When Apple COO Tim Cook and executives take to the stage at the Worldwide Developers Conference in Cupertino on June 10, there's one announcement that could herald a new phase in the AI wars: the tech giant is launching generative AI in iPhone-only apps like Siri and iMessage. But the real story at the developer conference will be how Apple plans to roll out AI capabilities to its vast 2.2 billion iOS devices. A focus on building trust with consumers around AI use could be a differentiator that Apple can leverage to avoid the pitfalls that Google and others fell into by haphazardly introducing AI to messaging.
“It's understandable that these big companies are desperately competing for some form of leadership in the development and deployment of AI,” Benjamin Kuipers, professor of computer science and engineering at the University of Michigan, told indianexpress.com. “That said, trust is a vital asset for individuals, companies and society at large. When trust is lost, plans become more difficult and executing those plans becomes riskier and more costly.”
In recent months, Google, OpenAI and Microsoft have been in the spotlight for all the wrong reasons, revealing a different side of AI that contrasts with the hype. First, Google had to roll back its AI Overviews search tool just two weeks after it was announced because the new feature produced inaccurate and bizarre results. Second, OpenAI came under fire for introducing a voice that sounded very similar to Hollywood actress Scarlett Johansson in ChatGPT, which ChatGPT's developers denied. But this led to a larger discussion about new threats in the AI era, such as “who owns your voice.” Microsoft's AI ambitions were also called into question when its new AI-powered Recall feature, which screenshots everything you do on your PC as part of its new Copilot Plus PC product line, was described as a “catastrophe” for cybersecurity. Microsoft walked back on its controversial “Recall” feature after a privacy uproar.
“The recent AI revolution based on deep neural networks, large-scale language models and generative AI has enabled new levels of performance at the cost of incomprehensible mechanisms and occasional serious illusory errors,” Kuipers said, citing the recent example of Google's AI summary search feature, where these errors led to a major loss of trust in the company.
What's often overlooked in the broader AI environment is a sense of urgency and a fear of not being able to keep up. This makes it hard not to get involved in the AI movement. This gives big tech companies an advantage because every AI feature or tool they release attracts users. But in reality, the information about how new technologies work and how they ultimately benefit users and society at large is often unclear.
While many question whether Apple is missing out and ignoring GenAI's potential, the reality is that Cupertino has a history of being very cautious about jumping into new technologies and has a clear stance on privacy. Experts say Apple's AI strategy is likely to be more focused on keeping users' information more private than other companies.
“Apple isn't moving at the same pace on AI as Google, and its entire business isn't at risk,” said Anisha Bhatia, senior technology analyst at GlobalData, “so Apple can and will approach generative AI cautiously.”
Bhatia agrees that with fierce competition from OpenAI and Microsoft, Google can't afford to wait before incorporating AI into its core products and services. But Bhatia says Google has a golden egg in Android, its leading mobile operating system used by billions of users. If Google rolls out its GenAI capabilities sooner, it can only benefit the search engine giant, as it can reach consumers directly through the products and services they're already familiar with.
Apple and Google have different business models. However, both are platform holders with huge install bases through their respective mobile operating systems and dominance through applications used by billions of people every day. Google has already demonstrated what AI can achieve through bespoke apps and products with Android devices. Apple is on a similar path by convincing the world that AI will become a bigger part of users' lives.
But unlike Google, which has partnered with a variety of smartphone makers, each with different timing and intent for introducing GenAI features to Android phones, Apple is likely to focus on AI tools that consumers can use and leverage in their daily lives, rather than flashy features that may not have as much practical impact. Apple also has the advantage of being able to leverage a huge user base that uses the iPhone, one of the most popular smartphones on the planet, which could go a long way in giving Cupertino an edge over its competitors, even as the company appears to be playing catch-up in the AI space.
Perhaps what will really spark the debate is how Apple solves AI's biggest problem for its own benefit by putting privacy at the center of everything it does. Apple doesn't sell ads like Google or Meta, nor does it have questionable security claims like Microsoft. Cupertino has used privacy as a differentiator in the past; in fact, many users buy Apple products because of the company's commitment to privacy. Experts say Apple will bring privacy back to the forefront, this time using artificial intelligence to highlight how the security of user data is under threat.
Apple will demonstrate how GenAI capabilities can now run natively on devices such as iPhones and Macs, instead of in the cloud. This maintains control over user data and security while also addressing response latency and cost factors associated with running AI models in the cloud, as well as the energy required to operate data centers. To do so, Apple will use its own Large Language Models (LLMs) and will not rely on user uploads to train the AI, as many platform owners do. Apple will highlight how its hardware and software are seamlessly integrated, and how it uses a “Neural Engine” as part of the chips in iPhones and Macs to run these AI models, keeping sensitive data only on the device.
But insiders have warned that to deploy AI more broadly on personal devices, Apple can't rely solely on on-device feature support. It will eventually need to add hybrid AI capabilities. Partnering with external resources such as OpenAI or Google could be an option to bring cloud AI capabilities to the iPhone. But how to incorporate external companies' technologies that use cloud processing while protecting user privacy is an issue that needs to be addressed.
“AI advances will be gradual and planned, especially in consumer tech, but rest assured there's no escaping AI. The biggest news will be what AI capabilities Apple will include in its devices and how quickly it will roll them out,” Bhatia said.
© IE Online Media Services, Inc.
First uploaded: August 6, 2024, 8:25 AM
