,” said Ari Lightman, a digital media professor at Carnegie Mellon University.
Apple in the past has stayed away from that and put its focus on selling products and services — and as long as they “adhere to their core tenants,” Lightman said he doesn’t think there will be a defection of iPhone users.
But a breach or privacy issues could cost Apple its high ground on consumer safety in the iOS system and affect the company, Lightman added.
The company announced two key AI updates coming to Apple devices — Apple Intelligence and an integration with ChatGPT — at its Worldwide Developer Conference last week.
The two separate updates seem to be aimed at complementing each other by giving users options for different AI use cases. Whereas Apple Intelligence is focused on a user’s personal information stored on their iPhone, the integration with ChatGPT would offer users a way to access the outside generative AI system’s information to assist with certain tasks.
For use of ChatGPT, users will be able to access the popular OpenAI chatbot through the Apple device for free without having to create a separate account. Users will be able to access ChatGPT through Apple’s voice assistant Siri, which will also be updated, or through writing tools.
Along with the AI updates, Apple announced privacy updates aimed at keeping users safe, including stricter standards for ChatGPT integration than OpenAI’s typical privacy policies.
Apple also announced Private Cloud Compute, a cloud intelligence system that is designed for private AI processing. Apple said the system brings its “industry-leading security and privacy” from Apple devices into the cloud, and the personal user data that is sent to the system is not accessible to anyone, including Apple, other than the user.
Given Apple’s focus on selling products rather than user data like some of its tech industry rivals, it is “unsurprising that they explicitly highlighted privacy, security, and transparency” in the presentation, said Valerie Wirtschafter, a fellow at Brookings Institution’s Artificial Intelligence and Emerging Technology Initiative.
“I’d suspect it’d be [a] pretty horrible business decision to renege on this type of commitment, but it will certainly be interesting to watch how it plays out in practice,” Wirtschafter said in an email.
The integration has already raised some questions about data privacy, including from Tesla and X owner Elon Musk, a fierce critic of OpenAI and CEO Sam Altman.
After Apple’s announcement last week, Musk threatened to ban Apple devices from his companies, calling its integration with OpenAI “an unacceptable security violation.”
Musk’s criticism came after he previously sought legal action against OpenAI. Musk filed and later dropped a lawsuit levied against OpenAI and Altman in March, alleging the company failed to adhere to its founding mission to develop AI for the benefit of humanity, a day after his threat to ban Apple devices over the partnership.
In announcing the integration with ChatGPT, though, Apple announced different rules that would be in place compared to OpenAI’s typical rules.
Notably, users’ requests sent to ChatGPT through integration will not be stored by OpenAI. And on top of a contractual obligation that OpenAI will not try to connect a user with their request, Apple will make it difficult to do so on a technical level by obfuscating a users’ IP addresses, an Apple spokesperson said.
If a user chooses to connect their ChatGPT account, though, they will be subject to OpenAI’s rules.
“When you look at Apple integrating OpenAI, I think actually it’s good thing for OpenAI … not a bad thing for Apple, because OpenAI is now going to, through partnership, [get the] benefit of all of Apple’s experience,” said Jason Hogg, an executive in residence at Great Hill Partners and a former FBI special agent.
Apple’s integration with OpenAI comes as the ChatGPT maker has enjoyed more than a year of public popularity with the chatbot on the market. That popularity has also brought scrutiny in the form of regulatory investigations and lawsuits based on copyright concerns.
Apple’s move also follows in the footsteps of AI announcements from rivals, including Microsoft, which has integrated OpenAI in its own services, and Google’s launch of rival chatbot Gemini.
Until last week’s announcement, Apple had been more hushed on AI updates and advancements as other tech giants pushed ahead.
“To what extent does Apple want to be in bed with a company that’s way deep sued by a variety of different organizations and associations?” Lightman said.
“It basically shows that they are willing to take that hit to try to level up in terms of AI capabilities, that’s how lucrative the space is,” he added.