Apple Intelligence is the big buzz at WWDC. But when it comes to AI and the cloud, if you aren’t a huge enterprise or well-funded government, privacy and data security have always been a challenge when using any cloud service. With the introduction of Private Cloud Compute (PCC), Apple just did cloud services right — and put a real competitive moat in place.

Apple seems to have solved the problem of offering cloud services without undermining user privacy or adding additional layers of insecurity. It had to do so, as Apple needed to create a cloud infrastructure on which to run generative AI (genAI) models that need more processing power than its devices could supply while also protecting user privacy. 

While you can also use ChatGPT with Apple Intelligence, you do not need to (and OpenAI has promised not to store your data under the Apple deal, I think); PCC helps you run Apple’s own GenAI models instead.

The Apple achievement explained

“You should not have to hand over all the details of your life to be warehoused and analyzed in someone’s AI cloud,” Apple Senior Vice President of Software Engineering Craig Federighi said when announcing the service Monday at WWDC.

To achieve this, Apple has poured what I imagine is more or less a nation-state level security budget into creating a highly secure cloud-based system that provides the computational power some problems will require to be resolved.

The introduction comes at a time when providers are rolling out a range of trusted cloud and data sovereignty solutions to answer similar challenges across enterprise IT; Apple’s Private Cloud Compute service represents the best attempt yet to provide trusted cloud access to the mass market. It comes as security experts warn against unconstrained use of cloud-based genAI services in the absence of a privacy guarantee. 

(Remarkably, Elon Musk’s first reaction on hearing of Apple Intelligence was to label it a security threat, when that is precisely what it has been built not to be — you don’t have to use OpenAI at all, and I expect device management tools will be able to close off access to doing so. Perhaps a PCC-style service will eventually form part of the ecosystem for autonomous vehicles that are actually safe?)

What is Private Cloud Compute?

Private Cloud Compute consists of a network of renewable-energy powered Apple Silicon servers Apple is deploying across US data centers. These servers run Apple’s own genAI models remotely when a query demands more computational power than is available on an Apple device. (We don’t expect some of the newly introduced Apple Intelligence services to be available outside the US until 2025, likely reflecting the time it will take to deploy servers locally to support them.)

The idea is that while many Apple Intelligence tasks will run quite happily at the edge, on your device, some queries will require more computational power — and that’s where the PCC kicks in. 

But what about the data you share when making a query? Apple says you don’t need to worry, promising that the information you provide isn’t accessible to anyone other than the user, not even to Apple. 

This has been achieved through a combination of hardware, software, and an all-new operating system. The latter has been specially tailored to support Large Language Model (LLM) workloads, while presenting a very limited potential attack surface. 

This is the power of what is at core server-side Unix, coupled with Apple’s own proprietary system security software and a range of on-device, on-system hardened and highly secure components.

How does this all fit together?

The hardware itself is built around Apple Silicon, which means the company has been able to protect the servers with built-in security protections such as Secure Enclave and Secure Boot. These systems are also protected by iOS security tools, such as Code Signing and sandboxing.

To provide additional protection, Apple has closed down traditional tools such as remote shells and replaced them with purpose-built proprietary tools. There is also something Apple calls Swift on Server on which the company’s cloud services exist. The use of Swift, Apple says, ensures memory safety, which helps further limit any attack surface.

This is what happens when you make an Apple Intelligence request:

• Your device figures out if it can process the request itself.
• If it needs more computational power, it will get help from PCC.
• In doing so, the request is routed through an Oblivious HTTP (OHTTP) relay operated by an independent third party, which helps conceal the IP address from which the request came.
• It will only send data relevant to your task to the PCC servers.
• Your data is not stored at any point, including in server metrics or error logs; is not accessible; and is destroyed once the request is fulfilled.
• That also means no data retention (unlike any other cloud provider), no privileged access, and masked user identity.

Where Apple really seems to have made big steps is in how it protects its users against being targeted. Attackers cannot compromise data that belongs to a specific Private Cloud user without compromising the entire PCC system. That doesn’t just extend to remote attacks, but also to attempts made on site, such as when an attacker has gained access to the data center. This makes it impossible to grab database credentials to mount an attack.

What about the hardware?

Apple has also made the entire system open to independent security and privacy review — indeed, unless the server identifies itself as being open to such review, the information will not be transmitted — so no spoof PCC for you. 

The company didn’t stop there. “We supplement the built-in protections of Apple Silicon with a hardened supply chain for PCC hardware, so that performing a hardware attack at scale would be both prohibitively expensive and likely to be discovered,” the company said. “Private Cloud Compute hardware security starts at manufacturing, where we inventory and perform high-resolution imaging of the components of the PCC node before each server is sealed and its tamper switch is activated.”

What that means: Apple has put protections in place to maintain server security that extend all the way from the factory where those servers are made. That’s a huge step on its own account.

What about Apple’s genAI software?

Apple also maintained a focus on user security while developing the tools it makes available within Apple Intelligence. These follow what it calls its “Responsible AI principles,” as explained on the company site. These are:

• “Empower users with intelligent tools: We identify areas where AI can be used responsibly to create tools for addressing specific user needs. We respect how our users choose to use these tools to accomplish their goals.
• “Represent our users: We build deeply personal products with the goal of representing users around the globe authentically. We work continuously to avoid perpetuating stereotypes and systemic biases across our AI tools and models.
• “Design with care: We take precautions at every stage of our process, including design, model training, feature development, and quality evaluation to identify how our AI tools may be misused or lead to potential harm. We will continuously and proactively improve our AI tools with the help of user feedback.
• “Protect privacy: We protect our users’ privacy with powerful on-device processing and groundbreaking infrastructure like Private Cloud Compute. We do not use our users’ private personal data or user interactions when training our foundation models.”

Apple has also thought about whose data it uses to train its models and promises that its Apple Intelligence LLMs are trained on licensed data, “as well as publicly available data collected by our web-crawler, AppleBot.” (If you don’t want your content crawled by AppleBot for use in training models you can opt out, as explained here.)

While researchers will kick Apple’s systems around, it looks very much like the company has crafted a highly secure approach to genAI from the device used to request a service all the way to the cloud, with software and hardware protections in place every step of the way. 

What Apple has achieved

There is a lot more to consider — take a look at this white paper — but Apple has achieved something potentially very good here: an ecosystem that provides private genAI services, and can be extended over time. 

I’m seeing positive reaction from across the security community to Apple’s news.

“If you gave an excellent team a huge pile of money and told them to build the best ‘private’ cloud in the world, it would probably look like this,” Johns Hopkins cryptography lecturer Mathew Green said. He also warned that the right to opt out of using Apple Intelligence should be more visible and suggested that the impact of Apple’s move will effectively lead toward more use of cloud services.

 “We believe this is the most advanced security architecture ever deployed for cloud AI compute at scale,” said Apple Head of Security Engineering and Architecture Ivan Krstić

Is that really the case? Perhaps. Apple has promised that additional transparency to confirm its security promise is on the way. Though I do wonder how this service will gel with those bandit nations (such as the UK) that legislate for pretty much constant data surveillance in order to protect nothing much at all.

But in terms of end user protection and a fully figured out system to support cloud services, Apple’s new offering shows what every cloud service should aspire to exceed. 

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.

It will take a while to process all of Apple’s many big WWDC announcements on Monday, but one set of improvements I can’t wait to use are the ones packed inside the Apple Intelligence-augmented Mail app. 

Based on what we’ve been told, these enhancements will really help any enterprise professional, knowledge worker, or frankly anybody who uses Mail. 

I ask you, who doesn’t struggle with over-full email boxes and important responses we don’t make because “life” gets in the way? It is also interesting that those improvements mean Apple devices will now offer the tools we used to have to pay through the nose for with Grammarly.

Ease the email pain points

As had been rumored, Apple took the wraps off its plans for artificial intelligence (“Apple Intelligence) at its big developers conference. We’ll be writing about them for weeks and months to come. I’m focused now on Apple’s improvements to Mail, which should (I hope) help push the tech out of the way and let us get on with what we need to do. They do so because they begin to answer the biggest problem with email: why is an industry standard solution that’s packed with our data — and that we use every day — now a more actively useful space.

That’s because, in tandem with the company’s on-device contextual intelligence, the information coming into your email box is made more easily actionable with the changes and additions Apple plans.

Together, these should go far toward turning Mail into a central focus space from which users can complete most communication and task-related projects. And by the time the changes ship this fall, I expect Apple will deliver a powerfully integrated mail experience that helps you get work out the way — so you can focus on making better genmojis.

Just take a look at the powerful features Apple Intelligence promises us.

Powerful and useful writing tools

Apple’s new Writing Tools (available in Mail, other Apple apps and to third-party developers via an API) offer a range of functions. They can rewrite and/or proofread what you have already written; summarize your message; bullet point the key points in your message; create tables and lists; and even change the tone of your mail.

The latter feature lets you take what you’ve written and, at the tap of a button, generate (for example) a professional tone that still says what you want to say, while not being overly informal. Better yet, these tools are available system-wide, and none of what you write leaves your system, unless you make and approve a request that wants to use tools provided by OpenAI’s Chat GPT. (Yes, that pre-WWDC rumored tie-up turned out to be correct, as well.)

Smart message prioritization 

Apple Intelligence will learn which of your incoming messages are most likely to be important and file them into the Priority Messages view in your Inbox. This will make it much easier to find them.

Apple explains: “On-device categorization organizes and sorts incoming email into Primary for personal and time-sensitive emails, Transactions for confirmations and receipts, Updates for news and social notifications, and Promotions for marketing emails and coupons. Mail also features a new digest view that pulls together all of the relevant emails from a business, allowing users to quickly scan for what’s important in the moment.”

It also handles important notifications the same way.

Summaries in messages and mail

Sometimes we have little time, and yet whoever we’re interacting with has a great deal of complex information to share; the result is a lengthy email. Sure, the whole message should probably be read, but if you’re short of time, you can use Summaries in Messages to get the gist of the entire diatribe. Yes, if you completely rely on summaries you’ll probably miss something, but if you are in a hurry and just need the basics, Apple Intelligence has your back.

Built-in transcription tools

Apple Intelligence lets you generate summaries and transcripts of audio recordings captured with the Notes app or during a phone call. This is going to be popular with a lot of people — particularly researchers, students, and journalists. (If you’re concerned about privacy, all parties in a phone call will be told this activity is taking place.)

The new audio transcription and summarization features in Notes enable a device to take notes for the user, Apple says. This lets them, “stay present in a situation where they need to capture details about what’s happening,” which means you can stay focused in that meeting and still have a useful and usable aide memoire.

Natural Language search

Have you ever spent time searching Mail for a message from a specific person that contains something you vaguely recall, but which you just can’t find using standard search in the app? Apple Intelligence brings natural language search, which should make it easier to find messages, documents, and other items when you can’t quite recall what they were called or where they are.

Contextual Siri

Siri now (finally?) gains semantic search. That means it will be able to understand information and relationships it couldn’t decipher before. Apple seems to promise this extends to understanding and creating language and images, acting across apps, and simplifying and accelerating everyday tasks. You can switch between text and voice to communicate with Siri in whatever way is most appropriate. 

What about Grammarly?

Many users rely on Grammarly to improve their writing. That seems like a less essential investment now that Apple’s writing tools exist, and not only do they exist, but they do so at no charge. One more thing? Apple’s writing tools do not collect the writing work you’ve done. To be fair, Grammarly says it protects your data; but as every hacker knows, the best data you can have if you want to stay secure is data that does not exist. Apple delivers.

Whatever next?

Apple execs, over the course of there presentations, introduced a horde of improvements, many of which will I think improve the user experience across the iOS, macOS and iPadOS platforms. Many of the richer experiences it highlighted mean its entire platform ecosystem (with the weird and probably temporary exception of visionOS), is now an AI platform. It also means that whatever Apple can’t yet do on the device can be outsourced easily to partners like OpenAI. 

The deal here seems to be that Apple maintains its hold on the intention and the experience, while also delivering access to genAI tools, preserving user privacy, and massively enhancing the email experience to the benefit of all knowledge workers. The integrated AI on your device can deliver highly personalized responses based on your specific data, without anyone other than you and the device knowing anything about you.

“It’s personal, powerful, and private — and integrated into the apps you use every day,” wrote Apple CEO Tim Cook.

That privacy promise is powerful. Though, there is a catch. 

As we had heard, you’ll need a Mac or iPad running an M1 or newer chip, or an iPhone 15 series device or newer to use Apple’s AI features, though anyone with a device running iOS 18 will probably enjoy the little blip they might see appear on screen when pressing a side button. Apple’s WWDC announcements show its continued attention to detail, as does the way it has chosen to bring genAI to the rest of us, while working to protect our privacy.

No other company does this as well — and no one else just hit the ground with hundreds of millions of users in one move.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.

It looks like artificial intelligence (AI) will be the heart of Apple’s announcements at WWDC this week. Now, we think we also know what devices will be needed to support the new — beta — features if you’re a developer given the right to install the new operating systems.

What is expected

Apple is preparing to shake a little of its reality distortion dust. That magic powder will be used to show how the company has already placed AI inside its devices, and to show how with the addition of generative AI (its own and from partners) it now offers the world’s first multi-platform (PC, smartphone, tablet, watch, spatial) AI computing ecosystem.

Apple will call it “Apple Intelligence,” according to Bloomberg.

The company might not fully realize its ambitions, but this is the direction in which it is going. It is also possible that not every feature Apple wants to talk about will be fully active yet — and some might not reach our devices until 2025. 

But they’re certainly coming. Apple has, after all, been informed by cutting-edge AI research since the company’s inception, though there’s no doubt that the rate of innovation has accelerated incredibly rapidly in the last two to three years.

The current thinking: Apple now sees this AI work as transformational and thinks it will form the foundations for another decade of product innovation — the USP of which will likely be Apple’s hardware and software excellence all supported by AI and trusted cloud services.

AI improvements will include…

Along with news of partnership with third-party genAI developers, these are some of the improvements coming:

• Mail becomes far smarter. Not only will it be able to recommend genuine seeming pre-written emails to help you get through them faster, but it will be capable of automatically filtering emails into specific channels and more. All of these features will be incredibly welcome, and some extend to Messages.
• Automatic on-device transcription tools that should make most enterprise professionals happy (most of the time). You’ll never need to argue over what was discussed during a meeting again.
• AI-powered tools in Keynote and Pages to help you swiftly put together presentations and documents.
• Safari will summarize web pages, Spotlight improves, new health and AI features will appear, notifications and news reports will be summarized, and Siri will get smarter and gain more capabilities.
• AI-boosted image editing in Photos and other additions to most of Apple’s own apps.

These tools and features effectively represent a quantum leap forward for what Apple’s hardware can achieve. Those AI-powered gadgets that raised a little attention earlier this year were interesting, but these tools get even better when you can access some of them via the Apple Watch you already wear on your wrist.

What you’ll need

Not every Apple Watch, iPhone or other device is expected to be able to run some or all of the new AI features, and the current speculation is that only the most relatively recent Apple Silicon processors will make the grade.

If that is the case, this is possibly because handling the kind of computational load created by genAI requests is extremely intensive. It takes a lot of computer power to ask ChatGPT for a recipe suggestion, and even some relatively recent devices might not yet be up to the task.

What this means is that to access these services, you will need to be running an iPad or Mac with an M1 chip, or later. These services will also require at least an iPhone 15 series device. I imagine this will spark an upgrade surge as Apple customers seek to try these new services.

In use, Apple is thought to have developed an on-device computational mechanism that will choose whether to process specific tasks on the device or via M2-based servers it is now thought to be putting into place. You’ll be able to opt-in to using these services, and the company will focus on privacy.

That’s the speculation so far, we’ll learn the real deal soon at WWDC 2024.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.

DuckDuckGo has released an AI-powered portal to some of the most popular chabots and said it will not disclose or otherwise use what users type into the window to train up large language models — the basis for the generative AI (genA) tech.

“Chats are private, anonymized by us,” DuckDuckGo lead designer Nirzar Pangarkar wrote in a blog post. “Our mission is to show the world that protecting your privacy online can be easy.”

DuckDuckGo AI Chat currently allows users to access four popular AI chatbots: Open AI’s GPT 3.5 Turbo, Anthropic’s Claude 3 Haiku, Meta Llama 3, and Mistral’s Mixtral 8x7B) — the latter two, open-source models. The optional AI Chat feature is free to use within a daily limit, and can easily be switched off.

The feature can be accessed through duck.ai or duckduckgo.com/chat on a user’s search results page under the Chat tab, or via the !ai and !chat bang shortcuts. “They all take you to the same place,” Pangarkar said.

The company is also working to add access to more chat models and browser entry points. “We’re also exploring a paid plan for access to higher daily usage limits and more advanced models,” Pangarkar said.

In its blog, the company cited a recent report from the Pew Research Center showing adults in the US have a negative view of AI’s impact on privacy, even as they’re feeling more positive about the technology’s potential impact in other areas. About 80% of those familiar with AI indicated its use by companies will lead to personal information being used in ways they won’t be comfortable or for which it wasn’t originally intended to be used.

At the same time, Pew research also shows a steady uptick in the share of US adults using chatbots for work, education, and entertainment.

Last year, the search engine released an AI-powered instant answer service called DuckAssist as part of a larger plan to integrate AI across its lineup. DuckAssist uses technology from ChatGPT maker OpenAI as well as Anthropic to generate its own answers to certain types of question.

DuckAssist, however, isn’t a chatbot; the company insists it was an upgrade to its Instant Answers feature, which allows users to submit a query and get comprehensive answers without the need to click on a result.

“We believe people should be able to use the Internet and other digital tools without feeling like they need to sacrifice their privacy in the process,” Pangarkar said. “So, we meet people where they are, developing products that add a layer of privacy to the everyday things they do online. That’s been our approach across the board — first with search, then browsing, email, and now with generative AI via AI Chat.”

Google has acquired Cameyo, the maker of an alternative to Citrix for virtual application delivery, to make legacy applications easier to access and manage on ChromeOS based devices like Chromebook.

The two companies have worked together before, but ownership  will enhance Google’s ability to offer a better experience for virtual applications and positions Chromebooks as a more attractive option for businesses that still rely on legacy applications.

“By bringing the Cameyo team’s expertise in-house, we are doubling down on our commitment to delivering a streamlined experience for virtualized applications,” Google’s head of product management for ChromeOS, Naveen Viswanatha, wrote in a  blog post announcing the acquisition.

Running legacy applications in new environments

For businesses that have specific apps that run only on Windows moving to a new platform or a device such as a Chromebook that does not support Windows can be a challenge, despite Chromebooks offering significant advantages including cost for some.

Traditionally, businesses have used virtual desktop infrastructure (VDI) to operate legacy applications without having to install them on the new device. However, this approach has its challenges, including the need for specialized skill sets to manage and secure it and, from the end users’ point of view, latency in application response.

Another approach, the one taken by Cameyo, is virtual application delivery (VAD), which also runs the application on a remote server, but in a self-contained environment, isolated from the underlying operating system.

“This innovative approach to virtualization streamlines the delivery of legacy applications to any device, eliminating the need for a full desktop environment. VAD significantly simplifies application management, making it easier for IT teams to keep software up-to-date and secure, all while improving the end-user experience by making all apps accessible without needing to sign in to a separate virtual desktop environment first. With VAD, users have seamless access to both web-based and legacy apps side-by-side without having to change their behavior,” the blog post said.

Cameyo offers this solution and is a pioneer in the VAD technology, the blog stated. With Cameyo, ChromeOS can run those Windows applications as if they are residing on the device itself.

Deeper relationship

Google first incorporated Cameyo’s software into ChromeOS in 2023, offering a way for enterprises to package needed legacy Windows applications as if they were installed locally on Chromebooks.

“Recognizing the potential of VAD, we partnered with Cameyo last year to launch a seamless virtual application delivery experience fully integrated with ChromeOS — with local file system integration, ability to deliver virtual apps as progressive web apps (PWAs) and enhanced clipboard support,” the blog added. “These features ensure users can seamlessly access data and files in a secure, easy, and familiar way within virtual apps.”

The acquisition of Cameyo promises easier access to Windows apps on Chromebooks, eliminating the hassles of “complex installations and updates,” the blog post said.

PCWorld tested an early version of Cameyo’s technology back in 2015, and Computerworld looked at its usefulness in the enterprise in 2019.

Apple has an excellent reputation for keeping its devices updated, but has never committed to a fixed term for updates. That’s now changed — but at just five years its commitment seems a little disappointing. 

A UK rule, the Product Security and Telecommunications Infrastructure (PSTI) regulation, recently went into effect. Under that law, smartphone manufacturers must commit to a defined support period pertaining to the security of their devices. 

It also requires manufacturers of connected devices sold in the UK to stop selling devices encumbered by insecure passwords, provide information on how to report security issues, and commit to a minimum period during which security updates will be made available. 

To comply with the law, manufacturers are expected to provide compliance statements; Apple provides its pledge via the company’s regulatory website, where it offers documents pertaining to Apple’s iPad, iPhone, Watch, TV and Home products.  In all the filings I checked, the company says it will support these devices for a “minimum five years from the first supply data.”

Five years, that’s what we’ve got

What I hope Apple is doing with its five-year promise is under-promising so it can deliver better. Company critics have already noted that the support commitment is on first sight lower than the promise made by Android smartphone manufacturers, Samsung and Google, both of which seem to promise five years. Except, that’s not quite true — Google’s compliance statement offers numerous schedules, from two years to seven, depending on which product you choose. Given the breadth of Samsung’s smartphone catalog, it’s possible the company offers different support periods for some of its products.

What makes the idea that the five-year term Apple promises should be seen as a minimum it will guarantee — and not the maximum it will deliver — is that iOS 17 (the current iteration of Apple’s smartphone software) supports devices up to six years old, including the iPhone XS. That’s a trend Apple set in motion a few years ago when it decided to make its operating system compatible with as many device iterations as it could. 

Under pressure

It is also true that customers want to use their devices for longer. Recent CIRP data showed that around a third of all iPhones in use are three or more years old, though 61% are over two years old. Consumers today will invest extra cash on the highest-end devices that they then try to keep using as long as they can; that’s the secret sauce spicing Apple’s iPhone Pro sales.

But the decision to give Android makers easy bragging rights by offering a shorter support commitment could be a problem for the Apple. Perhaps there’s a reason Apple needs to moderate expectations.

After all, current speculation claims some of the most interesting AI-driven features of iOS 18 will only be supported on the most recent iPhones. That’s not to say every new improvement in iOS 18 will be limited in that way, but it wouldn’t be the first time that some such features are only supported on a limited number of devices.

Golden years

We might learn whether that’s true next week, when Apple discusses iOS 18 with developers at WWDC 2024. However, if tomorrow’s AI smartphones do indeed need the most recent processors, it’s logical to expect some manufacturers might eventually begin to scale back the support promise they make under the UK law.

The regulation also means that if preventable security problems emerge in connected devices from any manufacturer, in the UK at least, there will be some potential to hold the makers responsible. If they can’t guarantee security for the full extent they promise, the onus will be on them.

One group of people I do think will benefit from the five-year Apple promise will be enterprise purchasers; they now have a solemn commitment on which to estimate their future IT spending budgets.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.