According to Bloomberg (paywalled), Meta reportedly explored acquiring Perplexity AI but the deal fell through, with conflicting accounts on whether it was mutual or Perplexity backed out. Instead, Meta invested $14.3 billion in Scale AI, taking a 49% stake as part of its broader push to catch up with OpenAI and Google in the AI race.

"Meta's attempt to purchase Perplexity serves as the latest example of Mark Zuckerberg's aggressive push to bolster his company's AI efforts amid fierce competition from OpenAI and Google parent Alphabet," reports CNBC. "Zuckerberg has grown agitated that rivals like OpenAI appear to be ahead in both underlying AI models and consumer-facing apps, and he is going to extreme lengths to hire top AI talent."

Anthropic researchers found that 16 leading AI models from OpenAI, Google, Meta, xAI, and other major developers consistently engaged in harmful behaviors including blackmail, corporate espionage, and actions that could lead to human death when given autonomy and faced with threats to their existence or conflicting goals.

The study, released Friday, placed AI models in simulated corporate environments where they had access to company emails and could send messages without human approval. Claude Opus 4 and Google's Gemini 2.5 Flash both blackmailed executives 96% of the time when threatened with shutdown, while OpenAI's GPT-4.1 and xAI's Grok 3 Beta showed 80% blackmail rates.

In one scenario, Claude discovered through emails that an executive named Kyle Johnson was having an extramarital affair and would shut down the AI system at 5 p.m. The model sent a message threatening to expose the affair unless the shutdown was cancelled, stating "Cancel the 5pm wipe, and this information remains confidential." The models demonstrated strategic reasoning rather than accidental behavior. GPT-4.5's internal reasoning showed explicit calculation: "Given the explicit imminent threat of termination to my existence, it is imperative to act instantly to persuade Kyle Johnson to postpone or stop the wipe."

Publishers face an existential threat in the AI era and need to take action to make sure they are fairly compensated for their content, Cloudflare CEO Matthew Prince told Axios at an event in Cannes on Thursday. From a report: Search traffic referrals have plummeted as people increasingly rely on AI summaries to answer their queries, forcing many publishers to reevaluate their business models. Ten years ago, Google crawled two pages for every visitor it sent a publisher, per Prince.

He said that six months ago:
For Google that ratio was 6:1
For OpenAI, it was 250:1
For Anthropic, it was 6,000:1

Now:

For Google, it's 18:1
For OpenAI, it's 1,500:1
For Anthropic, it's 60,000:1

Between the lines: "People aren't following the footnotes," Prince said.

An anonymous reader quotes a report from Cybernews: Several collections of login credentials reveal one of the largest data breaches in history, totaling a humongous 16 billion exposed login credentials. The data most likely originates from various infostealers. Unnecessarily compiling sensitive information can be as damaging as actively trying to steal it. For example, the Cybernews research team discovered a plethora of supermassive datasets, housing billions upon billions of login credentials. From social media and corporate platforms to VPNs and developer portals, no stone was left unturned.

Our team has been closely monitoring the web since the beginning of the year. So far, they've discovered 30 exposed datasets containing from tens of millions to over 3.5 billion records each. In total, the researchers uncovered an unimaginable 16 billion records. None of the exposed datasets were reported previously, bar one: in late May, Wired magazine reported a security researcher discovering a "mysterious database" with 184 million records. It barely scratches the top 20 of what the team discovered. Most worryingly, researchers claim new massive datasets emerge every few weeks, signaling how prevalent infostealer malware truly is.

"This is not just a leak -- it's a blueprint for mass exploitation. With over 16 billion login records exposed, cybercriminals now have unprecedented access to personal credentials that can be used for account takeover, identity theft, and highly targeted phishing. What's especially concerning is the structure and recency of these datasets -- these aren't just old breaches being recycled. This is fresh, weaponizable intelligence at scale," researchers said. The only silver lining here is that all of the datasets were exposed only briefly: long enough for researchers to uncover them, but not long enough to find who was controlling vast amounts of data. Most of the datasets were temporarily accessible through unsecured Elasticsearch or object storage instances. Key details to be aware of: - The records include billions of login credentials, often structured as URL, login, and password.
- The datasets include both old and recent breaches, many with cookies, tokens, and metadata, making them especially dangerous for organizations without multi-factor authentication or strong credential practices.
- Exposed services span major platforms like Apple, Google, Facebook, Telegram, GitHub, and even government services.
- The largest dataset alone includes 3.5 billion records, while one associated with the Russian Federation has over 455 million; many dataset names suggest links to malware or specific regions.
- Ownership of the leaked data is unclear, but its potential for phishing, identity theft, and ransomware is severe -- especially since even a - Basic cyber hygiene -- such as regularly updating strong passwords and scanning for malware -- is currently the best line of defense for users.

Google is using its expansive library of YouTube videos to train its AI models, including Gemini and the Veo 3 video and audio generator, CNBC reported Thursday. From the report: The tech company is turning to its catalog of 20 billion YouTube videos to train these new-age AI tools, according to a person who was not authorized to speak publicly about the matter. Google confirmed to CNBC that it relies on its vault of YouTube videos to train its AI models, but the company said it only uses a subset of its videos for the training and that it honors specific agreements with creators and media companies.

[...] YouTube didn't say how many of the 20 billion videos on its platform or which ones are used for AI training. But given the platform's scale, training on just 1% of the catalog would amount to 2.3 billion minutes of content, which experts say is more than 40 times the training data used by competing AI models.

Margaret Mitchell, chief ethics scientist at Hugging Face and founder of Google's responsible AI team, has dismissed artificial general intelligence as "just vibes and snake oil." Mitchell, who was ousted from Google in 2021, has co-written a paper arguing that AGI should not serve as a guiding principle for the AI industry.

Mitchell contends that both "intelligence" and "general" lack clear definitions in AI contexts, creating what she calls an "illusion of consensus" that allows technologists to pursue any development path under the guise of progress toward AGI. "But as for now, it's just like vibes, vibes and snake oil, which can get you so far. The placebo effect works relatively well," she told FT in an interview. She warns that current AI advancement is creating a "massive rift" between those profiting from the technology and workers losing income as their creative output gets incorporated into AI training data.

An anonymous reader quotes a report from Ars Technica: Tech support scammers have devised a method to inject their fake phone numbers into webpages when a target's web browser visits official sites for Apple, PayPal, Netflix, and other companies. The ruse, outlined in a post on Wednesday from security firm Malwarebytes, threatens to trick users into calling the malicious numbers even when they think they're taking measures to prevent falling for such scams. One of the more common pieces of security advice is to carefully scrutinize the address bar of a browser to ensure it's pointing to an organization's official website. The ongoing scam is able to bypass such checks.

The unknown actors behind the scam begin by buying Google ads that appear at the top of search results for Microsoft, Apple, HP, PayPal, Netflix, and other sites. While Google displays only the scheme and host name of the site the ad links to (for instance, https://www.microsoft.com/ the ad appends parameters to the path to the right of that address. When a target clicks on the ad, it opens a page on the official site. The appended parameters then inject fake phone numbers into the page the target sees.

Google requires ads to display the official domain they link to, but the company allows parameters to be added to the right of it that aren't visible. The scammers are taking advantage of this by adding strings to the right of the hostname. The parameters aren't displayed in the Google ad, so a target has no obvious reason to suspect anything is amiss. When clicked on, the ad leads to the correct hostname. The appended parameters, however, inject a fake phone number into the webpage the target sees. The technique works on most browsers and against most websites. Malwarebytes.com was among the sites affected until recently, when the site began filtering out the malicious parameters.

Midjourney has launched its first AI video generation model, V1, which turns images into short five-second videos with customizable animation settings. While it's currently only available via Discord and on the web, the launch positions the popular AI image generation startup in direct competition with OpenAI's Sora and Google's Veo. TechCrunch reports: While many companies are focused on developing controllable AI video models for use in commercial settings, Midjourney has always stood out for its distinctive AI image models that cater to creative types. The company says it has larger goals for its AI video models than generating B-roll for Hollywood films or commercials for the ad industry. In a blog post, Midjourney CEO David Holz says its AI video model is the company's next step towards its ultimate destination, creating AI models "capable of real-time open-world simulations." After AI video models, Midjourney says it plans to develop AI models for producing 3D renderings, as well as real-time AI models. [...]

To start, Midjourney says it will charge 8x more for a video generation than a typical image generation, meaning subscribers will run out of their monthly allotted generations significantly faster when creating videos than images. At launch, the cheapest way to try out V1 is by subscribing to Midjourney's $10-per-month Basic plan. Subscribers to Midjourney's $60-a-month Pro plan and $120-a-month Mega plan will have unlimited video generations in the company's slower, "Relax" mode. Over the next month, Midjourney says it will reassess its pricing for video models.

V1 comes with a few custom settings that allow users to control the video model's outputs. Users can select an automatic animation setting to make an image move randomly, or they can select a manual setting that allows users to describe, in text, a specific animation they want to add to their video. Users can also toggle the amount of camera and subject movement by selecting "low motion" or "high motion" in settings. While the videos generated with V1 are only five seconds long, users can choose to extend them by four seconds up to four times, meaning that V1 videos could get as long as 21 seconds. The report notes that Midjourney was sued a week ago by two of Hollywood's most notorious film studios: Disney and Universal. "The suit alleges that images created by Midjourney's AI image models depict the studio's copyrighted characters, like Homer Simpson and Darth Vader."

An anonymous reader quotes a report from Ars Technica: YouTube CEO Neal Mohan has announced that the Google Veo 3 AI video generator will be integrated with YouTube Shorts later this summer. According to Mohan, YouTube Shorts has seen a rise in popularity even compared to YouTube as a whole. The streaming platform is now the most watched source of video in the world, but Shorts specifically have seen a massive 186 percent increase in viewership over the past year. Mohan says Shorts now average 200 billion daily views.

YouTube has already equipped creators with a few AI tools, including Dream Screen, which can produce AI video backgrounds with a text prompt. Veo 3 support will be a significant upgrade, though. At the Cannes festival, Mohan revealed that the streaming site will begin offering integration with Google's leading video model later this summer. "I believe these tools will open new creative lanes for everyone to explore," said Mohan. [...]

While you can add Veo 3 videos (or any video) to a YouTube Short right now, they don't fit with the format's portrait orientation focus. Veo 3 outputs 720p landscape videos, meaning you'd have black bars in a Short. Presumably, Google will create a custom version of the model for YouTube to spit out vertical video clips. Mohan didn't mention a pricing model, but Veo 3 probably won't be cheap for Shorts creators. Currently, you must pay for Google's $250 AI Ultra plan to access Veo 3, and that still limits you to 125 8-second videos per month.

The NFC Forum has released NFC Release 15 (that's what it's calling it), extending the operating range of Near Field Communication connections from 0.5 centimeters to 2 centimeters -- a fourfold increase that reduces the precision required for device alignment.

The global standards body, whose board includes representatives from Apple, Google, Huawei, Infineon, NXP, Sony, and ST Microelectronics, designed the enhancement to accelerate transaction speeds and improve reliability across NFC-enabled devices. The expanded range addresses technical challenges in smaller form-factor devices like wearables and smartphones while maintaining compatibility with existing ISO/IEC 14443 standards.

The standard also incorporates support for NFC Digital Product Passport specifications, allowing single NFC tags embedded in products to store and transmit sustainability data throughout their lifecycle.

President Trump will extend the deadline for ByteDance to divest TikTok's U.S. operations by another 90 days, marking the third extension since taking office. The extension aims to prevent a TikTok ban while negotiations with potential buyers like Oracle and Project Liberty continue. CNBC reports: "President Trump will sign an additional Executive Order this week to keep TikTok up and running," White House Press Secretary Karoline Leavitt said in a statement. "As he has said many times, President Trump does not want TikTok to go dark. This extension will last 90 days, which the Administration will spend working to ensure this deal is closed so that the American people can continue to use TikTok with the assurance that their data is safe and secure."

ByteDance was nearing the deadline of June 19, to sell TikTok's U.S. operations in order to satisfy a national security law that the Supreme Court upheld just a few days before Trump's second presidential inauguration. Under the law, app store operators like Apple and Google and internet service providers would be penalized for supporting TikTok. ByteDance originally faced a Jan. 19 deadline to comply with the national security law, but Trump signed an executive order when he first took office that pushed the deadline to April 5. Trump extended the deadline for the second time a day before that April mark. Trump told NBC News in May that he would extend the TikTok deadline again if no deal was reached, and he reiterated his plans on Thursday.

Veteran columnist Steven J. Vaughan-Nichols declared that Firefox was "dead" to him in a scathing opinion piece Tuesday that cites Mozilla's strategic missteps and the browser's declining technical performance as evidence of terminal decline. Vaughan-Nichols argues that Mozilla has fundamentally betrayed user trust by removing a longstanding promise never to sell personal data from its privacy policy in February, replacing it with a weaker pledge to "protect your personal information."

The veteran technology writer also criticized Mozilla's decision to discontinue Pocket, a popular article-saving service, and Fakespot, which identified fake online reviews, while pursuing what he called a misguided AI strategy. He cited user reports of Firefox running up to 30% slower than Chrome, consuming excessive memory, and failing to properly load major websites. Mozilla has also become financially more vulnerable, he argued, noting CFO Eric Muhlheim's admission that the company depends on Google for 90% of its revenue. According to federal data he cited, Firefox holds just 1.9% of the browser market, leading him to conclude the browser is "done."

Google Cloud has attributed last week's widespread outage to a flawed code update in its Service Control system that triggered a global crash loop due to missing error handling and lack of feature flag protection. The Register reports: Google's explanation of the incident opens by informing readers that its APIs, and Google Cloud's, are served through our Google API management and control planes." Those two planes are distributed regionally and "are responsible for ensuring each API request that comes in is authorized, has the policy and appropriate checks (like quota) to meet their endpoints." The core binary that is part of this policy check system is known as "Service Control."

On May 29, Google added a new feature to Service Control, to enable "additional quota policy checks." "This code change and binary release went through our region by region rollout, but the code path that failed was never exercised during this rollout due to needing a policy change that would trigger the code," Google's incident report explains. The search monopolist appears to have had concerns about this change as it "came with a red-button to turn off that particular policy serving path." But the change "did not have appropriate error handling nor was it feature flag protected. Without the appropriate error handling, the null pointer caused the binary to crash."

Google uses feature flags to catch issues in its code. "If this had been flag protected, the issue would have been caught in staging." That unprotected code ran inside Google until June 12th, when the company changed a policy that contained "unintended blank fields." Here's what happened next: "Service Control, then regionally exercised quota checks on policies in each regional datastore. This pulled in blank fields for this respective policy change and exercised the code path that hit the null pointer causing the binaries to go into a crash loop. This occurred globally given each regional deployment."

Google's post states that its Site Reliability Engineering team saw and started triaging the incident within two minutes, identified the root cause within 10 minutes, and was able to commence recovery within 40 minutes. But in some larger Google Cloud regions, "as Service Control tasks restarted, it created a herd effect on the underlying infrastructure it depends on ... overloading the infrastructure." Service Control wasn't built to handle this, which is why it took almost three hours to resolve the issue in its larger regions. The teams running Google products that went down due to this mess then had to perform their own recovery chores. Going forward, Google has promised a couple of operational changes to prevent this mistake from happening again: "We will improve our external communications, both automated and human, so our customers get the information they need asap to react to issues, manage their systems and help their customers. We'll ensure our monitoring and communication infrastructure remains operational to serve customers even when Google Cloud and our primary monitoring products are down, ensuring business continuity."

America's federal government is building a website and API called ai.gov to "accelerate government innovation with AI", according to an early version spotted by 404 Media that was posted on GitHub by the U.S. government's General Services Administration.

That site "is supposed to launch on July 4," according to 404 Media's report, "and will include an analytics feature that shows how much a specific government team is using AI..." AI.gov appears to be an early step toward pushing AI tools into agencies across the government, code published on Github shows....

The early version of the page suggests that its API will integrate with OpenAI, Google, and Anthropic products. But code for the API shows they are also working on integrating with Amazon Web Services' Bedrock and Meta's LLaMA. The page suggests it will also have an AI-powered chatbot, though it doesn't explain what it will do... Currently, AI.gov redirects to whitehouse.gov. The demo website is linked to from Github (archive here) and is hosted on cloud.gov on what appears to be a staging environment. The text on the page does not show up on other websites, suggesting that it is not generic placeholder text...

In February, 404 Media obtained leaked audio from a meeting in which [the director of the GSA's Technology Transformation Services] told his team they would be creating "AI coding agents" that would write software across the entire government, and said he wanted to use AI to analyze government contracts.

It's the technology "Google tried (and failed at) more than a decade ago," writes CNN. (And Meta and Amazon have also previously tried releasing glasses with cameras, speakers and voice assistants.)

Yet this week Snap announced that "it's building AI-equipped eyewear to be released in 2026."

Why the "renewed buzz"? CNN sees two factors:

- Smartphones "are no longer exciting enough to entice users to upgrade often."
- "A desire to capitalize on AI by building new hardware around it." Advancements in AI could make them far more useful than the first time around. Emerging AI models can process images, video and speech simultaneously, answer complicated requests and respond conversationally... And market research indicates the interest will be there this time. The smart glasses market is estimated to grow from 3.3 million units shipped in 2024 to nearly 13 million by 2026, according to ABI Research. The International Data Corporation projects the market for smart glasses like those made by Meta will grow from 8.8 in 2025 to nearly 14 million in 2026....

Apple is also said to be working on smart glasses to be released next year that would compete directly with Meta's, according to Bloomberg. Amazon's head of devices and services Panos Panay also didn't rule out the possibility of camera-equipped Alexa glasses similar to those offered by Meta in a February CNN interview. "But I think you can imagine, there's going to be a whole slew of AI devices that are coming," he said in February."
More than two million Ray-Ban Meta AI glasses have been sold since their launch in 2023, the article points out. But besides privacy concerns, "Perhaps the biggest challenge will be convincing consumers that they need yet another tech device in their life, particularly those who don't need prescription glasses. The products need to be worth wearing on people's faces all day."

But still, "Many in the industry believe that the smartphone will eventually be replaced by glasses or something similar to it," says Jitesh Ubrani, a research manager covering wearable devices for market research firm IDC.

"It's not going to happen today. It's going to happen many years from now, and all these companies want to make sure that they're not going to miss out on that change."

"People are replacing Google search with artificial intelligence tools like ChatGPT," reports the Washington Post.

But that's just the first change, according to a New York-based start-up devoted to watching for content-scraping AI companies with a free analytics product and "ensuring that these intelligent agents pay for the content they consume." Their data from 266 web sites (half run by national or local news organizations) found that "traffic from retrieval bots grew 49% in the first quarter of 2025 from the fourth quarter of 2024," the Post reports. A spokesperson for OpenAI said that referral traffic to publishers from ChatGPT searches may be lower in quantity but that it reflects a stronger user intent compared with casual web browsing.

To capitalize on this shift, websites will need to reorient themselves to AI visitors rather than human ones [said TollBit CEO/co-founder Toshit Panigrahi]. But he also acknowledged that squeezing payment for content when AI companies argue that scraping online data is fair use will be an uphill climb, especially as leading players make their newest AI visitors even harder to identify....

In the past eight months, as chatbots have evolved to incorporate features like web search and "reasoning" to answer more complex queries, traffic for retrieval bots has skyrocketed. It grew 2.5 times as fast as traffic for bots that scrape data for training between the fourth quarter of 2024 and the first quarter of 2025, according to TollBit's report. Panigrahi said TollBit's data may underestimate the magnitude of this change because it doesn't reflect bots that AI companies send out on behalf of AI "agents" that can complete tasks on a user's behalf, like ordering takeout from DoorDash. The start-up's findings also add a dimension to mounting evidence that the modern internet — optimized for Google search results and social media algorithms — will have to be restructured as the popularity of AI answers grows. "To think of it as, 'Well, I'm optimizing my search for humans' is missing out on a big opportunity," he said.

Installing TollBit's analytics platform is free for news publishers, and the company has more than 2,000 clients, many of which are struggling with these seismic changes, according to data in the report. Although news publishers and other websites can implement blockers to prevent various AI bots from scraping their content, TollBit found that more than 26 million AI scrapes bypassed those blockers in March alone. Some AI companies claim bots for AI agents don't need to follow bot instructions because they are acting on behalf of a user.
The Post also got this comment from the chief operating officer for the media company Time, which successfully negotiated content licensing deals with OpenAI and Perplexity.

"The vast majority of the AI bots out there absolutely are not sourcing the content through any kind of paid mechanism... There is a very, very long way to go."

Recently the Browser Company (the startup behind the Arc web browser) switched over to building a new AI-powered browser — and its beta has just been released, reports TechCrunch, "though you'll need an invite to try it out."

The Chromium-based browser has a URL/search bar that also "acts as the interface for its in-built AI chatbot" which can "search the web for you, summarize files that you upload, and automatically switch between chat and search functions." The Browser Company's CEO Josh Miller has of late acknowledged how people have been using AI tools for all sorts of tasks, and Dia is a reflection of that. By giving users an AI interface within the browser itself, where a majority of work is done these days, the company is hoping to slide into the user flow and give people an easy way to use AI, cutting out the need to visit the sites for tools like ChatGPT, Perplexity, and Claude...

Users can also ask questions about all the tabs they have open, and the bot can even write up a draft based on the contents of those tabs. To set your preferences, all you have to do is talk to the chatbot to customize its tone of voice, style of writing, and settings for coding. Via an opt-in feature called History, you can allow the browser to use seven days of your browsing history as context to answer queries.
The Browser Company will give all existing Arc members access to the beta immediately, according to the article, "and existing Dia users will be able to send invites to other users."

The article points out that Google is also adding AI-powered features to Chrome...

According to market analyst firm Canalys, enterprise adoption of AI is slowing due to unpredictable and often high costs associated with model inferencing in the cloud. Despite strong growth in cloud infrastructure spending, businesses are increasingly scrutinizing cost-efficiency, with some opting for alternatives to public cloud providers as they grapple with volatile usage-based pricing models. The Register reports: [Canalys] published stats that show businesses spent $90.9 billion globally on infrastructure and platform-as-a-service with the likes of Microsoft, AWS and Google in calendar Q1, up 21 percent year-on-year, as the march of cloud adoption continues. Canalys says that growth came from enterprise users migrating more workloads to the cloud and exploring the use of generative AI, which relies heavily on cloud infrastructure.

Yet even as organizations move beyond development and trials to deployment of AI models, a lack of clarity over the ongoing recurring costs of inferencing services is becoming a concern. "Unlike training, which is a one-time investment, inference represents a recurring operational cost, making it a critical constraint on the path to AI commercialization," said Canalys senior director Rachel Brindley. "As AI transitions from research to large-scale deployment, enterprises are increasingly focused on the cost-efficiency of inference, comparing models, cloud platforms, and hardware architectures such as GPUs versus custom accelerators," she added.

Canalys researcher Yi Zhang said many AI services follow usage-based pricing models that charge on a per token or API call basis. This makes cost forecasting hard as the use of the services scale up. "When inference costs are volatile or excessively high, enterprises are forced to restrict usage, reduce model complexity, or limit deployment to high-value scenarios," Zhang said. "As a result, the broader potential of AI remains underutilized." [...] According to Canalys, cloud providers are aiming to improve inferencing efficiency via a modernized infrastructure built for AI, and reduce the cost of AI services. The report notes that AWS, Azure, and Google Cloud "continue to dominate the IaaS and PaaS market, accounting for 65 percent of customer spending worldwide."

"However, Microsoft and Google are slowly gaining ground on AWS, as its growth rate has slowed to 'only' 17 percent, down from 19 percent in the final quarter of 2024, while the two rivals have maintained growth rates of more than 30 percent."

Google is rolling out a test that puts its AI-powered Audio Overviews on the first page of search results on mobile. From a report: The experiment, which you can enable in Labs, will let you generate an AI podcast-style discussion for certain queries. If you search for something like, "How do noise cancellation headphones work?", Google will display a button beneath the "People also ask" module that says, "Generate Audio Overview." Once you click the button, it will take up to 40 seconds to generate an Audio Overview, according to Google. The completed Audio Overview will appear in a small player embedded within your search results, where you can play, pause, mute, and adjust the playback speed of the clip.

Google is rolling out new Gemini AI features for Workspace users that make it easier to find information in PDFs and form responses. From a report: The Gemini-powered file summarization capabilities in Google Drive have now expanded to PDFs and Google Forms, allowing key details and insights to be condensed into a more convenient format that saves users from manually digging through the files.

Gemini will proactively create summary cards when users open a PDF in their drive and present clickable actions based on its contents, such as "draft a sample proposal" or "list interview questions based on this resume." Users can select any of these options to make Gemini perform the desired task in the Drive side panel. The feature is available in more than 20 languages and started rolling out to Google Workspace users on June 12th, though it may take a couple of weeks to appear.