Apple Partners With ChatGPT: Everything Should Be Fine, Right?
On June 10th, 2024, the most important technology partnership in history was announced.
On this day, an alliance began between the most popular seller of consumer devices, with an estimated 150 million active units of iPhones in the US (making the company subject to lawsuits from the Justice Department for having a monopoly on the market) and OpenAI, the fastest-growing technology. This partnership will make generative AI tools more accessible and convenient while delivering a more personalized experience. A new era of powerful convenience that, once released, will be irreversible.
Apple Intelligence: Features Summary
Systemwide Writing Tools: Rewrite, proofread, and summarize text across apps with advanced AI tools.
Priority Messages and Smart Reply: Efficiently manage urgent emails and get smart response suggestions.
Priority Notifications: Highlight important notifications and summarize long ones for quick scanning.
Reduce Interruptions Focus: Minimize unnecessary notifications to stay focused on important tasks.
Audio Transcription and Summarization: Record and summarize audio notes and calls.
Image Playground: Create fun images with styles like Animation, Illustration, or Sketch.
Genmoji Creation: Generate custom emojis based on descriptions or photos.
Enhanced Photo Search and Clean-Up: Use natural language to find specific photos and remove unwanted objects.
Memory Movies: Create narrative-driven movies from your photos and videos with AI.
Advanced Siri Capabilities: Siri offers richer language understanding and on-screen awareness for more intuitive interactions.
Private Cloud Computer: Ensures privacy while leveraging cloud-based AI processing.
ChatGPT Integration: Access ChatGPT for enhanced content generation and document understanding within Apple’s ecosystem.
Consumers will now have access to Writing Tools to craft better messages. These tools will be able to proofread, check text for tone accuracy, and even rewritte to be more concise. User experience can be more personalized with the Image Playground and Genmoji. These tools generate custom images and emojis through the input of a prompt or from access to one’s image gallery.
One of the greatest updates will come to Siri, Apple’s digital assistant. Siri will be able to recognize voice more naturally and retain a memory of a voice chat, even text. This once-bemoaned assistant can now access information from calendar events, conversations and even recognize the names of individuals in photos. Siri will even have access to OpenAI’s ChatGPT.
AI Arms Race Summary
In December 2023, Google fired the first shot at OpenAI’s transformative generative AI platform ChatGPT with Google Gemini. Having held the reins, OpenAI has remained revered since 2018. On December 6th, 2023, Gemini placed that status in question, launching the arms race of generative artificial intelligence.
OpenAI History
OpenAI founded the company in 2015 and achieved one of its most significant milestones with the release of GPT-1 in 2018. This event established subsequent iterations, including GPT-2, released in 2019, and GPT-3 in 2020. Both of these models showed noted improvements in language learning and understanding. After years of careful testing and experimenting, OpenAI would release its “chatbot” to the public with the 3.5 model. This model received over one million members joining the platform in just five days. It would gain 100 million in under two months.
Comparably, no other technology has witnessed that level of growth. Social media platforms TikTok and Instagram took nine months and 2.5 years, respectively, to reach the same number of users. This growth did not go unnoticed, with other competitors, such as Google DeepMind, in attempts to rival OpenAI.
Competition Summary
Now, the aristocracy faces intimidating competition, facilitating the development of other worthy and formidable AI companies. The list includes Anthropic’s Claude, X’s Grok, and Microsoft’s Copilot. Although these developments may not have had the nuance Microsoft Windows had during its Windows 95 premiere, the integration of GenAI in modern-day life has blasted open the bunker doors. While new names are gradually being released and accepted, one name, notably associated with technological advancement, and the one with the second largest market share of the technology sector is absent – Apple.
Apple/OpenAI Marriage Summary
Despite the start of 2024 marking an era of controversy for artificial intelligence, some developments have remained underappreciated by the mainstream public, including one of the most recent – Apple partnering with OpenAI and ChatGPT. Rumors circulated following the release of Gemini and continued. Roumors were suggesting that Apple might be lagging on the latest technology, perhaps now making it an anachronism. Now, in this climate of business integration with technology, is this a matter of technological detente and attempts to maintain pace with larger names, perhaps missing the proverbial starting gun? Or is this a more strategic alliance akin to that of Horace Smith and Daniel Wesson aimed at addressing current limitations and harnessing modern technology ahead of potential competitors?
On Implications…
The outfall is evident but has yet to reach the ground from this monumental partnership, a second question arises. What are the implications for consumers? With industry giants such as Apple and OpenAI uniting, the landscape of technology use is poised for an unprecedented shift. This collaboration will usher in a new era of seamless AI integrations into consumer devices, enhancing experience and accessibility. It will also raise new concerns about data privacy and the reach of AI’s influence on people’s daily lives. The implications for the public are vast and unpredictable.
Can The Apple See Through The Forest?
Apple hosted its Apple’s Annual Developer Conference in June 2024. Prominent industry leaders were on the Apple Campus, including current and former Apple executives and Sam Altman of OpenAI. Speculations around Apple and a partnership with an AI company had been circulating, notably with Google and Gemini in May. There were earlier reports of Apple competing or partnering with an AI company and even developed its primitive AI Ferret.
The Developer Conference
Apple’s Annual Developer Conference usually impacts various facets of the technology industry. Apple unveils major announcements, including product launches, and new features, setting the tone for the direction of the year. Developers benefit from educational sessions, workshops, and networking opportunities, fostering collaboration and the exchange of ideas. The conference plays a crucial role in expanding the Apple ecosystem. The conference ensures the integration and compatibility of new technologies while encouraging third-party development by introducing new APIs and tools. The keynote address often outlines Apple’s strategic vision, influencing market trends and expectations. Media coverage amplifies Apple’s message, generating consumer excitement and impacting the stock market. The economic impact extends to the developer community, spurring innovation and growth.
Calculated Choices
Apple is far from new to the technology industry, rapidly approaching its 50th anniversary in 2026. Analogous to the argument behind the true founder of calculus, while other companies have an earlier history, Apple was the one to have the metaphorical fruit fall on its head. Apple became one of the most prominent companies worldwide, with a near-omnipotent status in the technology industry.
Apple does not take the acceptance nor integration of new technology lightly. The company is notoriously slow to adapt to new advancements and is taking cautious, calculated approaches. Apple conducted its own research on optimizing Large Language Models for smaller devices and developed its latest in-house computer chips.
However, AI is not an industry that can truly be bridled at the moment. How it grows and advances, restrictions are counterintuitive to growth and progress, elements necessary to dispel fears and preconceived notions and propel the new models that will serve as tools for the future. Unlike technologies of the past, AI develops consistently. It is dependent upon user experience education, and requires a magnitude of data to foster growth. Sleeping on AI is not an option for the magnates of technology, Apple included, thus the development of Apple Intelligence.
Previous Collaborations
In December 2023, Apple boasted an early AI development in conjunction with Cornell University of Ferret. The designers created this system to compete with ChatGPT, Gemini, Claude, and Microsoft Copilot. It did not appear to gain the traction that other established technologies had set. The news from the Apple Campus on June 10, 2024, was Apple Intelligence.
Ultimately, while Apple has been in talks with Google and its AI Gemini, it is currently integrating with ChatGPT. However, the company is leaving new opportunities and partnerships open. This latest advancement was made only hours into the meeting on the Apple campus when the partnership was news. Interestingly, while CEO Sam Altman was present, he was absent during the formal presentation in person or virtually. Altman was also absent in a private press event where the topic encompassed privacy and security. However, he did announce his optimism for the partnership, stating, “Apple shares our commitment to safety and innovation, and this partnership aligns with OpenAI’s mission to make advanced AI accessible to everyone.”. It was a topic largely untouched during the Aspen Ideas Festival, only two weeks later.
Treasures Along the Altman-Apple Trail
Although this partnership almost seems like a fairytale marriage of established technology and an emerging parvenu, the relationship is unique. However, it became necessary after Apple’s market share was temporarily superseded by chip manufacturer Nivida in June. Additionally, the company witnessed iPhone sales drop over 10% during the first quarter of 2024. Apple protects consumer information and data, but this relationship is a complimentary partnership.
Privacy Concerns
Per Apple CEO Tim Cook, “It had to be built with privacy from the ground up,” mentioning integrations need to align with familiar product experiences, reflecting Apple’s core principles. Apple has had experience with personal data breaches, including the release of photos from iCloud in 2014 and the confirmation of contractors listening to Siri calls in 2019. Reflective in this nature, Apple will not share personal user information with OpenAI. Instead, Apple will ensure privacy in its Apple Intelligence by primarily processing AI queries locally on user devices, requiring powerful hardware and the latest OS versions, iOS18, iPadOS 18, and macOS Sequoia. For complex requests, those requiring greater processing power are handled by its Private Cloud Computer; data is isolated, encrypted, and not retained post-processing, with robust security measures allowing external verification and incentivizing vulnerability reporting through the Apple Security Bounty program. With this process, the data that a user would input in a prompt with OpenAI can not be linked to their personal accounts; however, that same data will be used by OpenAI to further the development of ChatGPT.
The Prenup Advantage
This minimizes the risks associated with Apple, but they are not the only benefactors of this relationship. Instead, OpenAI holds a strategic advantage, a proverbial prenup. They can utilize this data discreetly through Apple to train and adjust its models. OpenAI will also retain full control over its proprietary models, with no interaction from Apple and data use. Financially, while Apple stock did witness a slight surge, it was not the only one in this deal that witnessed a similar pattern. Although OpenAI is not publicly traded, at the moment, it is powered by Microsoft, the largest shareholder in the technology industry, with Microsoft stock growing 0.5% after the announcement of this joint venture.
In recent news, Apple Fellow Phil Schiller, the head of App Store and events, who responds directly to Tim Cook, has been awarded a seat for an Observer Role at OpenAI. This seat places Apple in the boardroom of OpenAI along with Microsoft, but without voting rights, it will permit it to witness the inner workings of the company and job security against any major issues.
Hungry, Hungry Bytes
Researchers collect data for future purposes, and consumers have valid concerns about how either company intends to use their personal information. This is especially true when integrating OpenAI into personal Apple devices, where a query may seem convenient, but the information shared is very much private. Understandably, consumers will be wary about how companies process their data. Apple was able to mitigate itself despite facing prosecution from the Justice Department and 15 individual states, including Washington.
Data Privacy
Through the continuous growth of AI, one thing has been at the forethought of consumers, and rightfully so: the protection of information. The training of Large Language Models (LLMs), such as OpenAI’s ChatGPT, involves the meticulous use of data to develop sophisticated AI systems capable of understanding and generating human language. Initially, researchers collect vast and diverse datasets from various sources, including books, websites, articles, and social media. This extensive data collection exposes the model to a wide range of language patterns, topics, and contexts, which are crucial for comprehensive learning.
Data Collection
Following data collection, the model undergoes the preprocessing stage, where data is cleaned to remove irrelevant, duplicate, or harmful information. Subsequently, the text is tokenized, breaking down the language into smaller units, such as words or subwords, which the model can process. This preprocessing is essential for standardizing the data, enabling effective learning during the subsequent training phases, and serving as a preventative measure to minimize sharing false data, hallucinations, or responses with a form of bias.
Fine Tuning Methodology
Deep learning techniques are central to the training process, specifically through neural networks with multiple layers, known as transformers. During training, the model’s parameters are adjusted based on the data it processes, refining its ability to predict subsequent words in a sequence or generate contextually relevant responses. These learning methods can be supervised or unsupervised. Supervised learning involves training the model on data pairs with inputs and desired outputs. In contrast, unsupervised learning allows the model to learn from raw data without explicit instructions, enhancing its ability to understand nuanced and subtle language patterns.
Models then undergo reinforcement learning for fine-tuning, which involves evaluating the model’s outputs and providing feedback for improvement. Human reviewers often assess the model’s responses for accuracy and appropriateness, guiding the model toward generating more reliable and contextually appropriate answers.
Continuous learning and updates are integral to the process. The developers retrain or fine-tune the model as new data becomes available to incorporate recent information and evolving language patterns. This ensures that the model remains up-to-date and relevant, capable of understanding and responding to contemporary events and modern language usage.
Protecting Your Personal Orchard
Using personal data in training LLMs, such as OpenAI’s ChatGPT, raises significant concerns about privacy and data security. Training these models requires vast amounts of data, often including personal information inadvertently collected from various sources such as social media, websites, other digital platforms, and in this particular case, a user’s device. This raises the risk of exposing or misusing sensitive information.
The potential for using personal data without explicit consent is one primary concern. Users may be unaware that AI systems utilize their interactions and information for training. This lack of transparency can erode trust and violate privacy rights. Additionally, aggregating personal data from numerous sources increases the risk of creating detailed user profiles that malicious actors could exploit.
Data security is another critical issue. The storage and processing of vast amounts of personal data make these AI systems attractive targets for cyberattacks. Any breach could expose sensitive information, leading to identity theft, financial loss, and other serious consequences.
Using personal data in training LLMs can lead to biased or inaccurate outputs, reflecting the prejudices or errors present in the data. This can perpetuate harmful stereotypes and misinformation, affecting individuals and society. Therefore, integrating personal data in AI training necessitates stringent measures to ensure privacy, security, and ethical use.
Pruning The Serpents Out Of The Trees
Apple, however, has thought through this process. While OpenAI will ultimately be using your data to train its models, Apple is not without its privacy policy as a consumer safeguard. The implementation of Apple Intelligence aims to enhance user experiences while maintaining privacy and data security.
Apple plans to process most AI queries locally on users’ devices, reducing the need to send data to cloud servers. This local processing requires powerful hardware, such as the latest iPhone 15 Pro, Pro Max, or devices running the M1 chip and above, all operating on iOS 18, iPadOS 18, or macOS Sequoia. If a consumer has a request that needs to be addressed through OpenAI, the system asks for permission to direct them to ChatGPT, where they submit the request separately from their device, limiting any association.
Securing the Skies
However, Apple will employ its Private Cloud Compute to mitigate the typical privacy concerns associated with cloud processing for more complex requests. Apple’s Private Cloud Compute uses specialized servers with Apple silicon and a hardened operating system, physically secured by destroying modification keys. The system isolates and deletes data processed in the cloud after completing the request, utilizing end-to-end encryption for secure transmission.
The system architecture includes “stateless computation,” ensuring that data from one query is not mixed with others and is not stored post-processing. There are no specialized access routes for debugging or other privileged purposes, and external security experts can verify Apple’s security claims with every server-side software release published for security research. The Apple Security Bounty program encourages researchers to report any vulnerabilities they have discovered.
Apple’s approach integrates robust hardware and software security measures to ensure user data remains private and secure, even when processed in the cloud. It is a promising solution amidst growing privacy concerns.
Who Wants to Take a Byte Out of Me…
From an economic perspective, Apple would see a boost in its stock price after the announcement; within two and a half days of the partnership, Apple stock would rise approximately 10%, bolstering the market value of the company to more than $300 billion, re-surpassing Nvidia, who had displaced Apple in June for the second greatest market share next to Microsoft, perhaps reflecting a market that is receptive and ready for the integration of AI with consumer goods. Additionally, Microsoft, the company behind OpenAI, would see its shares rise 0.5%.
While the figures appeal to businesses, the consumer, not necessarily the market, draws the line in the sand. The latest technology might be a talking point, but general acceptance is another. Other companies, such as Samsung, will integrate AI into their platforms, as seen with Samsung’s Galaxy AI addition to mobile devices in early 2024. Apple, as history would attest, is a first. This will be the first time AI will be available on personal devices, from the latest iPhone to personal computers. This could be a turning point for the company, as sales of iPhone and Apple in general have declined; with the gradual acceptance of AI into daily life, these features could improve sales, should consumers opt-in to the functionality. Although it might appear optional at the moment, it is more of a matter of when than if, as companies have been working on and developing AI tools like Google Circle.
A Ripening Core
The partnership between Apple and OpenAI marks a significant milestone in the evolution of artificial intelligence and its integration into everyday technology. As Apple ventures into this new territory with “Apple Intelligence,” it brings promise and concern. On the one hand, the collaboration is set to enhance user experiences by leveraging advanced AI capabilities, improving the functionality of beloved Apple products, and ensuring that these advancements are delivered with Apple’s renowned emphasis on privacy and security. On the other hand, using personal data to train these AI models raises valid concerns about privacy, data security, and the potential misuse of information. The true test will be how consumers react to these changes and how they adopt these new functionalities. Standing on the brink of a new era in AI technology, the union of Apple and OpenAI is the dawn of unprecedented innovation and a harbinger of unforeseen challenges.
