My quick take on Apple Intelligence

Are We Emotionally Ready for Apple Intelligence?

Honestly, my head is still spinning after Apple’s WWDC24 but I would be remiss if I did not share my two cents before everyone else’s hot takes begin to cool down. There were so many major updates to all my favorite Apple products. However, the most impactful for me was Apple’s approach to AI.

Bottom line upfront, I am a big Apple fanboy. I have almost every current Apple product in my home and would have all the remaining products if my wife would allow it. I joined the Apple ecosystem with the original iPod (2001) and then my first iPhone, the iPhone 3 (2008), and haven’t looked back since. That being said, I have been very disappointed with Siri’s performance once I jumped on the ChatGPT bandwagon.

Now, seeing Apple embrace and rebrand AI, and partner with one of AI’s major players, has made me very excited. However, as a practitioner of cybersecurity, I have some concerns. Mentally walk with me as I work through these concerns and honestly determine if Apple Intelligence is the future of my dreams and if you and I should feel comfortable with our Apple devices “communicating more intelligently” on and off our Apple devices.

First, a quick rundown from Apple’s Machine Learning Research page. Don’t worry, I used ChatGPT to summarize the important parts…

Responsible AI Development

Apple emphasizes responsible AI development with four main pillars: user empowerment, representation, careful design, and privacy protection. This approach ensures AI tools are designed to respect user autonomy, promote fairness and inclusivity, and prioritize safety and security. The models aim to provide useful and safe user experiences while maintaining the highest standards of privacy.

Pre and Post Training

Apple’s approach involves sophisticated pre-training and fine-tuning processes. During pre-training, models are exposed to vast amounts of data to learn general patterns and representations. Post-training includes fine-tuning with specific, carefully curated datasets to specialize the models for particular tasks, ensuring they meet high standards of safety, performance, and relevance.

But what about the sending and receiving of private data?

I thought you would never ask. Here is another summary from Apple’s Newsroom page, titled “Apple extends its privacy leadership with new updates across its platforms,” summarized to remove a lot of the sales jargon and hone in on On-Device Privacy vs. Cloud Privacy and ChatGPT Usage.

Privacy on Device:

Apple ensures that interactions with ChatGPT prioritize user privacy through on-device processing. This approach means that sensitive data is processed directly on the user’s device rather than being sent to external servers. Features like on-device Siri processing and advanced encryption techniques protect personal information, ensuring that data such as messages, health information, and other sensitive content remain secure and private.

Privacy in the Cloud:

For tasks that require cloud processing, Apple uses advanced encryption and anonymization techniques to protect user data. The introduction of Private Cloud Compute allows complex AI tasks to be performed in the cloud without compromising user privacy.

User Control and Transparency: Apple’s communication with ChatGPT is designed to provide users with control and transparency over their data. This includes:

• Data Usage Transparency
• User Consent
• Regular Updates

By integrating these measures, Apple ensures that interactions with AI, including ChatGPT, are secure, private, and transparent, aligning with their commitment to user privacy.

Sounds great! But do I believe our Apple overlords and why?

Yes, only because Apple has consistently emphasized its commitment to user privacy, integrating robust privacy features into its products and services. Here are some examples of Apple’s dedication to privacy:

1. Data Encryption: Apple uses end-to-end encryption to ensure that data transmitted between devices and stored on iCloud is secure. This means that only the user and the people they choose can access their data, and not even Apple can decrypt this information.
2. App Tracking Transparency (ATT): Introduced in iOS 14.5, ATT requires apps to ask users for permission before tracking their activities across other companies’ apps and websites. This feature gives users more control over their data and how it’s used for advertising purposes.
3. Privacy Labels: Apple’s App Store now includes privacy labels, which provide users with a clear overview of the data an app collects and how it is used. This transparency helps users make informed decisions about the apps they download.
4. Secure Enclave: The Secure Enclave is a dedicated chip used in Apple devices to handle sensitive information, such as biometric data for Face ID and Touch ID. This chip ensures that this data is securely stored and never leaves the device.
5. Mail Privacy Protection: In iOS 15, Apple introduced Mail Privacy Protection, which prevents email senders from knowing when an email has been opened and masks IP addresses to prevent tracking.
6. Safari Privacy Features: Safari includes features like Intelligent Tracking Prevention, which uses machine learning to identify and block trackers, and a Privacy Report that shows which trackers have been blocked over the past 30 days.
7. Differential Privacy: Apple uses differential privacy techniques to gather useful data while ensuring that individual user information remains private. This approach adds random noise to data collected from users, making it difficult to identify specific individuals while still providing useful insights for improving services.
8. Sign in with Apple: “Sign in with Apple” allows users to create accounts with apps and websites using their Apple ID, reducing the need to create new accounts and share personal information. This feature includes an option to hide the user’s email address, further protecting their privacy.
9. Location Services Control: Apple provides granular controls for location services, allowing users to choose whether apps can access their location always, only while using the app, or never. This control ensures users have full visibility and control over how their location data is used.
10. Health Data Protection: Apple’s Health app and HealthKit framework store health and fitness data securely and ensure that this data is only shared with the user’s explicit consent. Apple does not have access to this data unless the user chooses to share it.
11. Messages and FaceTime Encryption: Messages and FaceTime use end-to-end encryption to ensure that conversations remain private and secure. This means that only the sender and receiver can read the messages, and not even Apple can access them.
12. Third-Party Code Audits: Apple conducts regular audits of third-party apps available on the App Store to ensure they comply with its strict privacy policies. Apps found to be in violation are removed from the store to protect users.

These initiatives highlight Apple’s proactive approach to safeguarding user privacy and setting industry standards for data protection.

But are these Apple Intelligence features as secure as advertised?

Prelaunch, no one outside Cupertino knows for sure, but I have faith. And faith is all I can hold on to until Apple’s MacOS Sequoia, watchOS 11, and visionOS 2 later this year or iOS 18 officially rolls out with the inevitable iPhone 16 this fall. I know I’ll be there on day one to replace my very old iPhone, so if things go wrong, you can blame me for ushering in real-life Terminator Judgment Day.

Ryan is a retired Air Force veteran who brings over 20 years of experience in network infrastructure, project management, and cybersecurity consulting to his current role at BuddoBot . At Buddobot, he is dedicated to supporting national security by helping organizations transition from costly, reactive, and automated IT and security practices to proactive and robust security solutions.