AI Weekly: Apple and others are democratizing AI tools, and that’s both good and bad

This week was Apple’s Worldwide Developers Conference in San Jose, California, an annual gathering of app developers, industry analysts, and media eager to get a glimpse at the Cupertino company’s newest innovations. It was a bit subdued by most accounts, but Apple’s characterization of artificial intelligence as an empowering, democratizing technology was undoubtedly one of the highlights.

During the keynote, Craig Federighi, head of software at Apple, announced Core ML 2, a new and improved version of its machine learning software development kit for iOS. It’s 30 percent faster thanks to batch prediction, and it can compress machine learning models by up to 75 percent with the help of quantization.

It’s not quite as flexible as Google’s ML Kit, an AI framework for the search giant’s Firebase development platform, or similar tools from Microsoft and Facebook. But Federighi claims its advantage is tight integration with existing developer workflows.

“We’re really unleashing this capability for this vast developer community,” he told Wired in an interview this week.

Apple’s other big AI-related announcement was Create ML, a GPU-accelerated tool for AI vision and natural language model training on Mac computers. It’s coded in Swift, Apple’s open source programming language, and it fully supports Xcode, the company’s integrated development environment (IDE) for macOS, iOS, and tvOS.

The power of Create ML is not to be understated. Because it supports Swift, developers can use simple drag-and-drop interfaces like Xcode Playgrounds. Training a basic algorithm is as simple as dragging and dropping a few files.

In a WWDC developer session following the keynote, Apple suggested a few use cases for Core ML and Create ML, like predicting the quality of wine from certain characteristics or sussing out the sentiment of online comments.

“We have such a vibrant community of developers,” Federighi said. “We saw that if we could give them a big leg up toward incorporating machine learning into their apps, they would do some really interesting things.”

Apple isn’t the only company this week that took an optimistic view of AI.

At Computex 2018 in Taiwan, Nvidia announced Isaac, a hardware and software platform that makes it easier for developers to design machine learning-powered robots.

“AI is the most powerful technology force of our time,” Nvidia CEO Jensen Huang said in a statement. “Its first phase will enable new levels of software automation that boost productivity in many industries. Next, AI, in combination with sensors and actuators, will be the brain of a new generation of autonomous machines.”

And on June 4, the Mozilla Foundation, the nonprofit behind the Firefox web browser, announced a $ 225,000 fund in partnership with the Knight Foundation, MacArthur Foundation, and Open Society Foundation for projects that “explore artificial intelligence and machine learning” and are accessible to “broad audiences.”

Apple, Nvidia, and Mozilla’s efforts stand in stark contrast to the Pentagon’s explorations in machine learning. This week, Reuters reported that the U.S. military is developing an AI algorithm that can track and target nuclear missiles.

An equally unnerving use of AI came from a team of researchers at the Massachusetts Institute of Technology, who trained an algorithm — nicknamed “Norman” after the murderous main character in Alfred Hitchcock’s classic horror film Psycho — on “data from the dark corners of the net.” When tasked with interpreting Rorschach-like ink blots, the AI described gruesome scenes of people being electrocuted, jumping from windows, getting pulled into dough machines, and being shot to death.

Tools like Apple’s Create ML are the harbinger of transformative apps and hardware. But autonomous tanks, military drone programs like Project Maven, and accidents involving self-driving cars AI are sobering reminders that the ethics around AI is sorely lagging behind the technology.

For AI coverage, send news tips to Kyle Wiggers and Khari Johnson, and guest post submissions to Cosette Jarrett — and be sure to bookmark our AI Channel.

Thanks for reading,

Kyle Wiggers
AI Staff Writer

P.S. Please enjoy the keynote presentation from Apple’s Worldwide Developers Conference 2018 (skip to 2:02:00 for the segment on AI):

[embedded content]

From VB

Amazon Echo speakers and Alexa will be available in France starting next week

Amazon today announced that Alexa and Echo speakers are now available in France. Echo and Echo Dot speakers officially go on sale in France next week, while Echo Spot will be available starting next month. The Alexa Skills Kit, which allows developers to make voice apps for Alexa, is now generally available in France.

Amazon introduces 4 pre-built PC form factors for Alexa

Amazon is making it easier for PC manufacturers to integrate Alexa into PCs. On Tuesday, the retailer announced four new pre-tested, white-box form factors designed with its voice assistant in mind from original design manufacturers (ODMs) partners.

Apple’s Core ML 2 vs. Google’s ML Kit: What’s the difference?

At Apple’s Worldwide Developers Conference on Tuesday, the Cupertino company announced Core ML 2, a new version of its machine learning software development kit (SDK) for iOS devices. Here’s how it’s different from Google’s ML Kit.

Apple’s shortcuts gives developers a reason to care about Siri

Apple introduced shortcuts with Siri at WWDC, and it’s giving iOS app developers a reason to build with Siri they didn’t have before.

The AI winter is well on its way

GUEST: Deep learning has been at the forefront of the AI revolution for years. Companies like Tesla and others talked big in 2014, 2015, and 2016 when technologies such as Alpha Go were pushing new boundaries. We are now in the middle of 2018 and things have changed.

Nvidia launches Isaac robot platform with Jetson Xavier

At Computex 2018 in Taiwan, Nvidia CEO Jensen Huang launched its Isaac robot platform to power the next generation of autonomous machines, bringing artificial intelligence capabilities to robots for manufacturing, logistics, agriculture, construction, and many other industries.