Paramount Skydance is accusing Netflix of maintaining a campaign against its proposed acquisition of Warner Bros. Discovery (WBD).
In a June 5 letter (PDF) addressed to Jared A. Hughes, acting section chief of the Media, Entertainment, and Communications Section of the US Department of Justice's (DOJ's) Antitrust Division, and A. Maya Kahn, a trial attorney for the Antitrust Division, and first reported on by Politico today, Paramount chief legal officer Makan Delrahim accused Netflix of trying to influence stakeholders about the merger. The letter reads:
Indeed, Netflix’s panic-level response and scorched-earth campaign to try and poison regulators and other stakeholders against the Transaction shows just how seriously Netflix takes Paramount as a scaled competitor.
The letter from Delrahim, a former assistant attorney general for the Antitrust Division, is a response to a letter that The International Brotherhood of Teamsters sent to the DOJ in March. The teamsters' letter argued that Paramount and WBD's merger would threaten film and TV workers. The union, which has 1.3 million members, asked the DOJ to block the merger "unless substantial and enforceable safeguards are put in place to increase domestic production and protect jobs," per an announcement from the group.
© Mike Cohen/Getty Images for The New York Times
One day after WIRED revealed that Meta had quietly embedded an unreleased face-recognition system into an app installed on more than 50 million phones, the company removed it, according to a WIRED analysis of the latest version’s code.
The most recent version of Meta AI, a companion app for its line of smart glasses, strips out the unactivated software components that powered the system Meta internally called NameTag. The version published the day of WIRED’s report included several code libraries explicitly named for face recognition. Friday’s release includes none of them.
Andy Stone, Meta's vice president of communications, told WIRED on Monday that the feature is purely exploratory, adding: “No final decision has been made on what to do here, if anything.”
© David Paul Morris/Bloomberg/Getty
A drone boat picked up two US Army pilots from waters near the Strait of Hormuz after their helicopter gunship went down, US military officials said. The incident apparently represents the first time the US military has used a drone for such a rescue mission at sea.
The two crew members from the US Army AH-64 Apache were “rescued by American forces” at 7:33 pm US Eastern Time after their helicopter went down off the coast of Oman on June 8, according to a US Central Command press release. That press release mentioned support from US Navy units including the US 5th Fleet’s Task Force 59, which is charged with integrating uncrewed aerial, surface, and underwater vehicles, alongside AI, into 5th Fleet maritime operations.
Anonymous US military officials initially told CBS News that the Apache air crew was rescued by an uncrewed surface drone operated by Task Force 59 from the US Fifth Fleet in Bahrain. The officials also described the incident as the first time the military had used a drone to rescue people from the water.
© Saronic Technologies
AST SpaceMobile decoupled from recent launch disappointments by scheduling the orbital deployment of its next three satellites on a SpaceX Falcon 9 rocket.
The space-based mobile network company hopes to blast its BlueBird 8, 9 and 10 satellites into low Earth orbit (LEO) on 17 June. As ever, launches are dependent on various environmental and other factors, so precise timing is fluid.
AST SpaceMobile expects the latest satellites to deliver almost twice the data rates of its initial BlueBird models, which it noted recently hit 98.9Mb/s in the downlink.
Company president Scott Wisniewski said the significance of the satellites goes beyond expanding its constellation and coverage: they represent the culmination of an in-house manufacturing drive and bolster claims to having birds with the largest phase-arrayed antennas at LEO heights.
AST SpaceMobile stated the trio of satellites are already stacked using a proprietary architecture and ready to be integrated with the SpaceX rocket.
The scheduled launch is something of a firing back by AST SpaceMobile at commentators who questioned whether a recent failure of a Blue Origin New Glenn rocket during a test would impact a plan to conduct regular launches throughout 2026.
The post AST SpaceMobile turns to SpaceX for next launches appeared first on Mobile World Live.
A new proposed Senate bill aims to establish clear guardrails for the U.S. military’s use of artificial intelligence (AI), drawing a definitive line on one of the most consequential questions in modern warfare: who has the final say when lethal force is used?
On June 8, Senator Adam Schiff (D-Calif.) introduced the proposed Human Authority in Lethal Operations Act of 2026, or “HALO Act,” which would impose sweeping requirements on the Department of War (DoW) to ensure that autonomous and semi-autonomous weapon systems remain under meaningful human control.
The legislation would establish new design standards, testing requirements, oversight procedures, and legal safeguards for military systems that use artificial intelligence to identify, recommend, prioritize, or engage targets.
The bill aims to ensure that responsibility for battlefield decisions remains with human commanders, rather than shifting to algorithms. The legislation would require the Pentagon to designate accountable commanders for any engagement involving AI-enabled autonomous or semi-autonomous weapons and mandates that those commanders retain “ultimate discretion, judgment, and control over the use of force.”
“The past few months have shown us that there is an urgent need for commonsense guardrails to ensure the Defense Department’s use of AI is in line with Americans’ national security and privacy priorities,” Sen. Schiff said in a press release. “There are good reasons to use AI technology to advance our national security,” although he emphasized that, like with any tool, AI should not be allowed solely to guide human enterprises, “particularly when the risks of harm can be fatal.”
The proposal arrives amid growing concern among policymakers, military planners, and technology experts about the increasing role of artificial intelligence in combat operations.
Significantly, these concerns are no longer theoretical. As The Debrief previously reported, Israel’s reported use of the AI targeting system “Lavender” in Gaza triggered renewed debate over how much trust militaries should place in algorithmic target selection, particularly when civilian lives may be at risk.
Other recent examples have added to the concern, including studies showing that AI models can escalate dangerously in simulated war games, as well as the Pentagon’s continued push toward autonomous drones, AI-controlled aircraft, and robotic battlefield systems.
Even Pope Leo XIV recently weighed in on the issue in Magnifica Humanitas, his first encyclical, or formal papal teaching letter, warning that autonomous weapons could make war more feasible and less subject to human control, and arguing that lethal decisions must not be delegated to opaque or automated processes.
While the Pentagon has repeatedly stressed its commitment to responsible AI development, the HALO Act would codify many of those principles into law and create new statutory requirements governing how military AI systems are designed, tested, approved, and monitored.
Among the bill’s most significant provisions are requirements that autonomous weapon systems operate within predefined geographic areas, target sets, and timeframes. If a system cannot complete an engagement within those authorized parameters, it would be required to terminate the engagement until additional human operator and commander review is completed.
The legislation also calls for extensive record-keeping, cybersecurity protections, and continuous oversight to guard against unintended engagements and hostile manipulation.
The bill would also require senior Pentagon officials to conduct formal reviews before AI-enabled weapons enter development and again before they are fielded. Those reviews would assess legal compliance, operational safety, reliability, cybersecurity, and risks to civilians and protected infrastructure under the Law of Armed Conflict.
Beyond battlefield autonomy, the legislation targets several controversial uses of artificial intelligence in national security. It would prohibit covered Defense Department AI capabilities from inferring a person’s religion, race, political identity, sexual orientation, or emotional state. It would also restrict the purchase or receipt of Americans’ personal data from data brokers or other third-party sources when that data was obtained in a manner that would violate the bill’s rules.
Perhaps most notably, the proposed HALO Act explicitly forbids removing humans from decision-making processes involving presidential actions related to nuclear weapons.
“No covered artificial intelligence capability may be used with the intent, purpose, or outcome of… removing a human from the chain of decisionmaking for actions critical to informing and executing decisions by the President, including initiating or terminating nuclear weapons employment,” the proposed bill reads.
The measure also includes whistleblower protections for personnel working on military AI systems, mandates regular reporting to Congress, and requires the Pentagon to identify and remediate any fielded systems that fail to meet the law’s requirements.
Existing autonomous systems would undergo a comprehensive review, and any system found to be noncompliant could face retirement, modification, or suspension pending corrective action.
While the bill includes exemptions for certain defensive systems, cyber capabilities, mines, and other categories of military technology, its wider message is that as artificial intelligence becomes more deeply integrated into national defense, Congress wants humans, not machines, to remain responsible for high-stakes decisions.
The HALO Act is not the only effort underway on Capitol Hill to establish guardrails around military AI.
In March, Senator Elissa Slotkin (D-MI) introduced the AI Guardrails Act. Last week, Senator Kirsten Gillibrand (D-N.Y.) introduced separate legislation that would similarly prohibit the Pentagon from using artificial intelligence to launch nuclear weapons, conduct warrantless surveillance of Americans, or deploy fully autonomous weapons.
“Right now, the Pentagon is moving toward deploying incredibly powerful AI technology without commonsense guardrails in place, which could have catastrophic consequences that make all of us less safe,” Sen. Gillibrand said in a statement. “We must act now – not to stifle technological progress, but to establish clear rules of the road that keep humans in charge and keep AI’s use in warfare smart and safe.”
The competing proposals could set the stage for a broader congressional debate over the role of artificial intelligence in future military operations, particularly if lawmakers seek to incorporate them into the Fiscal Year 2027 National Defense Authorization Act (NDAA).
Before that can happen, however, the HALO Act and similar proposals must navigate the legislative process. The bills will first be considered by the relevant Senate committee, where lawmakers may hold hearings, debate their provisions, amend the text, or decline to advance them. If approved, it would then need to pass both the Senate and the House before being sent to the president for signature, either as a standalone measure or as part of a larger NDAA.
“My legislation would protect Americans from unlawful domestic surveillance, ensure that humans in the chain of command exercise responsibility for the use of any lethal technology, and maintain strong ethical protections in the deployment of autonomous and semi-autonomous weapons,” Sen. Schiff said in a statement.
Tim McMillan is a retired law enforcement executive, investigative reporter and co-founder of The Debrief. His writing typically focuses on defense, national security, the Intelligence Community and topics related to psychology. You can follow Tim on Twitter: @LtTimMcMillan. Tim can be reached by email: tim@thedebrief.org or through encrypted email: LtTimMcMillan@protonmail.com
As Apple announced last year, this year's macOS release will end support for Intel Macs. The macOS 27 Golden Gate release will require a Mac with an Apple Silicon chip inside, including the original M1 that launched in the MacBook Air, MacBook Pro, and Mac mini back in late 2020.
Intel Macs running macOS 26 Tahoe can expect security and Safari patches for about two more years after the release of macOS 27 Golden Gate. Macs running macOS 15 Sequoia will receive one more year of updates. Apple Silicon Macs will still be able to run Intel Mac apps via the Rosetta 2 compatibility layer in macOS 27, but future releases will begin to limit the technology (Apple has said it will mainly be used to support older games that still use Intel code).
This change has been a long time coming, and every new macOS release has left a longer and longer list of Intel Macs behind. But many Mac owners who purchased late-model Intel machines in 2019 and 2020 could still run the latest version of the operating system, and third-party utilities like the OpenCore Legacy Patcher helped more adventurous Mac owners use their unsupported hardware a bit longer.
© Apple
If you're using older iPhone or iPad hardware and you're hoping to keep running Apple's latest operating systems, this year's releases bring mostly good news. The iOS 27 update will run on all iPhones that can run iOS 26, all the way back to the iPhone 11 and second-generation iPhone SE. The iPadOS 27 update is slightly less generous, dropping support for the 3rd-generation iPad Air, 8th-generation iPad, and 5th-generation iPad mini (all of these devices used an older A12 Bionic chip; supported devices now use an A13 or better).
Apple says owners of older devices should see performance improvements in iOS 27, thanks in part to an updated CPU scheduler. This scheduler was apparently already included with newer iPhones but has been ported back to older devices with this release.
But many of the new features Apple mentioned require support for Apple Intelligence, which remains confined to newer devices with at least 8GB of RAM. Apple Intelligence still requires an iPhone 15 Pro or newer, an iPhone 16 or newer, or an iPhone Air. On the iPad, support requires an iPad Air or iPad Pro with an M1 or newer.
© Apple
Google's NotebookLM was one of the company's first forays into generative AI technology, and in un-Googley fashion, it hasn't been shut down yet. In fact, NotebookLM is getting one of its biggest updates, ever, today, moving to the latest Gemini 3.5 model, support for more file types, and streamlined web source integration. Google also says NotebookLM will be able to do more with all those queries thanks to embedded support for Antigravity.
Gemini 3.5 Flash debuted at Google I/O this year, promising much faster and more efficient processing. Google has claimed that companies worried about token costs can save big by moving their projects to the new Flash model while also getting outputs that are of similar or better quality. Those improvements are now filtering down to other Google products. NotebookLM, which launched in 2023 at the very beginning of the AI boom, lets you analyze specific sources like documents and webpages with Google's latest AI models.
The upgraded NotebookLM beats the old version in all of Google's "core evaluation dimensions."
Credit:
Google
Google conducted side-by-side evaluations of NotebookLM on the old Gemini 3.1 branch and with the updated 3.5. The company is being somewhat vague about the nature of the tests, breaking things up into "top five core evaluation dimensions," which are Accuracy and Quality, Multilingual Support, Large Document Analysis, Document Creation, and Advanced Research. In these tests, Google says NotebookLM averaged a 65 percent win rate versus the older model.
Apple has taken the wraps off of its next-generation operating system updates at its Worldwide Developers Conference today, including iOS 27, iPadOS 27, and macOS 27 Golden Gate. And while the long-awaited Siri AI update is the headliner, Apple also emphasized its efforts to optimize its software, improving its performance and reliability.
For starters, the company is continuing to refine the Liquid Glass design language that it introduced last year. A slider in the Settings will allow users fine-grained control over the translucency of the Liquid Glass effect, ranging from maximally transparent and glassy to fully tinted. Last year's redesigned icons are also being re-redesigned with more glass layers, which Apple says will make them sharper and more distinctive.
On macOS, Apple has also changed the way app toolbars and sidebars work, making toolbars more distinct, making the contents of sidebars extend all the way to the edge of the window, and reintroducing color to sidebar icons. Mac windows are also getting a "tighter corner radius," to address complaints about the way window resizing works in macOS 26 Tahoe.
© Apple
The UK government gave technology companies three months to switch on device-level tools blocking children from taking, sending or viewing nude images, warning action will be taken if they fail to do so.
The UK Home Office said companies including Apple and Google must implement built-in protections across smartphones and tablets used by children in a move it claimed would make Britain “the first country in the world where it is impossible for children to take, share or view naked pictures on their devices”
If tech companies do not act within the deadline, the government plans to enforce legislation mandating the activation of the technology. Proposed penalties could include fines, with criminal liability for tech bosses under consideration “as a last resort”.
On privacy, the government framed the proposal as on-device blocking, stating there would be “no data collection, no monitoring and no reporting”. Users over the age of 18 would still be able to access adult content by providing proof of age.
The initiative aims to stop predators from exploiting underage victims through their devices and limiting children’s access to pornography. The Home Office noted 91% of online child sexual abuse reports recorded in 2024 contained self-generated content, adding children as young as five were being groomed or coerced into creating explicit images.
Speaking at London Tech Week, Prime Minister Keir Starmer (pictured) said the industry must address the issue as a matter of urgency. “If we are serious about unlocking the opportunities that tech can bring, then we must also be serious about preventing our children from those who look to abuse it – the online predators,” he explained, arguing technology should “adapt to the needs of society, not the other way round”.
Starmer warned if technology players “choose not, then we will act and we will change the law”.
The post Big tech faces UK child safety ultimatum appeared first on Mobile World Live.
Interview: AT&T completed a controlled introduction phase of its cloud RAN deployment, with roughly 25 sites running live traffic on Intel’s Sapphire Rapids processors as it plans to aggressively scale deployment when Intel’s next-generation Granite platform arrives over the coming months.
The Sapphire deployment is currently handling the TDD layers of the network, which represent the largest share of traffic, but not yet the full stack.
Rob Soni (pictured), VP of the operator’s RAN architecture, told Mobile World Live (MWL) Granite, which is already in AT&T’s labs, will enable a full stack covering FDD, TDD, NR and LTE from a single server design.
“This is all coming in phased pieces through the fall and early next year,” he said. “We plan to scale more aggressively on Granite than we did on Sapphire.”
He said AT&T expects to move from a few hundred Sapphire sites to thousands on Granite through 2026 and into 2028.
Performance on the Sapphire sites is matching traditional RAN on call drop rates and throughput, Soni said, and the team is beginning to utilise the platform’s promise of faster feature delivery.
The leading example is AI-driven link adaptation, developed with Ericsson, which replaces rules-based scheduling algorithms with a deep neural network.
Trials are showing roughly 10% gains in spectrum efficiency and 15% improvements in individual user throughput, with further benefits expected as AT&T tunes the global Ericsson model to its own network morphology.
“We’ve done a lot of trailing with the kind of global model of Ericsson,” Soni explained. “We’re now in the process of tuning the model to the AT&T morphology, the AT&T usage. That will change the nature, and we expect there to be more gain.”
Halfway point
In cellular networks, Soni stated a base station must determine on every transmission frame which users to serve, what modulation and coding scheme to use and how to balance competing demands across the spectrum.
It is a computationally intense optimisation problem, and the traditional approach has been to encode engineering judgment into static rule sets. The AI model Ericsson developed learns from network behaviour and adapts more fluidly to changing conditions.
Link adaptation is a critical algorithm for determining the overall multiple user performance in a cellular network,” he stated.
Rival T-Mobile US stated in May 2026 it is also trialling Ericsson’s AI-native scheduler with link adaptation on live 5G-Advanced traffic, but Soni noted AT&T has been trialling it since December 2025.
Soni told MWL AT&T is just past the halfway point for switching out Nokia’s RAN gear for Ericsson’s as part of a five year plan to transition 70% of its network traffic to run on open platforms by end of year.
The post AT&T scales cloud RAN as open RAN swap passes 50% appeared first on Mobile World Live.
Northwestern University engineers have unveiled an extremely energy-efficient 3D camera, called SpiderCam, inspired by how jumping spiders judge depth before making a calculated leap.
The research team behind the novel device said their technology could help engineers build wearable technologies, assistive devices, robots, drones, and other futuristic technologies that benefit from quickly and efficiently gauging their surroundings.
According to a statement announcing SpiderCam, 3D cameras typically estimate image depth in two ways. The first approach involves captured images from multiple viewpoints and comparing them, and the second involves projecting a light source, often an infrared (IR) laser, and then measuring the reflection.
Although both methods can produce realistic three-dimensional images, they can require significant computing power, additional operational energy, and often expensive hardware such as lasers.
Fortunately, millions of years of evolution have resulted in a creature that possesses a hyper-efficient way of estimating depth: the jumping spider.
“Jumping spiders jump to catch prey, to avoid predators and to get around, and that requires excellent vision,” said Northwestern’s Emma Alexander, the study’s corresponding author and an expert in bio-inspired computer vision. However, the researcher notes that their brains are extremely small, requiring the ability to calculate distances in an energy- and computationally efficient way.
According to Alexander, jumping spider eyes have multiple layers of retinas in each eye, rather than just one as in human eyes. When measuring distance, each retina focuses on an image at a slightly different distance, resulting in some images being blurrier than others.
“They see multiple levels of focus at all times,” Alexander explained. “So, they are always collecting pairs of images.”
After collecting these images in real time, the fly’s poppyseed-sized brain compares their sharpness, yielding a surprisingly accurate depth estimate, as evidenced by the creature’s ability to jump with precision.
Inspired by nature’s hyper-efficient, extremely accurate depth estimation method, Alexander’s team decided to build a 3D camera using the same basic approach.
“We wanted to understand whether we could borrow some of the same principles to create an extremely energy-efficient depth sensor that could be used in resource-constrained situations where users don’t have unlimited access to power,” Alexander explained.
Instead of capturing images at multiple distances, as the jumping spider’s eye does, the Northwestern team’s initial design captures two images of the same scene at different focal lengths. Next, the system uses a custom algorithm to analyze the sharpness of the edges and textures captured in each image. The team said this analysis acts as a “translator between blur and distance” in the fly’s brain, resulting in “depth measurements in real time.”
To best approximate the naturally evolved energy efficiency of a jumping spider’s eye, Alexander’s team built the algorithm directly into a customizable computer chip specifically optimized for energy efficiency called a field-programmable gate array (FPGA).
The final prototype integrating the team’s dual-lens system generates depth maps at 32.5 frames per second. Critically, it generates this level of accurate 3D maps while consuming 624 milliwatts of power. For comparison, the team notes that 624 milliwatts is “less energy than used by a standard nightlight.”
Although the team said that the current version of the SpiderCam is the “first passive FPGA-based 3D camera system to operate below one watt,” they hope to improve its optics and expand its field of view. Once they have a commercial-grade version, the team hopes to integrate SpiderCam into wearable devices and small robots.
Beyond proposed applications, the Northwestern team said that they “envision” a future, ultra-energy-efficient version of the chip capable of “bringing 3D vision to applications where conventional depth sensors are impractical.”
“I’m very interested in settings where you’re very resource constrained and can’t just plug a camera into a wall,” Alexander explained. “For example, it could be deployed in field settings with limited power.”
“Separately, I also think it’s particularly exciting for applications like augmented reality where you’re interfacing with the physical world and need to know the locations of objects around you,” the researcher added.
The study’s co-first authors, Marcos Ferreira and Tianao Li, presented this work at the Computer Vision Foundation’s Conference on Computer Vision and Pattern Recognition (CVPR) in Denver. Complimentary registration is available for members of the media.
Christopher Plain is a Science Fiction and Fantasy novelist and Head Science Writer at The Debrief. Follow and connect with him on X, learn about his books at plainfiction.com, or email him directly at christopher@thedebrief.org.
Login