Tag Archives: Internet & Networking Technology

Meta wants its Llama AI in Britain’s public healthcare system

Posted on by

Meta is making a pitch to get its AI into the UK’s public health system. The Guardian reported on Tuesday that the company held a hackathon in Europe, tasking over 200 developers to use its Llama AI to improve the country’s health services. The company awarded funds for developing AI that shortens wait times in Britain’s A&E rooms (ERs in the US).

The UK’s AI minister, Feryal Clark, told The Guardian that the “government can adopt AI, such as Meta’s open-source model, to support our key missions.” Earlier this month, Meta CEO Mark Zuckerberg gave the green light for Llama to work with the US government. Bloomberg reported on the same day that the company was also working with governments and contractors in Canada, Australia, New Zealand and — as we can see movement toward now — the UK.

Given that Britain’s use of the open-source AI model wouldn’t provide a direct or immediate financial pipeline back to the company, The Guardian asked Meta’s president of global affairs, Nick Clegg, what was in it for Meta. “It is in the long run indirectly in our interest to see this ecosystem of Llama-based innovation because it then makes it much easier for us to reincorporate innovations that are out there into our own products,” Clegg, a former UK deputy prime minister, said.

Mark Zuckerberg said in August that Llama was approaching 350 million downloads, which he credited to the AI model’s open-source nature.

Clegg sounded dismissive — perhaps mockingly so — of fears surrounding AI infiltrating governments. “Who knows, maybe AI will start developing a mind of its own and will start wanting to turn us into paper clips by next Tuesday,” he told The Guardian. “But I think that right now, the technology is way more primitive than a lot of the fears suggest.”

To be fair to critics, “right now” is less of a concern than later.

The UK AI minister Clark told The Guardian that the government wouldn’t shy away from the significant risk AI represents. Instead, it would “make sure that any regulation we introduce is proportionate, supports innovation and does not place undue burden on business.”

This article originally appeared on Engadget at https://www.engadget.com/ai/meta-wants-its-llama-ai-in-britains-public-healthcare-system-174119281.html?src=rss

The EU publishes the first draft of regulatory guidance for general purpose AI models

Posted on by

On Thursday, the European Union published its first draft of a Code of Practice for general purpose AI (GPAI) models. The document, which won’t be finalized until May, lays out guidelines for managing risks — and giving companies a blueprint to comply and avoid hefty penalties. The EU’s AI Act came into force on August 1, but it left room to nail down the specifics of GPAI regulations down the road. This draft (via TechCrunch) is the first attempt to clarify what’s expected of those more advanced models, giving stakeholders time to submit feedback and refine them before they kick in.

GPAIs are those trained with a total computing power of over 10²⁵ FLOPs. Companies expected to fall under the EU’s guidelines include OpenAI, Google, Meta, Anthropic and Mistral. But that list could grow.

The document addresses several core areas for GPAI makers: transparency, copyright compliance, risk assessment and technical / governance risk mitigation. This 36-page draft covers a lot of ground (and will likely balloon much more before it’s finalized), but several highlights stand out.

The code emphasizes transparency in AI development and requires AI companies to provide information about the web crawlers they used to train their models — a key concern for copyright holders and creators. The risk assessment section aims to prevent cyber offenses, widespread discrimination and loss of control over AI (the “it’s gone rogue” sentient moment in a million bad sci-fi movies).

AI makers are expected to adopt a Safety and Security Framework (SSF) to break down their risk management policies and mitigate them proportionately to their systemic risks. The rules also cover technical areas like protecting model data, providing failsafe access controls and continually reassessing their effectiveness. Finally, the governance section strives for accountability within the companies themselves, requiring ongoing risk assessment and bringing in outside experts where needed.

Like the EU’s other tech-related regulations, companies that run afoul of the AI Act can expect steep penalties. They can be fined up to €35 million (currently $36.8 million) or up to seven percent of their global annual profits, whichever is higher.

Stakeholders are invited to submit feedback through the dedicated Futurium platform by November 28 to help refine the next draft. The rules are expected to be finalized by May 1, 2025.

This article originally appeared on Engadget at https://www.engadget.com/ai/the-eu-publishes-the-first-draft-of-regulatory-guidance-for-general-purpose-ai-models-223447394.html?src=rss

A Google ‘test’ will omit EU publishers from news links

Posted on by

Google is conducting a “test” that will omit results from EU-based news publishers. The company says the time-limited trial will only affect a small portion of users in nine EU countries and will help “assess how results from EU news publishers impact the search experience for our users and traffic to publishers.” But given the fragile state of the news media — and the company’s history of threatening to pull its services in the face of news-related regulations — it’s tempting to view it as the equivalent of a mob boss conducting a “little test” to see how the corner laundromat fares without its protection.

Google describes the experiment (via The Verge) as a “small, time-limited test” to omit EU results from search, Google News and the personalized Discover feed. It will only affect one percent of users in Belgium, Croatia, Denmark, France, Greece, Italy, the Netherlands, Poland and Spain. Those users will still see results from other websites, including non-EU news publications.

The company says news results will reappear as usual once the test concludes. (It didn’t list a specific timeframe.) Google stresses that the experiment won’t impact the publisher payments it makes under the European Copyright Directive (EUCD), under which the company has inked deals with over 4,000 EU publishers.

Google does have a history of using the potential withdrawal of its visibility as a negotiating stick in similar situations. In some cases, the tactic has helped it draw concessions.

Last year, Google pulled its news links from Canada in response to Bill C-18 (the Online News Act), which required tech companies to negotiate compensation with online publishers for linked content. After months of negotiations, Google said Canada had addressed its concerns and given it a path to an exemption. Canada said it granted one to Google last month, with the company agreeing to pay $100 million annually to news organizations.

In April of this year, Google briefly removed links to California news outlets in response to the proposed California Journalism Preservation Act (CJPA), which would require Google to pay news publishers in exchange for continuing to link to their websites. Although the bill’s fate is still up in the air, Google struck a deal with state lawmakers this summer, committing tens of millions of dollars to a fund supporting local news.

In 2021, the company threatened to remove its entire search engine from Australia in response to a then-proposed law requiring tech companies to share royalties with news publishers. The nation’s then-Prime Minister stood firm. “Let me be clear. Australia makes our rules for things you can do in Australia,” Scott Morrison said. After the bill was passed and enacted, Google struck deals with Australian media companies to license content.

Google says it hopes the data analysis tools it provides publishers will help them use the EU test to “understand traffic patterns.”

This article originally appeared on Engadget at https://www.engadget.com/big-tech/a-google-test-will-omit-eu-publishers-from-news-links-184536615.html?src=rss

Meta cuts the price of its ad-free plan by 40 percent in a bid to sate EU regulators

Posted on by

Meta has long been at loggerheads with European Union officials over its approach to targeted Facebook and Instagram ads. The company is hoping to placate regulators with some changes to its ad model in the bloc, which includes lowering the price of its ad-free subscription. Starting November 13, the plan will cost 40 percent less — €6 ($6.36) per month for signups via the web and €8 ($8.48) for those who subscribe on an iOS or Android device. The fee for each additional Facebook and Instagram account is €4 per month on the web and €5 via mobile.

The company will automatically drop current subscribers down to the lower pricing. It says that it will once again ask users in the bloc if they'd like to sign up.

When they see this prompt (which can only be ignored for a certain period of time), there will be a third option for EU Facebook and Instagram users to choose from. Those who don't want to pay for a subscription can instead opt to only view ads that are based on what they see in a given session in the apps. Meta will also factor in a few key data markers such as "a person’s age, location, gender and how a person engages with ads."

These less-personalized ads naturally won't be as tailored to a given user's interests, the company notes. As such, people are perhaps less likely to click on such ads. To make up for that (and make sure this option doesn't hit Meta in the pocket too hard), folks who choose the less-personalized ads option will sometimes encounter unskippable ads. According to The Wall Street Journal, these will be displayed full screen.

"Such ad breaks are common across other services, and are already offered by many of our competitors," Meta argues. "This change will help us continue to provide value to advertisers which ensures we can offer people a less personalized ads experience at no charge."

Targeted ads are Meta's biggest revenue driver, but EU officials have reportedly been pressuring the company to offer a free, less-personalized option in its apps. Meta has argued that would negatively impact its bottom line. Although it has seemingly caved to officials' requests, the unskippable ad aspect may be construed as malicious compliance, as it worsens the user experience.

Meta claims that these changes to its ad model "meet EU regulator demands and go beyond what’s required" by the bloc's laws. The company introduced its ad-free subscription a year ago to comply with laws such as the Digital Markets Act (DMA), as well as stricter interpretations of the General Data Protection Regulation. It was previously ordered to seek permission from users in the bloc before showing them personalized ads.

The EU didn't take too kindly to the paid ad-free approach, however. An investigation into the "consent or pay" model is ongoing. In July, the EU said that in its preliminary findings, Meta was violating the DMA with this plan.

These latest changes are said to be Meta's attempt to settle the case, but according to the Journal, the EU's discussions with the company haven't concluded. The bloc's regulatory body has until late March to finish its investigation and make a final decision. If it determines that Meta has indeed violated the DMA, the company could be on the hook for a fine of up to 10 percent of its annual global revenue. Based on its total revenue for 2023, it could have to pay up as much as $13 billion or so.

This article originally appeared on Engadget at https://www.engadget.com/big-tech/meta-cuts-the-price-of-its-ad-free-plan-by-40-percent-in-a-bid-to-sate-eu-regulators-174926790.html?src=rss

Google’s Nest Wi-Fi Pro 6E is cheaper than ever in this early Black Friday deal

Posted on by

For many homes, a single router just won't cut it. Unless the device is in a prime position at the center of your home and there are few walls and other obstacles for the signal to deal with, your Wi-Fi network may not reach the outer limits of your house or property. As such, a mesh network might be what you need. A three-pack of Google's Nest Wi-Fi Pro 6E is worth considering, especially because it has dropped to its lowest price to date.

You can pick up this bundle for $279 at both Amazon and the Google Store. The three-pack normally retails for $400.

This is our pick for the best mesh Wi-Fi system for folks who are using the tech for the first time. There isn't a great deal of customization here and there are more powerful options out there. But the Nest mesh system is easy to set up and then just let it do its thing.

Together, these three nodes provide up to 6,600 sq. ft. of Wi-Fi coverage and speeds of up to 2,402 Mbps. That's more than enough for most people. Though if you're gaming, you'll probably want to connect to Ethernet if possible to minimize lag.

As you might expect, the Wi-Fi Pro 6E integrates with other Google services. It's compatible with the Matter and Thread smart home standards too.

Check out all of the latest Black Friday and Cyber Monday deals here.

This article originally appeared on Engadget at https://www.engadget.com/deals/googles-nest-wi-fi-pro-6e-is-cheaper-than-ever-in-this-early-black-friday-deal-191945827.html?src=rss

Matter’s latest version will try to smooth out the smart home platform’s rough edges

Posted on by

The latest version of Matter, the open smart home standard that largely hasn’t yet lived up to its lofty promises for seamless device-agnostic integration, is now available. Hardware makers and platforms can integrate Matter 1.4 into their products on Thursday. The new version’s biggest feature may be Enhanced Multi-Admin, which aims to improve interoperability between different device ecosystems.

The Connectivity Standards Alliance, which manages Matter, describes version 1.4 as “a significant step forward in the Matter ecosystem.” It says the update resulted from the coalition’s hundreds of member companies (including Apple, Amazon, Google and Samsung) working with thousands of engineers and product experts. Given the platform’s struggles to deliver on its promise of a smooth and headache-free smart home, it could use all the help it can get.

One of the new highlights, Enhanced Multi-Admin, strives for less setup when connecting to multiple ecosystems (like Apple HomeKit, Google Home or Alexa). Matter 1.4 only requires user consent once. After that (at least in theory), new and existing devices will connect to multiple ecosystems without making you approve the same device for each additional platform. If hardware manufacturers implement the feature without issues (again, a big ask based on Matter’s rocky first two years), this could improve on a big pain point and move the platform’s big promises of seamless integration closer to reality.

The update also opens the door to Matter-certified routers and access points that securely store your Thread protocol credentials. The Alliance says the networking hardware will “provide the foundational infrastructure of smart homes by combining both a Wi-Fi access point and a Thread Border Router, ensuring these ubiquitous devices have the necessary infrastructure for Matter products using either of these technologies.”

Following the addition of basic energy management features in version 1.3, Matter 1.4 expands on that with support for new device types. Solar panels, batteries (including battery walls, storage units and Battery Energy Storage Systems), heat pumps and water heaters can now work with the standard. In addition, it now covers Electric Vehicle Supply Equipment (EVSE). So, EV owners can use the management features to specify when they want their car to be charged, choosing times based on convenience or cost.

Matter 1.4 also adds protocols to lengthen battery life for low-powered devices like switches, buttons and sensors that only connect intermittently. More flexible automation options for products like in-wall switches that control lights, fans and other non-smart appliances are also on the menu. This addresses yet another pain point in previous versions, which treated all of these device types as “lights.”

This article originally appeared on Engadget at https://www.engadget.com/home/smart-home/matters-latest-version-will-try-to-smooth-out-the-smart-home-platforms-rough-edges-130003372.html?src=rss

The Internet Archive’s Wayback Machine is fully back in action with saving pages

Posted on by

The Internet Archive is continuing the recovery process after a series of DDoS attacks that took down its servers in early October. On Monday, the nonprofit digital library posted on X that its 'Save Page Now' service has been restored to the Wayback Machine.

The Wayback Machine resumed operation in read-only mode on October 14; now users can upload new web pages to record their information and access them later. As the X post notes, the Wayback Machine will begin collecting web pages that have been archived since October 9 when the entire site was taken down.

The October DDoS attacks coincided with the Internet Archive's move to disclose a data breach that saw more than 31 million records taken. Security researcher Troy Hunt, who runs the Have I Been Pwned? service for monitoring compromised accounts, said that the two actions against the Internet Archive were "entirely coincidental" and likely taken by "multiple parties."

This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/the-internet-archives-wayback-machine-is-fully-back-in-action-with-saving-pages-223736576.html?src=rss

Mozilla Foundation cuts 30 percent of its staff

Posted on by

The Mozilla Foundation has made steep cuts to its team. The 30 percent staff reduction will completely eliminate the nonprofit's advocacy and global programs divisions, according to an internal email sent by Mozilla Foundation Executive Director Nabiha Syed. TechCrunch reported on the news, and Mozilla confirmed the cuts in a statement to that publication as well as to Engadget. 

"The Mozilla Foundation is reorganizing teams to increase agility and impact as we accelerate our work to ensure a more open and equitable technical future for us all," according to the statement from Brandon Borrman, vice president of communications at Mozilla. "That unfortunately means ending some of the work we have historically pursued and eliminating associated roles to bring more focus going forward." 

While the standalone advocacy division will end, Mozilla did affirm its commitment to that work. "We want to clarify that the restructuring has not dropped advocacy; on the contrary, advocacy is still a central tenet of Mozilla Foundation’s work and we are in the process of revisiting our approach to it," Borrman said in an additional comment.

The Mozilla Foundation is a nonprofit arm under the bigger Mozilla umbrella; Mozilla Corporation is the arm responsible for the Firefox browser. In February, Mozilla announced that about 60 workers would be laid off, primarily on the product development team.

Update, November 5, 2024, 4:56PM ET: Post has been updated with additional comment from Mozilla.

This article originally appeared on Engadget at https://www.engadget.com/computing/mozilla-foundation-cuts-30-percent-of-its-staff-203951504.html?src=rss

Meta opens its Llama AI models to government agencies for national security

Posted on by

Meta is opening up its Llama AI models to government agencies and contractors working on national security, the company said in an update. The group includes more than a dozen private sector companies that partner with the US government, including Amazon Web Services, Oracle and Microsoft, as well as defense contractors like Palantir and Lockheed Martin.

Mark Zuckerberg hinted at the move last week during Meta’s earnings call, when he said the company was “working with the public sector to adopt Llama across the US government.” Now, Meta is offering more details about the extent of that work.

Oracle, for example, is “building on Llama to synthesize aircraft maintenance documents so technicians can more quickly and accurately diagnose problems, speeding up repair time and getting critical aircraft back in service.” Amazon Web Services and Microsoft, according to Meta, are “using Llama to support governments by hosting our models on their secure cloud solutions for sensitive data.”

Meta is also providing similar access to Llama to governments and contractors in the UK, Canada, Australia and New Zealand, Bloomberg reported. In a blog post, Meta’s President of Global Affairs, Nick Clegg, suggested the partnerships will help the US compete with China in the global arms race over artificial intelligence. “We believe it is in both America and the wider democratic world’s interest for American open source models to excel and succeed over models from China and elsewhere,” he wrote. “As an American company, and one that owes its success in no small part to the entrepreneurial spirit and democratic values the United States upholds, Meta wants to play its part to support the safety, security and economic prosperity of America – and of its closest allies too.”

This article originally appeared on Engadget at https://www.engadget.com/ai/meta-opens-its-llama-ai-models-to-government-agencies-for-national-security-182355077.html?src=rss

How we test VPNs

Posted on by

VPNs, or virtual private networks, are everywhere. If you’ve spent any time at all on YouTube, chances are you’ve seen an ad selling VPNs — there have been literally hundreds and thousands of them. 

We test and review VPNs so you don't have to — but what even is a VPN? The pitch is fairly simple: a VPN can mask your IP address and the identity of your device by routing your traffic through a remote server. Sites and services will see a different IP address to your own, and your internet service provider will only be able to see that you’re connected to a VPN server, rather than what sites you’re visiting. That means a VPN can be an important tool to keep your data private and secure, especially if you need to connect to an unsecured network. If you’re on public WiFi, for example, you can’t be sure what they’re tracking while you’re surfing the internet. With a VPN, there’s an additional layer of privacy — as long as you make sure the VPN provider can be trusted. But a VPN should not be the end of your security journey. Complex passwords, multifactor authentication and other security basics should be way higher on your checklist.

Another commonly advertised feature of VPNs is evading geoblocking techniques that companies use to prevent you from accessing certain content. Because you’re connecting to a remote server, a VPN can effectively change your location, allowing you to watch another country’s Netflix content or, say, watch Doctor Who on the UK-only BBC iPlayer. If you’re in the EU, it can allow you to access sites that are otherwise blocked due to the region’s strict data protection regulations. In the past, VPNs have even been used to skirt political firewalls put up to censor online access.

With many VPN makers making sweeping claims about military-grade encryption, digital invisibility and lightning-fast speeds, it’s tough to know what’s true and what’s marketing hype. That’s why Engadget started evaluating VPNs in 2023, looking at a number of factors such as security, speed, latency, usage limits and price. And while we don't expect to ramp up to more VPN testing until 2025, our best VPNs guide will continue to be the home base of our coverage. In the meantime, this is our current VPN testing methodology. 

Before we install or sign up for a service, our work begins by investigating each product’s lineage. We reference privacy policies, transparency reports and security audits that are publicly available, and note the security specs of each service. We also look into each company’s history of security incidents like data breaches. We rely on some pre-existing academic work from bodies like Consumer Reports, VPNalyzer and others when looking into security specs.

The next step in our review process involves a close look at what it’s like to get set up on the VPN. This involves looking at the various platforms each VPN supports, from certain web browsers to Smart TVs, to better understand how it can be used. We also look to see just how easy it is to get started on a new VPN. Some will automatically connect to a secure server every time you use your device, while others make you jump through hoops to stay signed in. We seek to understand how user friendly and intuitive these VPNs can be.

While many VPN providers promise that their service works across every device all the time, there is often small print to consider. We scan the terms of service for each of the company’s plans to understand what they are offering. Is there a limit to the number of devices a user can have connected at once? Is there a “fair use” limit that kicks in after a certain amount of traffic? Are advanced features multi-hop connections locked to only the priciest plans? We investigate all of these things so you don’t have to. We also conduct many of our more specific tests at the same time, which allows us to verify companies’ claims about simultaneous device use.

The first step in testing speed and latency is to ascertain the baseline of the connection before connecting to a VPN. We then use the “quick connect” feature on VPN apps to connect to the “fastest” provider available when testing internet speed. Once connected, we run internet speed tests by Ookla and ping tests with meter.net. When relaying these results, we will note the average speed and ping across these results, as well as the difference between that average and our baseline when not connected to a VPN.

We test all VPNs from within the US. To check that a service is able to effectively avoid geoblocking, we connect to a Canada-based server and attempt to access content on Netflix that is only available to users in that country. We then connect to a Hong Kong-based server to attempt to watch a news livestream on a YouTube channel that is exclusively available to users in that region. Finally, we conduct a gaming test by playing on servers in the United Kingdom. We’re looking to see not only if we can access the same content from all these servers, but also to test for lag or any other headaches that you could run into when using something.

A DNS leak happens when your device sends an unencrypted DNS query outside of your VPN service’s encrypted tunnel to an ISP’s DNS servers. This effectively nullifies the use of a VPN in the first place by allowing third parties to see and potentially track your browsing activity and IP address.

A WebRTC leak is an issue that can occur in modern web browsers that can similarly expose your personal IP address to a website you’re on. While you can avoid this by configuring your browser to not send such data, a VPN should be able to prevent this from happening.

To check for both of these issues, we use publicly available tools such as NordVPN’s IP address lookup tool and ExpressVPN’s leak tests before and after connecting to a VPN and make sure neither fault is occurring.

This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/vpn/how-we-test-vpns-175845189.html?src=rss