Feed aggregator

On March 14, skywatchers in the Americas will be able to see the first of 2025’s two total lunar eclipses.

Michigan's first-term senator, Slotkin — a centrist with deep national security credentials — will deliver the Democrats' rebuttal to President Trump's first speech to a joint session of Congress.

(Image credit: Mandel Ngan)

Rural school districts depend on the state to fund construction and maintenance projects. But over the past 25 years, Alaska lawmakers have ignored hundreds of requests for public schools that primarily serve Indigenous children.

(Image credit: Photo illustration by Shoshana Gordon/ProPublica. Photos by Emily Schwing/KYUK, Michael Grabell/ProPublica)

A Public Citizen report obtained exclusively by NPR shows the new administration has halted or tried to dismiss cases or investigations against at least 89 companies in the last month.

(Image credit: Andrew Harnik)

The country claims Smith & Wesson and other gunmakers are turning a blind eye to hundreds of thousands of high-powered weapons made in the U.S that are illegally trafficked into in the hands of Mexican cartels.

(Image credit: Kevin Dietsch)

Last fall, thousands of people took part in a stress reduction study, learning skills that can improve mood and reduce anxiety. The results are in: Here's what they show.

Celebrated Nigerian writer Chimamanda Ngozi Adichie speaks with NPR's Michel Martin about her new book, "Dream Count."

(Image credit: Ulf Andersen)

The prospect of a federal strategic reserve is a major step in President Trump's vision to establish the U.S. as the crypto capital of the world and could be a game changer for the industry.

(Image credit: Jonathan Raa)

For businesses, data is their one constant. In the age of AI, data itself has not changed, but how it is delivered, and what companies can do with it has. Think of the arrival of Netflix and its effect on Blockbuster. The core product remained the same, but the way it was delivered and consumed was completely new, and that meant Blockbuster ceased to exist.

Data is becoming a key tool for businesses hoping to weather today’s tough economic climate and overcome common industry challenges, from inflation to supply chain issues. Those who have not yet got to grips with their data are more at risk from economic difficulties, as data, underpinned by a robust data strategy and paired with artificial intelligence (AI), can deliver the insight and foresight needed to remain competitive and manage market fluctuations. Companies that fail to engage with their data or those that implement AI ineffectively may face the same fate as Blockbuster.

Building the foundations

Data is as valuable to a company as any other asset on a balance sheet. Organizations cannot be valued by revenue alone, but also by the data they own and the impact it has on the business. All companies have data at their fingertips, but to turn that into business value, it has to be correlated, analyzed and used correctly. Organizations can’t make the most of their data if it is sitting in siloes.

Businesses must have a strong data foundation in place: this is the jumping-off point for data exploration and analysis. Once organizations have onboarded onto a modern data platform to serve as this foundation, it’s time for leaders to think carefully about what sort of AI they want to use. As a first step, they should consider enterprise AI tools, which can be applied to organizations to enhance various business functions. To enable this, businesses need to build a consolidated and open data strategy to run these AI models.

Generative AI (genAI) excels at creating new content based on inputs, often fueled by huge data models. While it has many uses, the vastness of the data required can result in issues within an organization's data foundation if they are unable to ingest and analyze such quantities of data. While there is mass excitement around genAI, companies should avoid rushing into implementation, as the models will not be properly trained on all the data available. This can result in inaccuracies and hallucinations, turning AI investments sour and leading business leaders to cut back on spend.

Enterprise AI should instead be the first port of call for businesses, as it is based on smaller data sets, for a certain purpose or benefit as opposed to having potentially infinite uses. For example, since building an enterprise data foundation, Zoom has built enterprise-grade AI applications for everyone within the company, enabling smarter business decisions as teams have access to the right data at the right time.

Now, anyone within the business can chat with their data in natural language to glean key business insights and increase productivity. Enterprise AI is also more energy efficient, using AI for purpose and leveraging the technology where it makes the most sense. While GenAI will see more mainstream use, the two are not mutually exclusive and businesses will likely need to consider both on their AI journey.

The right culture

Technology, by its very nature, is disruptive and aims to drive automation within an organization. This can have a big impact on people, but business leaders need to minimize any uncertainty around AI as much as possible and work to dispel the natural fear that the technology might herald an era of job losses. The latest research from PwC shows that the opposite will happen: AI will bring an era of gradual job growth, helping to ensure there are enough workers so that organizations can find the people they need. What will change are the skills required by workers.

Traditional skills are already disappearing from job adverts, and new skills are appearing, with this change happening 25% faster in roles where AI automation is having an impact. Businesses need to consider the right people for the job and hire them into roles where they can develop their existing skills while giving them learning opportunities to gain new skills.

Culture is hugely important when it comes to implementing technologies such as AI. Business leaders need to first ensure they have access to education and resources so they can get to grips with the technology themselves. Then and only then should they aim to educate their workforce on the benefits of the technology and the skills they will need to engage with it.

This will help establish an open dialogue for employees and leaders to discuss AI use within a company and its purpose, resulting in an increased appetite for engaging with the technology. Businesses can then assess the right tools, policies and data processes and how they will use AI.

Data and survival

Today, the goal for many businesses is to implement AI and reap its rewards. While AI is the current destination, it’ll become a stepping stone to the next destination that all companies are vying to reach. This is a cycle all industries go through, some organizations survive, and some don’t. To be one of the companies that thrive in the age of AI, leaders need to steer their organizations in the right direction, and ensure there is a strong foundation in place to support a consolidated data strategy, enabling them to differentiate themselves from the competition and deliver for customers.

Doing so requires careful planning, and a shift away from siloed approaches, as well as a step change in company culture. By building this foundation, and avoiding the pitfalls of adopting AI too rapidly, business leaders can build companies fit to grow into an AI-powered future.

We list the best resume builder.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

• Polaroid's Now instant cameras get a refresh with new third-gen models
• Now+ model can connect to Polaroid's app for wireless control
• Available from March 4 in new colors and made from 40% recycled materials

Polaroid has refreshed its Now instant camera series with third-gen Now 3 and Now 3+ models. They are the latest in the line of low-cost instant cameras that shoot onto Polaroid's gorgeous (and largest) square film, like the premium I-2 does. Consider them super-sized Polaroid Go's.

Little changed between the original Now and Now+ from 2020 and second-gen models that came two-years later and, truth be told, these latest models are pretty familiar too.

That's no bad thing, we really rate the Now+ and it has featured in our best instant cameras buying guide for several years. But what exactly is new in the Now 3 and Now 3+?

New features are neatly summarized by Polaroid, who describes its latest Now 3 and Now 3+ as, "a refined instant camera with an upgraded two-lens autofocus system, improved ranging sensor, and better light metering for sharper images, even on the sunniest days."

In other words, upgraders should expect sharper focusing and better-exposed prints, whatever the weather, wherever you are.

What's also new is that both cameras are made with 40% recycled materials – kudos Polaroid for that – and come in a range of new colorways. They are available immediately (from March 4, 2025) at the Polaroid website and select retailers, with prices starting from $119.99 / £119.99 for the Now 3 and $139.99 / £139.99 for the Now 3+ (Australia pricing TBC).

Image 1 of 2

(Image credit: Polaroid)Image 2 of 2

(Image credit: Polaroid) + = ?

There has always been two coexisting Polaroid Now models; the regular Now and the Now +. So what's the difference between the two versions? Well, they are identical in every way save for one thing; the Now+ version can connect to your smart device for wireless control.

The Now 3+ can be controlled through the Polaroid app, with creative tools such as double exposure, remote timer and an aperture priority mode. Combine those skills with colored lens filters (available separately and in various bundles), and there's much creative fun to be had.

It's also possible to scan and share prints through the app. Naturally, the plus model costs a little extra, but not by much and we think it's worth the extra outlay over the standard version.

As cameras, both are excellent value, but when considering an instant camera you need to factor in the running costs of buying film, and Polaroid's premium film is costlier than Instax (and with good reason – it's stunning).

We are yet to get our hands on the new versions, but we'll conduct a full review when we do. Our expectations are tempered – we don't expect to see much difference between the Now+ and Now 3+ but, again, that's no bad thing. We'll let you know soon.

You might also like

• The best instant camera 2025: top retro prints from Polaroid, Instax and more
• I love my new Polaroid, but I wish I didn’t instantly make these 6 common rookie mistakes
• Polaroid Now+ review

This review first appeared in issue 354 of PC Pro.

Jabra’s PanaCast 50 video bar ensures that no-one feels left out, with its triple 13MP 4K camera turret presenting a full 180° horizontal field of view (FoV). What’s more, lurking inside this 650mm wide cylinder is an 8-microphone beamforming array teamed up with pairs of 50mm woofers and 20mm tweeters.

Speaker tracking is a cut above the rest, too, as the PanaCast 50 incorporates no fewer than nine Edge processors, including two Edge AI chips. These allow it to provide integral video analytics, apply always-on people counting and use Jabra’s Virtual Director technology to focus on the active speaker and zero in on them with its automatic 6x digital zoom.

The camera provides a USB-C port for BYOD meetings or connection to a permanent room computer. Jabra also offers two-room system solutions where one partners the camera with its Android-powered touchscreen tablet while another comes with Lenovo’s ThinkSmart Core + Controller devices.

The kit includes a wall-mounting bracket, with the optional aluminum table stand costing £63. We also checked out Jabra’s Bluetooth remote pad (£39), which provides camera and audio controls plus direct access to whiteboard sharing.

We reviewed the model with a grey fabric cover, but Jabra also offers a black version for the same price. Whichever model you choose, you’ll find integral Wi-Fi 5 services plus a 10/100 Ethernet port at the back for remote management using Jabra’s free Xpress web portal.

BYOD installation is easy: you simply connect the camera to a USB port on a Windows or macOS host computer and wait for the drivers to load. It will work happily with any UVC-compliant VC app, but don’t forget to download Jabra’s Direct app otherwise you’ll miss out on a wealth of features.

The app’s camera controller window allows you to play with the image quality, set a zoom level, move the camera view and save two presets that can be accessed from the remote. From the general settings tab, you can control people counting, enable the Virtual Director and turn on the new dynamic composition feature, which puts the four most recent speakers in a split screen.

The PanaCast 50 can be remotely managed and delivers superb video quality (Image credit: Future)

A large monitor is recommended, since the highest resolution is a very wide 3,840 x 1,080 pixels. The PanaCast’s army of Edge CPUs come into play here as they apply Jabra’s patented video-stitching technology to produce a single image from the three camera feeds.

This works perfectly as we couldn’t see any joins and the cameras deliver a pin-sharp image with great color balance and contrast. Speaker tracking is very responsive, too: we could walk around our meeting room and, even without speaking, the camera followed us while we were moving.

Jabra’s microphone expertise shines through. Remote participants could hear us clearly at a five-metre distance, and the quad speaker combo was just as impressive. Jabra won’t beat Biamp’s 2023 Excellence award-winning Parle VBC 2500 as the PanaCast 50 lacks a little in the bass department, but it delivers a clean and clear sound quality with a 65% volume level quite sufficient for our 24m2meeting room.

Remote monitoring and management are good, with the Xpress portal providing analytics on camera and room usage. “Packages” group camera settings together and, when you install the local Jabra Direct app, just copy a package URL to it.

You can remotely apply settings that override the local app. During room creation, you add a device serial number and apply maximum and safety participant capacities. The camera uses its people-counting skills to provide room usage details and will warn you if the room is over capacity.

This sleek cylinder delivers great video and audio quality, fast speaker tracking and a wealth of advanced features. Jabra’s Xpress web portal offers smart remote management services, and the super-wide view helps make the PanaCast 50 ideal for all-inclusive meetings.

We've also ranked the best alternatives to Skype.

This review first appeared in issue 354 of PC Pro.

Fedora Linux is refreshed every six months, with version 39 shipping in November 2023, 20 years (and one day) since the first iteration’s debut. It was originally a spin-off of Red Hat Linux, but the tables have been turned and it now forms the basis of Red Hat Enterprise Linux and CentOS Stream.

Of the five versions on offer, we reviewed the desktop build, which sits alongside server, cloud, containerized and IoT editions. Each release receives support for 13 months, with version 40 scheduled to appear in April 2024, and build 39 reaching end of life in November 2024.

Fedora has a free-to-download media creation tool, much like Microsoft’s equivalent for Windows, and the Raspberry Pi imager for the single-board computers. This writes the latest build to a bootable thumb drive. Use this to start up, and you’ll encounter one of the best installers we’ve come across. It’s simple, straightforward and painless.

System requirements are 4GB of memory and a 40GB SSD, although Fedora Project notes that it’s possible to run the OS on less than this.

Our installation featured a slim selection of pre-installed applications, including Firefox 119, Rhythmbox Music Player, Boxes virtual machine environment and the latest edition of LibreOffice – release 7.6. However, we needed to install our own email client, as well as common creativity tools such as GIMP and Inkscape. You can do this through the integrated software tool, where we found 61 updates waiting to be processed on first launch. This isn’t as drastic as it sounds, as many of them were fonts and codecs.

The Flatpak package manager is enabled and, if you prefer to update manually, Fedora uses the DNF package manager for RPM.

Where some distributions are just now managing the transition from the X11 display server technology to Wayland, Fedora began that process several releases back, and release 39 marks the 14th edition since it completed the transition. Wayland’s frequently touted benefits, not just by Fedora but in general, include greater security and better performance.

The selection of pre-installed apps includes the latest edition of LibreOffice (Image credit: Future)

The default desktop environment is Gnome 45, which isn’t a huge step up from 44. There are some welcome touches, though, such as subtly redesigned window elements, where two-tone colorways and full-height sidebars tidy things up. There’s also an improved workspace indicator in the top left corner. Click it once and you get an overview of your open windows, some of which might otherwise be hidden, alongside a quick way to switch between desktops. It’s useful, but you can achieve the same result by pressing the Super key.

There’s also a new Image Viewer, which Fedora notes has also been rewritten for high performance, while Gnome search has also been reworked with a focus on speed. The improvements don’t only apply in the Files app, but across several core Gnome tools, such as Software and Characters.

If you don’t get on with Gnome, there are several alternative builds – Spins in Fedora parlance – running the lightweight XFCE desktop, KDE Plasma or Cinnamon, among others. Cinnamon, as used by Linux Mint, is often touted by advocates of Windows-to-Linux switching as a reason to choose that distro.

There’s also a handful of immutable deployments, which keep OS code and apps separate, as with Nitrix. By making the core of the OS read-only, it can’t be hijacked by malicious actors or corrupted by a bad or incomplete update. The result is a more secure environment, for use in sensitive workplaces such as finance and government.

Fedora scored 1,105 in our Geekbench single-core test and 3,053 in the multicore section. These figures were broadly similar to what we saw in the Red Hat Enterprise Linux-based Rocky Linux, which, while slightly lagging in the single-core tests, was around 1.8% faster on the multicore tests. In neither case should it make any noticeable difference in day-to-day use.

With one of the best thought-out installers and a wide choice of desktop environments, there’s much to like about Fedora, which is why it’s our runner-up to Ubuntu. The default installation was minimal, but many will appreciate this, and all the tools you could possibly need are waiting in the software manager.

That Fedora is a primary source for Red Hat Enterprise Linux should fill you with confidence, and the fact it’s available with a variety of desktops will make it immediately familiar, whether switching from Windows or a Debian-based rival.

We also the rated the best free office software.

This review first appeared in issue 354 of PC Pro.

Many distributions are available in full-fat and light editions. Zorin and Linux Mint are good examples, each giving a choice of Gnome- or Xfce-based ISOs. Debian with Raspberry Desktop, which closely resembles Raspberry Pi OS, is different. Available for PCs after a decade as the default OS for the eponymous single-board computer, it uses neither Gnome nor Xfce, but PIXEL, the Pi Improved Xwindows Environment, Lightweight, which itself is based on the lightweight LDXE desktop manager.

Don’t let the “lightweight” moniker put you off. PIXEL sports many features familiar from rival environments, including a menu bar, application menu, desktop icons and context menus. It does lack some frills, though. There are no widgets, as there are in Mint, you can’t snap windows to the edges of the screen for easy arrangement, and it lacks Zorin’s visual flair.

However, it does boot extremely quickly, it has every feature you’re likely to need, and its options and settings are as easy to find as they are to understand.

The operating system’s original name – Raspbian – hint at its Debian roots. In fact, you could be forgiven for thinking you were installing stock Debian when working your way through the installer, as there’s no mention of Raspberry Pi to be found.

It’s built on Bullseye (Debian 11) and the Linux 5.10 kernel, both of which are starting to show their age. Debian is currently at 12.2 (Bookworm), on which the regular Raspberry Pi OS for single-board computers is based, and it supports the 6.1 kernel.

Where Raspberry Pi has produced its own imager for setting up Raspberry Pi OS, which also handles user creation, configuring SSH, establishing Wi-Fi credentials and so on, the process for installing Debian with Raspberry Pi Desktop is much like that for most other distributions. Download the ISO, write it to a bootable USB drive, and boot the installer from there. Around ten minutes later, we were sitting in front of the desktop, with a full complement of essential applications ready to be used.

The pre-installed office suite is LibreOffice 7.0.4 (the latest is 7.6), Claws is installed to manage mail, and the default web browser is Chromium. Other applications can be installed via the Add/Remove Software tool, or using APT through the Terminal. We opted for the latter to install Firefox and Thunderbird, and in each case it set up version 115. This is particularly welcome where Thunderbird is concerned, as 115 marked a significant interface refresh, which makes the suite a more pleasant environment in which to spend the working day.

The Add/ Remove Software tool isn’t as friendly as some others (Image credit: Future)

Elsewhere, the list of pre-installed apps is a reminder that Raspberry Pi is popular in STEM environments, coding and automation. Both Geany Programmer’s Editor and the excellent Thonny IDE are in evidence, alongside Scratch and Mu. So is the SmartSim circuit designer.

VLC and an image viewer are both pre-installed, but GIMP (for bitmap graphics), Inkscape (vectors) and Shotwell (photos) all need to be installed manually – if you use them. Again, this can be done through the Add/Remove Software utility, but this isn’t as friendly as the equivalent installers in Ubuntu, Mint and co, presenting sometimes extensive lists of options in response to a search, each accompanied by the same default icon. You might occasionally find yourself scratching your head, wondering which you need.

PIXEL includes a Recommended Software tool, which does allow you to sidestep both the software installer and APT for a handful of common applications, but the three mentioned in the previous paragraph are absent. We’d like to see them included in the next refresh.

While we’ve griped about a few aspects of Debian with Raspberry Pi Desktop, there’s much to recommend it. It’s extremely fast to boot, the PIXEL interface is refreshingly distraction-free, and it’s compact enough to be a reasonable option for running full-time from a USB thumb drive.

Yes, it’s starting to look outdated in places (note that it was released in July 2022), but Raspberry Pi tells us an update is planned, which is expected, although not guaranteed, to arrive before Easter. In the meantime, if you’re looking to recycle some older hardware as a no-frills workhorse, or want to use the same environment on your single-board computer and desktop, this distro could be just what you’re after.

We've ranked the best driver updater.

This review first appeared in issue 354 of PC Pro.

AI seems to be everywhere lately, and that includes the world of home networking: D-Link’s latest “smart router” proudly wears its AI credentials on its sleeve. As we’ll see, this is definitely a case where the buzzword oversells the reality, but if the R32 is short on smarts it makes up for that in value. As we went to press, it cost only £73 inc. VAT on Amazon, making it one of the cheapest Wi-Fi 6 routers around.

The Eagle Pro AI R32 is the successor to last year’s Eagle Pro AI R15, which was similarly cheap and looked almost identical. The main difference between the two is what you might guess from the name: the new model is just over twice as fast as the old one, in terms of wireless bandwidth. It supports connections of up to 800Mbits/sec on the 2.4GHz band, while the 5GHz radio goes up to a maximum speed of 2.4Gbits/sec. There’s also more wired bandwidth available, as the R32 gains an extra Ethernet socket at the back, to make up a full quartet of gigabit LAN ports.

Aside from those sockets there’s not much to see: no USB, no multi-gig connectors and only four LEDs on the front to show you the status of your power, internet and Wi-Fi. But that’s fine – do you really need more?

Four LEDs on the front show power, internet and Wi-Fi statuses (Image credit: Future)

Getting set up is delightfully simple, via either D-Link’s Eagle Pro AI mobile app or the router’s built-in web management interface. There isn’t a huge amount to configure, but D-Link builds in more features than you might expect from such a cheap device. For example, a basic quality-of-service tool lets you assign different priority levels to individual clients, and you can also enforce internet access schedules and time limits for kids’ devices.

For added security, you can flick a switch to replace your ISP’s default DNS with secure DNS from Google or Cloudflare, offering protection from hijacking attacks. There’s a basic configurable firewall, too, and an unexpected treat is an incoming VPN server, plus integrations with no-ip. com and dyndns.com to provide easy external access to your home network.

If you want to use the R32 as a Wi-Fi extender for an existing network you can alternatively switch it into bridge mode – or use mesh mode to connect two or more units together and spread your wireless signal over a wider area. Remember, though, that beaming mesh traffic back and forth between wireless stations eats into the bandwidth available for your devices, so you’re trading off performance for range.

Finally, let’s not forget the R32’s promised “AI” capabilities. In practice, this simply means the router periodically checks its Wi-Fi channels and tunes its beamforming settings to get the best connection to your clients. These are welcome features, to be sure, but it’s a pretty egregious overreach of the term AI.

Never mind; you’re probably not buying a £73 router in the expectation of a world-class feature suite. The real question is, how does the thing perform? And the answer is very well – at close range. I tested the R32 by hooking up a NAS drive to one of its Ethernet ports, then hawking a laptop around my home and measuring upload and download speeds as I copied a set of 100MB test files to and from the NAS.

(Image credit: Future)

Initial findings were very positive. I was delighted to see an average download speed of 85.5MB/sec in the same room as the router, and when I moved downstairs to the living-room I still got a speedy 72.9MB/sec.

Unfortunately, the R32’s four little antennas aren’t beefy enough to keep up that sort of performance all through my home. Speeds fell to 25.8MB/sec in the kitchen and 21.1MB/sec in the bedroom; that’s still ample bandwidth for typical internet tasks – a 4K HDR video stream requires about 4MB/sec – but it’s still a pretty precipitous drop.

Unsurprisingly, the R32 fared worst in the bathroom at the back of the house. Here I could clearly see my file transfers momentarily hang and resume several times during my tests, so while the eventual average download speed of 5.5MB/sec might sound fine for web browsing or cloud syncing, it’s patchy: I’d rather not rely on it for video calls or online gaming.

Does this mean you should steer clear of the R32? It really depends on what you want from a router. Those who need reliable wide-area coverage should consider spending the extra £50 on the Asus RT-AX59U: as well as more stable long-range coverage, it will give you a broader range of software features. However, if you just want basic connectivity across a few rooms, the R32 will fit the bill very nicely – there’s no need to pay more.

We also rated the best encryption software.

A key challenge in conservation and decarbonization today is gaining real-time, verifiable insights into environmental impact from the first-mile. This is the origin point for commodities like agriculture, timber, and minerals— and often where the most significant environmental and ecological risks exist, yet it has traditionally been the most opaque.

AI is transforming this by making massive datasets—satellite imagery, IoT sensors, and environmental risk models—actionable in near real time. With the ability to analyze everything from land-use changes to methane emissions, AI enables companies and regulators to detect deforestation, illegal activities, and sustainability risks with unprecedented accuracy.

Beyond monitoring, AI plays a role in risk forecasting and compliance. It can model climate risks—such as drought, wildfires, or extreme weather—and help organizations proactively adapt their operations and sourcing strategies. As regulations like the EU Deforestation Regulation (EUDR) become more stringent, businesses will increasingly rely on AI-powered analysis to ensure compliance and mitigate supply chain vulnerabilities.

How does AI technology improve the accuracy and reliability of carbon offset measurements compared to traditional methods?

Historically, carbon markets have relied on manual verification and estimates based on projections rather than real-world impact. This has led to concerns about credibility and market integrity.

AI and remote sensing technologies revolutionize this process by enabling near real-time measurement, reporting, and verification (MRV). For example, AI can detect deforestation and land-use changes, ensuring that credits for forest conservation are genuinely additional and permanent. Advanced models can quantify carbon tracking - such as methane emissions and reductions - particularly in agriculture and landfill projects. AI-powered analysis of soil carbon sequestration ensures that carbon farming and regenerative agriculture credits are measurable and defensible.

This milestone-based approach, where credits are issued based on verified progress rather than speculative claims, is helping shift the market toward greater transparency and trust.

What are the biggest challenges in applying AI and machine learning to forest conservation and carbon credit verification, and how can these be addressed?

One of the biggest challenges is data integrity. AI models are only as good as the data they are trained on, and for environmental applications, gaps in first-mile data have historically led to inefficiencies and unverifiable claims.

To address this, there is a growing emphasis on combining multiple data sources – satellite imagery, LiDAR scans, ground-truth observations, and machine learning models – to ensure that carbon sequestration, deforestation, and biodiversity impacts are measured accurately.

Another challenge is the time and cost of project verification. Traditional methods can take years for a carbon credit project to be audited and approved. AI-powered automation is now reducing project registration times from multiple years to just weeks, significantly accelerating climate action.

Regulation is also catching up. Emerging policies are increasingly requiring verifiable, high-resolution environmental data to ensure that credits issued in the market represent real, additional carbon reductions.

How will AI shape the fight against climate change over the next 5-10 years?

AI tools are already proving its value in climate risk management and emissions reduction, and over the next decade, its impact will only expand.

For example, AI will improve carbon tracking, particularly for Scope 3 emissions, which remain the hardest to quantify and manage. It will also optimize nature-based solutions, such as regenerative agriculture and reforestation projects, ensuring they deliver measurable carbon benefits. AI can also enhance climate risk forecasting, helping businesses and governments anticipate disruptions and adapt before crises hit.

We are also seeing a growing intersection of AI and blockchain in sustainability. By integrating AI-powered measurement and monitoring with immutable records on blockchain, businesses can create auditable, tamper-proof sustainability claims – critical for regulatory compliance and investor confidence.

How has technology directly contributed to more effective decarbonization efforts or improved sustainability practices?

Transparency has long been a challenge in sustainability efforts. While satellites and AI can provide visibility into environmental impact, the real issue is verification and accountability.

Historically, registering and verifying a carbon project – whether a reforestation effort or a methane reduction initiative – was an expensive, slow process. But with AI and blockchain-backed registries, project validation timelines have been cut from multiple years to just a few months.

This acceleration is critical because it increases the speed at which capital can flow into high-impact climate projects. Whether it’s reducing emissions from land use, improving soil carbon storage, or transitioning to regenerative supply chains, technology is making it possible to measure, verify, and scale these efforts faster than ever before.

We list the best IT management tool.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

China will impose additional tariffs of up to 15% on imports of U.S. farm products, including chicken, pork, and soy. They follow Trump's order to raise tariffs on imports of Chinese products to 20%.

(Image credit: Andy Wong)

Here are the answers for The New York Times Mini Crossword for March 4.

Interest in AI agents that can peruse the internet on your behalf has ticked up of late, partly thanks to OpenAI's Operator, along with Browser Use and the recently released Proxy 1.0. Popular browser provider Opera has joined in to offer a tool that will actively do tasks on the web for you. If this catches on, the way we interact with the internet might never be the same, especially if it's baked right into the browser like Opera's is.

The idea for the Operator is that, instead of just answering questions or giving you information like Opera's Aria AI assistant, it can go out onto the web and do things on your behalf. You can tell it to buy concert tickets, book a hotel, track down the best deals on a gadget, or research a topic, and it will navigate sites, fill out forms, and complete tasks for you while keeping you updated.

You’re still in control, but now you have a digital errand runner handling the tedious stuff. And that could change a lot about how we use the internet. Here are five ways it could shake things up.

Shop for me

(Image credit: Opera)

Online shopping could become radically simpler with Opera's Operator. Right now, finding the perfect item can require having ten tabs full of reviews, price comparisons, and shipping data. With the Browser Operator, you could just say, “Find me the best-rated wireless earbuds under $150 and order them to my address.”

Instead of doom-scrolling product pages for an hour, you’d get a curated recommendation, approve it, and be done in seconds. Imagine never dealing with sketchy third-party sellers again because your AI assistant already weeded them out for you.

Trip AI

The Operator could make planning travel actually fun instead of stressful. A simple weekend trip requires juggling flight options, hotel rates, rental car bookings, and activity reservations, all while praying you don’t accidentally book a 10-hour layover.

With the Browser Operator, you could say, “Plan a weekend getaway to Chicago with a hotel near downtown and a rental car,” and it would handle the legwork, presenting you with an itinerary to approve. No more agonizing over which travel site has the best deals or whether that budget hotel actually has walls. The AI would be able to do the tedious searching; you just decide what sounds good.

Subscribe AI

Managing subscriptions and online accounts could stop being a nightmare using AI. These days, half of our digital lives are spent trying to remember where we signed up for what, why we’re still being charged for something we don’t use, and how to cancel a subscription before it renews for another year. Normally, you have to dig through emails, track down obscure account settings, and fight a desperate battle with “Are you sure you want to cancel?” pop-ups.

With the Browser Operator, you could say, “Find all my active subscriptions and show me what I should cancel.” It could even handle the cancellations for you, sparing you from guilt-tripping retention prompts. Suddenly, your bank account isn’t a graveyard of forgotten free trials turned full-priced commitments.

Bills to pay

(Image credit: Opera)

Even the most mundane online tasks could become hands-free with the Operator's help. Paying bills, managing subscriptions, downloading bank statements are all the little things that chip away at your day. The annoyance of having to remember which day you need to log in, navigate a website, and remember your passwords is just part of life.

However, you can set up the Browser Operator to handle routine tasks automatically. Imagine just getting a notification that your phone bill has been paid instead of remembering to do it yourself. That’s not just convenience but fewer minor annoyances cluttering your brain.

Information filter

Keeping up with the internet’s relentless firehose of content might actually become manageable using the AI Operator. Staying informed today means subscribing to newsletters, following a bunch of blogs, and hoping the algorithm decides to show you nothing but cat videos. But the Browser Operator could act as your own personal news curator.

You could tell it, “Keep me updated on the latest breakthroughs in space exploration,” and it would regularly collect and summarize the most relevant articles. Instead of wading through an endless news feed, you’d get just what matters to you, neatly packaged. That way, you can stay informed without feeling like the internet’s infinite scroll has hijacked your entire day.

Operator opening

The internet has always required us to be the operators; clicking, searching, navigating, managing. But with AI tools like this, that might be changing. Opera’s Browser Operator takes the first real step toward making the browser an active participant instead of a passive tool. It doesn’t just give you a new way to browse; it changes what browsing is. Sure, this could make everything more efficient, but it also raises questions about what happens when we offload so much of our online activity to AI. If the internet can browse itself for us, how much do we really need to engage with it? Will we still know how to search for things manually in a few years, or will that start to feel as outdated as dialing a rotary phone?

For now, though, it’s hard not to be excited. This is the kind of innovation that makes you wonder how we ever lived without it. If AI can start handling the tedious parts of the internet, maybe we’ll finally have time for the things we actually enjoy. Or, more realistically, maybe we’ll just use that extra time to doom-scroll even more efficiently. Either way, the future of browsing just got a lot more interesting.

You might also like

• OpenAI's first AI Agent is here, and Operator can make a dinner reservation and complete other tasks on the web for you
• I discovered a surprising difference between DeepSeek and ChatGPT search capabilities
• Opera brings its web browser and AI assistant to iOS

A spring storm is bringing various weather conditions across the U.S. Some regions are facing snow, possible tornadoes, and critical fire conditions, according to the National Weather Service (NWS).

(Image credit: NWS/NOAA)

The news comes just days after a disastrous meeting between President Trump and Ukrainian President Volodymyr Zelenskyy.

(Image credit: Andrew Harnik)