Apple had been among the tech giants most conspicuously absent from — or at least, low-key about — the generative AI craze, at least...

Apple had been among the tech giants most conspicuously absent from — or at least, low-key about — the generative AI craze, at least until today.

At its annual Worldwide Developer Conference (WWDC 2024) in Cupertino, California, the company unveiled its biggest push into generative AI so far: a new service called Apple Intelligence, which will offer a variety of features across Apple devices including Mac computers, iPhones, and iPads.

The service is not an app per se, rather, it is a set of features embedded within other popular apps, from web browser Safari (where you can summarize articles) to Mail (where it can rewrite and suggest grammar improvements) to Photos (auto generate photo albums on specific subjects and topics set to music based on a text prompt) to Messages (where it can create custom AI generated emoji and photos of your contacts, as well as event and group photos).

As with nearly all new Apple announcements of the company’s storied history, the Apple Intelligence announcement was watched by a large audience of tech workers and journalists, as well as creatives, and some notable entrepreneurs and executives from rival firms.

It also inspired a wide range of responses, from some interpreting the announcement as underwhelming or undermining of Apple’s reputation as a company where minimalistic and clean designs are prioritized, while others viewed it as one of, if not the best examples of generative AI done right. Here are some of the most interesting reactions I saw:

High praise from former rivals

Steven Sinofsky, the former president of the Windows Division at Microsoft and current board partner at Andreessen Horowitz, called Apple Intelligence “really excellent work.”

He also said he thought the idea of weaving Apple Intelligence through its various Apple-branded apps was “exactly right and even more so when combined with privacy/on device.”

And, as if that wasn’t enough of a favorable review, Sinofsky also took the opportunity to ding Google and his own former employer Microsoft in comparison to Apple’s approach:

Similarly, Andrej Karpathy, an esteemed researcher who was previously director of artificial intelligence and Autopilot Vision at Tesla (where he competed with Apple’s abandoned self-driving car project) and a co-founder of OpenAI, said in a post on X that he found Apple Intelligence “super exciting.”

Double standard?

Bilawal Sidhu, host of the TED Talks AI Show and a former Google Maps AR/VR engineer, wrote a lengthy post on X comparing how Apple Intelligence leverages personal data on the device in which it operates, as well as virtual private clouds, to serve up AI responses — a tack that he saw as similar to Microsoft’s new Recall feature for Windows Copilot + PCs that faced intense backlash from some users and researchers for possible data security risks. Microsoft Recall was, as of last week, disabled by default and now must be turned on by the user during setup.

AI images and Genmoji: love/hate?

One of the most immediately obvious use cases for Apple Intelligence for regular users is in its ability to create custom imagery and emoji based on their text prompts within Messages and other apps.

Open source software developer and AI influencer Simon Willison took to his blog to commend Apple’s approach toward AI image generation, writing:

“This feels like a clever way to address some of the ethical objections people have to this specific category of AI tool:

  1. If you can’t create photorealistic images, you can’t generate deepfakes or offensive photos of people
  2. By having obvious visual styles you ensure that AI generated images are instantly recognizable as such, without watermarks or similar
  3. Avoiding the ability to clone specific artist’s styles further helps sidestep ethical issues about plagiarism and copyright infringement

The social implications of this are interesting too. Will people be more likely to share AI-generated images if there are no awkward questions or doubts about how they were created, and will that help it more become socially acceptable to use them?

Others criticized the look and feel of the cartoonish AI generated images in Messages:

A third-party app killer

Various users pointed out that by integrating a number of AI features across its native apps, Apple was essentially killing third-party AI-powered apps and services that sought to offer similar functionality prior to the news today and the absence of Apple Intelligence.

Questions about training data

Other users on X, including some visual artists and tech workers opposed to the practices of generative AI model providers training without express consent on vast swaths of artwork and creative work posted to the web, questioned exactly how Apple had trained its underlying Apple Intelligence AI models — the company mentioned both language and diffusion models in its keynote announcement — and on what specific data.

One Apple executive present at WWDC told Axios’s Ina Fried that the models were trained on “data from the public web” combined with licensed, or paid, data.

AI is a feature not a product?

Apple’s choice to weave the Apple Intelligence service throughout its apps also had The Information founder and CEO Jessica Lessin musing that the approach was likely to be seen as influential.

Clearly, a wide range of reactions and they’re still rolling in. What do you think about Apple Intelligence so far?

Source link