Apple is considering enlisting the help of OpenAI or Anthropic to power its AI-upgraded Siri, according to a report from Bloomberg’s Mark Gurman. As Apple continues to struggle with the development of an upgraded “LLM Siri,” it reportedly asked OpenAI and Anthropic to create versions of their large language models to test on the company’s private cloud infrastructure.

For months, Apple has been working to get its AI-enhanced Siri back on track after delaying the overhauled assistant’s launch in March. Apple later appointed Vision Pro head Mike Rockwell as the leader of AI and Siri after CEO Tim Cook “lost confidence” in the team’s former chief, John Giannandrea.

As reported by Bloomberg, Rockwell asked his team to test whether Anthropic’s Claude, OpenAI’s ChatGPT, or Google’s Gemini performs better at handling basic requests compared to its own models, with Anthropic’s apparently seen as the most promising. While Google has Gemini AI features for Android and its Pixel lineup, Samsung licenses Google’s AI model for its phones. It is also reportedly close to cutting a deal with Perplexity, which already has a tie-up with Motorola. Earlier this month, Bloomberg reported Apple executives had considered acquiring Perplexity to help boost its AI ambitions.

LLM Siri was largely absent from Apple’s Worldwide Developers Conference earlier this month, where SVP of worldwide marketing Greg Joswiak admitted that the technology “didn’t hit our quality standard.” Bloomberg notes that Apple’s plans to incorporate a third-party AI model into Siri are still at an “early stage” and that it’s still considering using in-house models.

In a memo to Meta staff, CEO Mark Zuckerberg announced the company’s new “Meta Superintelligence Labs” group that will head up its AI work, Bloomberg reports.

Former Scale AI CEO Alexandr Wang, who joined Meta as part of a multibillion dollar deal earlier this month, will head up the group as the company’s chief AI officer. Former GitHub CEO Nat Friedman will “partner with” Wang to help lead the division, according to Bloomberg. Meta has also made 11 new AI-focused hires, including former Anthropic, Google DeepMind, and OpenAI employees. 

Zuckerberg has been making huge offers to potential hires – The Verge’s Alex Heath has reported that they’re “well into the eight-figure range” – to bolster its AI efforts, which have fallen behind other big AI companies. Meta has also held talks to buy companies like Mira Murati’s Thinking Machines Lab, Perplexity, and Ilya Sutskever’s Safe Superintelligence, though “none of these talks progressed to the formal offer stage for various reasons,” Heath reported.

Meta is also planning to “start research on our next generation of models to get to the frontier in the next year or so,” Zuckerberg said in the memo, which was published in full by CNBC.

Google announced a variety of new features for its Classroom software suite, including free Gemini AI tools for educators and NotebookLM for users under 18 — the first time the tool has been available to minors.

Teachers with a Google Workspace account will have a new dedicated Gemini tab in their Google Classroom, offering tools that can help brainstorm lesson plans, make math problems, or draft emails, among others. For those with Education Plus Edition accounts, Gemini audio lessons will also be offered as add-ons. Another new tab for Analytics will allow teachers to track student performance, highlighting any who have been showing recent improvement or who are missing assignments and may need more support.

Google will also be opening up its NotebookLM research and note-taking tool to users under 18 “in the coming months.” Students will be able to use NotebookLM to access things like interactive study guides and podcast-style audio overviews based on source materials their teachers upload. And educators can create their own custom Gems, mini custom Gemini agents, based on their curriculum to share with students. Students will be able to interact with a Gem to get extra help or learn more about a topic it’s based on. This feature goes one step further, as well, as administrators will also be able to share Gems with the teachers they oversee in their district, to help with lesson planning or whatever else administrators might like to micromanage.

A few other new tools coming to educators include Google Vids for generative AI video creation, a new “help me create a form” feature in Gemini, and the launch of Google’s Class Tools. The Class Tools feature was announced earlier this year, and now that it’s live, it will allow teachers to manage their students’ Chromebooks — like sharing content to their screens and minimizing distractions — during a lesson. By connecting directly to their students’ accounts, teachers can push a workbook, video, presentation, or even a quiz directly to their screens.

Speaking of Chromebooks, Google is also highlighting that it has more than 12 new Chromebook, Chromebook Plus, and Chromebook OPS (Intel’s Open Pluggable Specification) devices for front-of-classroom displays coming — at some point. There’s no specific timeframe given, though the list includes the recently announced Lenovo Chromebook Plus 14 that I got a preview of, which is launching today.

With all these new Gemini tools for teachers and students, Google is also announcing that Gemini Education is now the standard tier for its base-level education accounts. It offers Gemini 2.5 Pro AI models, which educators can access with higher limits than base-level consumer accounts. Existing Gemini Education and Gemini Education Premium add-ons will be unified as Google AI Pro for Education.

When Ela Darling began performing in porn at the tail end of the 2000s, tube sites like Pornhub were relatively new, and major studios like Vivid and Digital Playground still dominated the market. For performers, the job was straightforward: show up to set, turn in a good performance, and collect a check. Everything else - from scripting the scene to editing and marketing to makeup - was generally handled by the studio's team of professionals.

Fast forward to 2025. Vivid and Digital Playground are distant memories. Sites like OnlyFans - and competitors like JustForFans, Clips4Sale, and ManyVids - now rule the roost. And while that shift has empowered many performers to create careers that would have been unimaginable in earlier eras, it's also come with a cost. While some jobs, like broadcasting and distribution, are now handled by third-party platforms like OnlyFans, a great deal of the work that used to be handled by a studio is now the performers' responsibility. In addition, these new sites heavily encourage interaction with fans. The most successful creators aren't just turning out hot porn scenes. They're constantly chatting with their subscribers and creating custom scenes …

Read the full story at The Verge.

It has become almost impossible to browse the internet without having an AI-generated video thrust upon you. Open basically any social media platform, and it won't be long until an uncanny-looking clip of a fake natural disaster or animals doing impossible things slides across your screen. Most of the videos look absolutely terrible. But they're almost always accompanied by hundreds, if not thousands, of likes and comments from people insisting that AI-generated content is a new art form that's going to change the world.

That has been especially true of AI clips that are meant to appear realistic. No matter how strange or aesthetically inconsistent the footage may be, there is usually someone proclaiming that it's something the entertainment industry should be afraid of. The idea that AI-generated video is both the future of filmmaking and an existential threat to Hollywood has caught on like wildfire among boosters for the relatively new technology.

The thought of major studios embracing this technology as is feels dubious when you consider that, oftentimes, AI models' output simply isn't the kind of stuff that could be fashioned into a quality movie or series. That's an impre …

Read the full story at The Verge.

Reddit has become known as the place to go for unfiltered answers from real, human users. But as the site celebrates its 20th anniversary this week, the company is increasingly thinking about how it can augment that human work with AI.

The initial rollout of AI tools, like Reddit Answers, is "going really well," CTO Chris Slowe tells The Verge. At a time when Google and its AI tools are going to Reddit for human answers, Reddit is going to its own human answers to power AI features, hoping they're the key to letting people unlock useful information from its huge trove of posts and communities.

Reddit Answers is the first big user-facing piece of the company's AI push. Like other AI search tools, Reddit Answers will show an AI-generated summary to a query. But Reddit Answers also very prominently links to where the content came from - and as a user, you also know that the link will point you to another place on Reddit instead of some SEO-driven garbage. It also helps that the citations feel much more prominent than on tools like Google's AI Mode - a tool that news publishers have criticized as "theft."

"If you just want the short summary, it's there," Slowe says. "If you want to …

Read the full story at The Verge.

In the past week, big AI companies have - in theory - chalked up two big legal wins. But things are not quite as straightforward as they may seem, and copyright law hasn't been this exciting since last month's showdown at the Library of Congress.

First, Judge William Alsup ruled it was fair use for Anthropic to train on a series of authors' books. Then, Judge Vince Chhabria dismissed another group of authors' complaint against Meta for training on their books. Yet far from settling the legal conundrums around modern AI, these rulings might have just made things even more complicated.

Both cases are indeed qualified victories for Meta and Anthropic. And at least one judge - Alsup - seems sympathetic to some of the AI industry's core arguments about copyright. But that same ruling railed against the startup's use of pirated media, leaving it potentially on the hook for massive financial damage. (Anthropic even admitted it did not initially purchase a copy of every book it used.) Meanwhile, the Meta ruling asserted that because a flood of AI content could crowd out human artists, the entire field of AI system training might be fundamentally at odds with fair use. And neither case a …

Read the full story at The Verge.

For years, Meta trained its AI programs using the billions of public images uploaded by users onto Facebook and Instagram's servers. Now, it's also hoping to access the billions of images that users haven't uploaded to those servers. Meta tells The Verge that it's not currently training its AI models on those photos, but it would not answer our questions about whether it might do so in future, or what rights it will hold over your camera roll images.

On Friday, TechCrunch reported that Facebook users trying to post something on the Story feature have encountered pop-up messages asking if they'd like to opt into "cloud processing", which would allow Facebook to "select media from your camera roll and upload it to our cloud on a regular basis", to generate "ideas like collages, recaps, AI restyling or themes like birthdays or graduations."

By allowing this feature, the message continues, users are agreeing to Meta AI terms, which allows their AI to analyze "media and facial features" of those unpublished photos, as well as the date said photos were taken, and the presence of other people or objects in them. You further grant Meta the right to "retain and use" that personal inform …

Read the full story at The Verge.

So far, Runway is known for bringing generative AI to Hollywood. Now, the $3 billion startup is setting its sights on the gaming industry.

This week, I was granted access to a new interactive gaming experience that Runway plans to make available to everyone as soon as next week, according to CEO Cristóbal Valenzuela. The consumer-facing product is currently quite barebones, with a chat interface that supports only text and image generation, but Valenzuela says that generated video games are coming later this year. He says that Runway is also in talks with gaming companies about both using its technology and accessing their datasets for training.

Based on his recent conversations, Valenzuela believes the gaming industry is in a similar position to Hollywood when it was first introduced to generative AI. There was considerable resistance, but over time, AI has been gradually adopted in more areas of the production process. Valenzuela says Amazon's recent show, House of David, was made in part with Runway's technology, and that his company is working with "pretty much every major studio" and "most of the Fortune 100 companies."

"If we can help a studio make a movie 40 percent fas …

Read the full story at The Verge.

Google’s 2025 sustainability report emphasizes that its “ambition-based emissions” grew 11 percent last year to reach 11.5 million metric tons of carbon dioxide pollution, marking a 51 percent increase compared to 2019. In its report, Google cites several factors outside its “direct control” that are making it more challenging to achieve its climate goals. That includes the “rapid evolution of AI” that could make the company’s “future energy needs and emissions trajectories more difficult to predict.”

That puts Google farther away from its goal of slashing planet-heating pollution in half by 2030 compared to a 2019 baseline. But what it calls “ambition-based” emissions excludes certain categories of supply chain pollution it considers out of its control or“peripheral” to the core business of Alphabet, Google’s parent company. That includes certain purchased goods and services and food programs.

AI is driving up energy usage at other companies like Microsoft and Meta, and the technology is on track to consume more power than Bitcoin by the end of this year. Though AI companies like DeepSeek are aiming to create more energy-efficient models, that prospect isn’t stopping companies like Meta from building a massive, gas-powered data center in Louisiana. President Donald Trump has also signed an executive order promoting the use of coal to power AI data centers.

A table buried in the report’s appendix shows that its total emissions actually reached 15,185,200 metric tons of carbon dioxide. That’s roughly equivalent to the emissions from nearly 40 gas-fired power plants over a year.

Google’s report also highlights a 12 percent reduction in carbon emissions from data centers, even though its energy usage continues to rise thanks to AI. Despite the decrease, carbon emissions from data centers are still far higher when compared to 2019.

The report also calls out changes to climate and energy-related policies, “slower-than-needed” deployment of carbon-free energy technologies, and a lack of carbon-free energy solutions in certain markets. “These external factors could affect the cost, feasibility, and timeline of our progress — and navigating them requires flexibility,” Google writes.