TECH

Craig Federighi and John Giannandrea talk about Apple Intelligence at WWDC

iJustin, John Giannandrea, and Craig Federighi (left to right)

Privacy at the Core

The battle took place in the Steve Jobs Theater and was hosted by the ever-popular iJustine, who quizzed Craig Federighi and John Giannandrea on the announcements.

While Apple is focusing heavily on Apple Intelligence, it’s easy to see that it’s not standing still. Your iPhone already has more than 200 machine learning models that do everything from detecting a car crash to suggesting carefully chosen memories.

“We don’t want AI to replace our users, we want it to augment them,” Federighi said. He’s pushing the idea that these aren’t the same AI features we’ve seen before.

Traditional chatbots don't know much about you, but they do know more about the world around you. This is what Apple calls “personal intelligence.”

For example, if you ask a simple question about how long it will take to get to your son's baseball game, the chatbot will not answer. He doesn't know how you travel — car, public transport or walk — he also doesn't know about the baseball game or who your son is.

Giannandrea believes it's Apple's focus on privacy that allows them to use so much of your information. Information that other chatbots won't even touch due to privacy implications.

How Apple uses AI

Chatbots, according to Federighi, are the most open way to study AI and its weaknesses. Apple's approach was to focus on a set of solutions that it knew would be reliable and work best.

This results in a more curated set of use cases for Apple users, even if some feel it's limited.

Despite this, the AI ​​uses quite a lot of resources. That's why it's limited to Apple's M series and A17 Pro chips, which are found in the latest iPhone 15 Pro and iPhone 15 Pro Max. They require too much processing power and consume too much power to be implemented on anything else.

Federighi notes that the Neural Engine in the A17 Pro is twice as powerful as the previous version and has other advantages that suit Apple Intelligence.

Training Apple's AI models was actually ethical

Besides which devices are supported, another commonly asked feature is what data Apple used to train its AI models.

Giannandrea says Apple uses a variety of sources. The company includes publicly available web data — where publishers can freely opt out of licensed data such as stock photos, textbooks, etc.

To be clear, Apple paid for the licensed data, as we said in April. Other AI companies have said they shouldn't do it because it's too difficult to get the data right.

The fine-tuning and training dataset was created by Apple itself.

Private Cloud Computing

“We're very excited about this because it solves a big problem around with many artificial intelligence capabilities,” Federighi said when talking about the new private cloud computing infrastructure.The idea is that most of what you do will be processed locally on your phone, but in some cases, when the topic is too broad, it's better to use larger servers. That's what Apple tried to do, but with a focus on privacy.

The pair of executives were keen to highlight security concerns with other cloud providers. Even if a company says it won't do anything with your data, things can change at any time.

Only those who manage this server really knew how secure your data was and what could happen to it. It can be saved in a log file, in a database, or linked to your profile.

When using private cloud computing, only a small part of the data is uploaded to the servers, which, of course, is also anonymized. These servers are not capable of logging or storing anything.

“It's very important to know that no one, not even Apple, has access to the data used to process your request,” Federighi notes.

What's more, your devices won't even communicate with servers that don't have publicly available software. This allows independent third parties to verify Apple's privacy claims, setting a high bar.

These servers, by the way, run on Apple Silicon. This is not only because it's incredibly energy efficient, which is critical to Apple's environmental efforts, but also because they're secure from the start with features like Secure Enclave, Secure Boot, and more.

And yes, they all run on 100% renewable energy.

Siri & ChatGPT

AppleInsider reported news ahead of WWDC about many of the new features we'll see in Siri. For now, this is just the beginning of Apple's efforts.

“This is a huge journey for us and our developers,” adds Giannandrea as they talked about how developers can leverage these features by adding app intent.

App intents are individual app functions, such as applying a filter to a photo, sending an email, or purchasing a game in the Playstation app. They require developer work, but can unlock many new AI features.

Siri can also forward certain requests to other LLMs such as ChatGPT. At launch it will offer integration with ChatGPT 4, but they are open to working with others, including Google's Gemini.

Federighi sees this as a way to feed Siri data that Apple isn't an expert at. For example, we could see models full of medical data, legal data, models for knowledge encoding or creative writing.

Users will be able to add models of their choice that suit what they need to do. However, when this happens, it becomes clear that you are transferring this data to another provider and what their data policies are.

More to come from Apple Intelligence

Apple Intelligence is just the beginning. Technically, it doesn't even exist yet, as it's not in the first beta versions of Apple's new operating systems.

It will appear in a later beta version and will launch as a beta feature this fall.

“We're just starting to look at these major trends. The advent of the Internet, mobility. Now we have the iPhone… 18 is a number in the operating system,” he muses. Federighi. “I think it's been an incredible start for the iPhone to see where it's come over the years, and it's the start of a long and exciting journey.

“What excites me is that this Developer Tool, as a platform that brings users and developers together as a nexus, I think will create incredible value and be exciting for years to come,” Federighi concluded.

Follow AppleInsider on Google News.

Leave a Reply