AI

Astute AI copyright observer Michael Weinberg raises some good questions about the Common Pile, an AI training dataset billed as being composed of only “openly licensed text”:

Koray Kavukcuoglu, Google DeepMind’s chief technology officer, will work to combine Google’s AI models with its products “with the goal of more seamless integration, faster iteration, and greater efficiency,” according to a memo seen by Semafor. Kavukcuoglu will reportedly remain Google DeepMind CTO while in his new role.

The Verge’s new senior AI reporter, Hayden Field, noticed we didn’t hear the name of Apple’s assistant very much during the WWDC 2025 keynote on Monday.

CEO Sam Altman says to expect it “later this summer but not June,” as reported by TechCrunch. Altman teased the model earlier this year.

AI tools from companies like Google and OpenAI are replacing traditional web searches, starving news sites of traffic. Similarweb data reported by The Wall Street Journal shows that HuffPost’s organic search traffic fell by just over half in the past three years, and by nearly that much at The Washington Post.

Founder Ren Zhengfei says Huawei’s Ascend chips are still “one generation” behind US rivals like Nvidia, and that the “US has exaggerated Huawei’s capabilities — we’re not that strong yet.”

Apple is pushing visual intelligence features that build on Apple Intelligence, allowing users to go beyond searching for context using their device’s camera and now also “search and take action on anything they’re viewing across their apps.” The feature can also recognize when a user is looking at something they may like to attend and pre-populate a calendar event with time and place.

Apple is updating Genmoji and Image Playground with new styles, powered in part by ChatGPT. With Image Playground, users can now tap into ChatGPT to change a friend’s photo into the style of an oil painting, for instance.

Apple announced at WWDC 2025 that it’s debuting live translation Apple Intelligence features that allow you to translate between languages in Messages, FaceTime, and phone calls.

This year at WWDC, Apple announced it’s opening up access for any app to tap into the on-device large language model at the core of Apple Intelligence, giving developers direct access. It will “ignite a whole new wave of intelligence experiences” in the apps users frequent, per Apple, and it cuts out cloud API costs due to on-device access.

There’s a flood of reports about college students apparently using AI to cheat their way through college, and debating whether or not that’s the actual problem. Now, NBC4 in Columbus points out Ohio State’s new AI Fluency plan that targets the class of 2029 (with some level of assumptions that our current view of generative AI will still be relevant by then), focusing on a few steps:

Right before WWDC 2025, Apple researchers published a paper called The Illusion of Thinking (PDF) that made waves. The researchers wrote that popular and buzzy AI models “face a complete accuracy collapse beyond certain complexities,” especially with things they’ve never seen before.

Our sister site Eater just did a great deep dive into why you might be seeing people comment “meatball” and other foodstuffs on cooking creators’ social media. The short of it is AI recipe and link-sharing automations that help users avoid that pesky link in bio. That said, creators have mixed feelings over chatbots taking over their comment section.

Google’s new feature in Labs will create the visualizations to “help bring financial data to life for questions on stocks and mutual funds,” according to a blog post. You can ask Google a follow-up question, too, and “AI Mode understands what to research for you.”

On Thursday, the Oversight Board overturned Meta’s decision to leave up a Facebook post showing an AI deepfake of Brazilian soccer star Ronaldo Nazário in an ad for a gambling app. The ad was viewed more than 600,000 times and received more than 50 reports.