Category: Technology

The Information reported that an AI agent within Meta took unauthorized action that led to an employee creating a security breach at the social company last week. According to the publication, an employee used an in-house agentic AI to analyze a query from a second employee on an internal forum. The AI agent posted a response to the second employee with advice even though the first person did not direct it to do so. 

The second employee took the agent’s recommended action, sparking a domino effect that led to some engineers having access to Meta systems that they shouldn’t have permission to see. A representative from the company confirmed the incident to The Information and said that “no user data was mishandled.” Meta’s internal report indicated that there were unspecified additional issues that led to the breach. A source said that there was no evidence that anyone took advantage of the sudden access or that the data was made public during the two hours when the security breach was active. However, that may be the result of dumb luck more than anything else. 

Many tech leaders and companies have touted the benefits of artificial intelligence, this is just the latest incident where human employees have lost control over an AI agent. Amazon Web Services experienced a 13-hour outage earlier this year that also (apparently coincidentally) involved its Kiro agentic AI coding tool. Moltbook, the social network for AI agents recently acquired by Meta, had a security flaw that exposed user information thanks to an oversight in the vibe-coded platform.

This article originally appeared on Engadget at https://www.engadget.com/ai/a-meta-agentic-ai-sparked-a-security-incident-by-acting-without-permission-224013384.html?src=rss 

During a Senate hearing, FBI Director Kash Patel confirmed that his agency has bought information that could be used to track individuals’ movement and location. “We do purchase commercially available information that’s consistent with the Constitution and the laws under the Electronic Communications Privacy Act, and it has led to some valuable intelligence for us,” he said. 

Law enforcement is required to obtain a warrant in order to get location data from cell service providers following the Carpenter v United States ruling from 2018. But why bother with all that hassle when they can just buy the information from the open market?

“Doing that without a warrant is an outrageous end run around the Fourth Amendment, it’s particularly dangerous given the use of artificial intelligence to comb through massive amounts of private information,” Sen. Ron Wyden, (D-Ore.) said during the Intelligence Committee hearing. Wyden is one of several lawmakers pushing for an overhaul of when and how the government can obtain citizens’ personal information. 

It’s an overhaul that’s badly needed. Patel already has a history of dubious use of government resources, such as ordering SWAT protections for his girlfriend and somehow horning in on men’s hockey victory celebrations at the recent winter Olympics, so one would hope he’s not also stretching the limits of the few privacy protections that do exist. Then outside the FBI, we have the Department of Homeland Security being sued for illegally tracking immigration raid protestors and the Pentagon’s labeling of Anthropic as a supply-chain risk after the AI company refused to let its products be used for mass surveillance of Americans.

This article originally appeared on Engadget at https://www.engadget.com/big-tech/the-fbi-confirms-its-buying-americans-location-data-230835196.html?src=rss 

The White House has been promising a set of national rules to guide artificial intelligence since late last year, and today Sen. Marsha Blackburn (R-Tenn.) fired the first volley. The senator shared a discussion draft for codifying the executive order signed by President Donald Trump in December calling for an AI bill. Her stated goal is a policy that “protects children, creators, conservatives and communities from harm.”

Blackburn has called for tougher policies for AI safety, and one of the core messages in this discussion draft is that it “places a duty of care on AI developers in the design, development and operation of AI platforms to prevent and mitigate foreseeable harm to users.” It also draws a line on the many copyright infringement questions raised by creative industries: “an AI model’s unauthorized reproduction, copying, or processing of copyrighted works for the purpose of training, fine-tuning, developing, or creating AI does not constitute fair use under the Copyright Act.” 

Some of the other notable provisions are:

Requires covered online platforms, including social media platforms, to implement tools and safeguards to protect users under the age of 17 against online harms.

Protects the voice and visual likenesses of individuals and creators from the proliferation of digital replicas without their consent.

Sets new federal transparency guidelines for marking, authenticating and detecting AI-generated content.

Requires certain companies and federal agencies to issue reports on AI-related job effects, including layoffs and job displacement to the U.S. Department of Labor (DOL) on a quarterly basis.

It includes ending Section 230, marking the latest attempt to retire a law that has been questioned as a possible loophole for AI companies to escape liability when their tools cause harm. While AI critics might see positive signs here, remember that this is just the initial version of the framework. Lawmakers will likely spend a lot of time negotiating over the eventual result, which may be notably de-fanged from its current state. It could wind up with a lot more requirements echoing this Republican complaint: “Combats the consistent pattern of bias against conservative figures demonstrated by AI systems by requiring third-party audits to prevent discrimination based on political affiliation.” Despite the claims of suppression and censorship, we’ve consistently seen this conservative argument to be false — or at the very least misleading.

This article originally appeared on Engadget at https://www.engadget.com/ai/senator-blackburn-introduces-the-first-draft-of-a-federal-ai-bill-202509852.html?src=rss 

Google and cybersecurity companies Lookout and iVerify have detailed a new hacking technique that potentially puts a significant portion of iPhone users in danger, just by visiting the wrong web page. The hack is called “DarkSword” and since it specifically targets several different versions of iOS 18, it could affect “close to a quarter of iPhones,” Wired writes.

DarkSword is a “fileless” hack that leverages a collection of exploits to access sensitive data when an iPhone visits an infected website. Rather than install spyware that hangs around on a user’s phone after messages and other private information are stolen, fileless hacks like DarkSword take control of “the legitimate processes in an iPhone’s operating system to steal data,” according to Wired. Even more troubling, DarkSword deletes any evidence it was running on an iPhone after it finishes stealing your information.

The hack starts as soon as an iOS device encounters an “malicious iframe embedded in a web page,” after which it works its way through your iPhone, gathering sensitive information like passwords before deleting itself. DarkSword can abscond with things like messages and iCloud content, but it’s also specifically designed to access crypto currency wallets, Lookout says, which could indicate who was using DarkSword before it became widely available.

DarkSword has reportedly been used in Ukraine, Saudi Arabia, Malaysia, Turkey and Russia, and its origins could be tied to a different hacking toolkit called Coruna that TechCrunch reports may have been created for the US government by a company called Trenchant. Regardless of where DarkSword came from, the tool didn’t become widely available until its Russian users left DarkSword’s source code on a website for anyone to access, “complete with explanatory comments in English that describe each component and include the ‘DarkSword’ name for the tool,” Wired writes.

Apple patched the exploits that DarkSword and Coruna used in recent updates to iOS 26, the yearly software release from 2025 that followed iOS 18. The problem is that not everyone is using Apple’s latest update. DarkSword targets iOS 18 releases between iOS 18.4 and iOS 18.6.2, and according to Apple’s latest iOS usage stats for developers, around 24 percent of iOS devices are still on iOS 18. Without more detail, it’s hard to know how many people that leaves exposed, but as a rule of thumb, if your iOS device can update to a newer software release, you should do so as soon as possible to stay secure.

This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/a-new-iphone-hacking-tool-puts-anyone-still-on-ios-18-at-risk-203745949.html?src=rss 

Perplexity’s Comet AI browser is now available as a standalone app for iPhone users. The tool initially debuted last summer on PCs, but cost $200 a month. The new app is free, as is the recently-released Android version.

Comet is an AI-powered browser, which has become a hot product category throughout the past year or so. This is basically a web browser combined with a chatbot that can perform some tasks on a user’s behalf. Like related tools, people can ask it to summarize a webpage or conduct additional research for more context about a subject.

Perplexity says that Comet “acts as a personal assistant and thinking partner” to “turn curiosity into momentum.” Those are certainly words. The company boasts that folks can use the tool to shop and make schedules. I advise some caution for both of these tasks, given that AI browsers are notorious marks that easily fall for various online scams.

One cool thing here is that the app takes full advantage of Apple’s Liquid Glass technology. It certainly looks easy on the eyes and that address bar really pops. The iOS version does suffer from some limitations inherent to Apple’s walled-garden approach, according to a preview by MacStories. Users cannot install third-party extensions here, though it can be made the default browser.

Finally, there’s a privacy concern worth considering. Perplexity has been open about the fact that it uses browsers in part to collect customer data for ad targeting. There’s a reason why something that used to cost $200 is now available for free.

The Comet browser is now available for iPhones, Android devices, Windows PCs and Macs. There isn’t a native iPad app yet, but the standard Perplexity app is available for Apple’s tablets so maybe the browser will follow suit.

This article originally appeared on Engadget at https://www.engadget.com/apps/perplexitys-comet-ai-browser-is-now-available-for-iphones-183947569.html?src=rss 

Microsoft is rolling out a number of Xbox Series X|S updates to those on its Xbox Insiders program, and one of them has been requested by the community for years. You’ll soon be able to disable the Quick Resume feature for specific games, meaning they’ll boot up fresh each time you launch them.

As a reminder, Quick Resume is a feature that Microsoft introduced with its current generation of consoles. It lets you suspend a number of games simultaneously so they’re running in the background even when the console is turned off. When you want to return to a game that supports Quick Resume, it’ll drop you back in exactly where you left it within a few seconds, often even when it’s been months since you last played that game. It’s a pretty neat feature, and still impressive more than five years since the Series X|S launched.

But over the years, players have discovered that Quick Resume isn’t ideal for every title, particularly always-online games. When you take these out of their suspended state, they’ll often shout at you for letting them disconnect, forcing you to manually close the game down properly and then relaunch it. Other games just don’t seem to get on with it for a variety of other reasons. You can manually close any game you want at any time, but it takes a few more button presses than Quick Resume, so it’s easy to forget.

The latest Xbox update finally addresses this, letting you turn off the feature on a game-by-game basis by selecting “More Options” and then “Manage Quick Resume” on a game tile. That way you can ensure it’s only enabled for single-player titles in which the ability to instantly pick up where you left off is very appealing. It’s a change that has gone down well with ex-Xbox icon Larry “Major Nelson” Hryb, who said on X that he first requested it five years ago when still employed by Microsoft.

Microsoft is also adding more customization options to the Xbox dashboard in the upcoming update. You’ll be able to create custom background colors, rather than having to select one of the existing color options, while the amount of Groups you can add to your Home screen has been increased from two to 10. These can be reordered in the same way you reorder individual game tiles.

Select Xbox Insiders can start using the new features now, before they’re eventually made available to everyone.

This article originally appeared on Engadget at https://www.engadget.com/gaming/xbox/microsoft-will-finally-let-you-turn-off-quick-resume-for-individual-games-174639730.html?src=rss