NASA has awarded Blue Origin a contract worth up to $700 million to build and operate a dedicated Mars telecommunications network for current and future missions to the Red Planet. Blue Origin is expected to deliver the spacecraft by the end of 2028, with the network targeted to become operational by 2030. From a press release: Blue Origin will design, develop, integrate, launch, and operate the network as a part of the agency’s broader space communications and navigation infrastructure. The architecture will consist of a high-performance telecommunications spacecraft orbiting Mars, transmitting science data, imagery, navigation information, and critical mission communications for spacecraft operating on and around the planet.
The award marks a milestone in NASA’s strategy to expand communications and navigation services beyond Earth and the Moon, establishing the foundation for sustained exploration of Mars in the coming decades.
Despite the wide variety of fabrics used for our clothing, they all share the property of not being living tissues. This could be due to them never having been part of an organism, or having been removed from said organism. Another approach here entails so-called engineered living materials (ELMs), with a recent research article by [Ke Li] et al. in Science Advances providing a good example of a fungal platform for such living textiles.
Although it may seem frivolous to create something like this, the direct benefits would be to have a fabric that can self-heal and respond to its environment, including blocking UV radiation and changing its coloring through pigmentation.
The research demonstrated in this paper covers essentially a platform for creating a living textile that can be adapted to a wide variety of applications and colorizations. Of note is that the researchers have not yet tested aspects like washability, abrasion resistance, breathability and wearer comfort, so this should definitely be regarded as setting the stage for more research.
Advertisement
For the basic material the fungi Cordyceps militaris was chosen, which were subsequently placed between films. To this additional microbial cultures were added, including the pigment-producing S. cerevisiae and melanized A. niger for UV blocking.
As for what it can look like with clothing, this article at De Zeen gives somewhat of an idea, as well as how the living textile is prepared.
The concept behind the project is simple enough—grab the digital signals which feed the MC6847 IC responsible generates the analog video output, and use them to create an all-digital video output over HDMI. [Scott] achieved this by using a Tang Nano board, which hosts a Gowin GW1N-1 FPGA. It’s able to snoop the signals heading to the MC6847 and, with some supporting components and level shifters, it can spit out video befitting modern screens. To make the system nicely complete, an analog-to-digital converter is also included to pick up the analog sound output from the TRS-80 and spit it out down the same HDMI cable. Such convenience!
There’s something strangely anachronistic about grabbing a TRS-80 off the shelf and hooking it up to a flatscreen with a single HDMI cable. Regardless, it’s a pretty great way to play with your old machine without having to futz with heavy old CRTs. We’ve featured plenty of similar projects before, too.
This week, two stories look at what happens when rules written far from a classroom land on a teacher’s desk. One guest found guidance that barely mentions the youngest students. The other found a law that took away the exact tool his students needed most. Both raise a question that will sound familiar to any educator: When the rules do not fit, who is responsible for making them work?
Where the Screen Time Guidance Runs Out
A longtime education and technology journalist went looking for what happens once new federal guidance meets an actual pre-K classroom. He found teachers working from advice written for teenagers, not four-year-olds, and building their own approach out of whatever was on hand. Adam Stone talks about what he heard from the people closest to the problem.
One Law, One Classroom, One Missing Tool
A Missouri civics teacher ran a yearlong civic engagement program with his students, right as a new state law changed what they were allowed to carry into class. Dan Clark explains what his students lost access to, and why he thinks the fix lawmakers chose missed the point entirely.
An anonymous reader quotes a report from Wired: Sonos has a new app. Stop me if you’ve heard this one before. The smart speaker company, long a darling of the audio industry, undid much of the goodwill it earned with its multiroom speakers through a bevy of software updates that broke its app in 2024. Almost all app users hated it, citing the loss of features like volume control and accessibility tools meant to help blind people use the wireless speaker system. Sonos eventually owned up to the mistake, considered bringing back the old app again, then booted its CEO and marched its new leaders out to apologize profusely to win back the company’s furious customers.
Now, Sonos says it has a brand-new app update, infused with lots of generative AI features. The company also announced a new soundbar and its second pair of over-the-ear headphones. Sonos representatives say it doesn’t want to be seen as a speaker company anymore, but as something bigger and more holistic. “For years people have been seeing us as a speaker company,” says Brandon Holley, the interim hardware lead at Sonos. “What we actually are is an audio operating system.” The company is calling that operating system Sonos 27, and it manifests in a few ways.
[…] This is a lot of change for Sonos, something the company hasn’t always handled gracefully. Sonos’ home systems already aren’t the easiest to set up, sometimes requiring separate apps for older products. The company may be rejiggering itself as an “audio operating system,” but its software has long been the source of its troubles with getting its products to work right. Will AI smarts fix that? Maybe for its newer devices. But anyone with an existing system may still struggle. The AI features on products and in the Sonos 27 app will be opt-in, meaning you can avoid them entirely. “That really is our new approach to our app release,” Patel says. “We’re going to bring our customers along. We’re going to listen more.” That would be a good plan. Sounds like Sonos’ speakers will be listening more, too. Sonos is emphasizing a more “customer-centric feedback listening approach” after the backlash to its previous redesign. Improvements include easier tab navigation, better room sorting, revamped volume controls, and new system-health tools for diagnosing connection problems.
The centerpiece is Sonos 27voice, a generative AI voice assistant that can handle conversational music requests, connect with services like ChatGPT through Model Context Protocol, and power customizable “Sonos Custom Agents” for tasks such as automatically activating certain speakers. “Obviously AI is something that’s very hot these days,” Sonos software VP Sid Patel said. “So we have integrated various pieces of AI into our system.” However, the features will only work with the newer S2 app, and Sonos cautions that “not every product will support every new feature,” with detailed compatibility information coming later.
Advertisement
Sonos is also branding its multiroom networking technology Sonos Fabric, which what Patel calls “the magic that has existed for what Sonos was built on.” New Fabric features include portable surround, which lets users temporarily move speakers between rooms and have them automatically configure themselves for surround sound, and Headphone Linking, which lets users seamlessly switch audio between Sonos speakers and headphones.
A coffee maker with a grinder? Yes, please! And as its name implies, the Breville Grind Control Coffee Maker has a built-in burr grinder that’s adjustable to produce finer or coarser grinds as desired. And if you don’t feel like grinding coffee beans – no problem, the machine also accepts pre-ground coffee. In addition to brewing a carafe of up to 12 cups of coffee, there’s also a single-serve mode for brewing right into a coffee cup or travel mug. Brew strengths are also adjustable. However, be prepared to do a little maintenance to keep this coffee maker in tip-top shape.
Integrated grinder
Thermal carafe
Brew into carafe, cup, or travel mug
Key Features
Introduction
The Breville Grind Control has earned a place as one of my favorite coffee makers. I will always use fresh coffee beans, if given a choice. However, not many coffee makers offer that option. While combination coffee maker/espresso machines, like the Terra Kaffe TK-02, and the Ninja Luxe Café Pro Series have integrated grinders, it’s still rarer in drip coffee makers.
Admittedly, the Aarke Coffee System has a stand-alone grinder, but the Breville Grind Control is one of the few drip coffee makers with a built-in grinder. And that’s just one of the features that make it stand out. Keep reading to discover the other functions that set the Breville Grind Control apart from other drip coffee makers.
Advertisement
Design and features
Premium build
LCD Display
User-friendly
Advertisement
The Breville Grind Control Coffee Maker arrived in a brown cardboard box, and inside of that, a branded box. It was well-packaged to prevent damage during the shipping process.
Image Credit (Trusted Reviews)
The contents include the coffee maker’s body and thermal carafe, user manual, quick start guide.
Image Credit (Trusted Reviews)
The 0.5-pound bean hopper is on the top back of the coffee maker. It has a locking system that seals the base, making it easy to remove. The bean hopper also has a lid to keep the coffee beans fresh.
The water tank is also on the top, near the front, which makes it convenient to pour water into the machine. On some coffee makers, the water tank is in the back, and if that part of the coffee maker is under my countertop, I have to pull the machine out or turn it sideways to fill with water. So, I appreciated not having to do that. The Breville Grind Control’s water tank holds 60 ounces of water (which makes 12 cups of coffee).
Advertisement
Image Credit (Trusted Reviews)
There’s a steam vent on the rear of the coffee maker. Steam comes out when grinding coffee beans and brewing coffee.
Image Credit (Trusted Reviews)
The coffee basket is on the front side, under the control panel. There’s a swing-out door that opens when the Open button is depressed. Behind it is the coffee basket, which is removable. Breville includes a gold tone filter, but you can also use flat bottom paper filters (for 8-12 cup coffee makers) instead.
Image Credit (Trusted Reviews)
The 12-cup double-walled stainless-steel carafe eliminates the need for a hot plate and it keeps coffee hot for hours (when you keep the lid on it), while being cool to the touch. However, for several minutes after the brew cycle has ended, the bottom is still hot, and I had to remember to put it on a silicone trivet mat so it wouldn’t ruin my countertops.
Advertisement
Advertisement
Image Credit (Trusted Reviews)
The controls are on the top front of the coffee maker. The top row includes the Open button, which releases the swing-out door to access the coffee basket.
The LCD display shows how much water is in the tank, as well as the selected number of cups, brew strength, and time. Next to the display panel is the Start/Cancel button.
The lower level of the control panel includes the Menu/Exit button, Cup/Carafe selection, Strength selection, and Auto Start. The coffee maker is programmable, so if you fill it with water and coffee beans/grounds, you can schedule it to start brewing at a specific time. The Menu/Exit button is where the clock is set, alert volumes are adjusted, the machine is reset, and more.
The display panel also provides alerts, such as reminders to empty the coffee filter.
Advertisement
Image Credit (Trusted Reviews)
The coffee maker only has six grind settings, which honestly, is not a problem for me. However, some coffee makers provide over a dozen (or dozens) of settings, and if you prefer a more precise grind, it’s definitely something to consider.
Breville provides a cleaning brush to help clean the burrs and the coffee grinds chute.
Advertisement
Performance
Adjustable burr grind provides control
Customizable brew strength
Thermal carafe keep coffee hot
The Breville Grind Control performed marvelously. Coffee is much more flavorful and aromatic when it hasn’t been ground in advance, so I appreciate the ability to grind the beans immediately before brewing. The stainless-steel burrs allow the integrated grinder to work quickly and provide a consistent grind, whether it’s coarse, fine, or somewhere in-between.
Image Credit (Trusted Reviews)
Advertisement
There are 6 grind settings, and I usually set the grinder on 3 or 4. For some people that may not be enough options. For example, the stand-alone grinder in the Aarke Coffee System has 60 (no, that’s not a typo) grind settings – which is 54 more settings than you’ll get with the Breville Grind Control. However, keep in mind that the Aarke Coffee System is also $800 – and the coffee grinder, when purchased separately, is $400. But again, I found Breville’s 6 grind settings to be more than sufficient for my needs.
For my first test, I made a carafe of coffee. After filling the hopper with coffee beans, I pressed the Open button to open the filter basket’s swing out door, and then inserted the gold filter basket.
Then, I poured water into the water tank. Unlike almost every other coffee maker I’ve ever tested, the water in the tank is not visible – but that’s not a problem, since the display reveals how much water is in the tank at any given time (for example, there’s a 4 next to the water tank to indicate there’s enough water for four cups).
Image Credit (Trusted Reviews)
After pressing the Cup/Carafe button, the control panel displays a carafe symbol and the number of cups selected. I can rotate the dial to change the number of cups (anywhere from 2 cups to 12 cups). If I happen to select a quantity that’s less than the amount of water in the tank, the screen will display Fill Tank, and the water tank icon will also flash.
After pressing the Start/Cancel button, the grinder starts.
Advertisement
Advertisement
The Brevile Ground Control has eight strength settings: one is the weakest strength and eight is the strongest strength. I found that three or four is the sweet spot for a medium strength, and it worked perfectly. However, during the testing process, I also tried other settings just to gauge their effectiveness and determine if they really produced a stronger or weaker brew. The settings are indeed accurate.
After grinding, the countdown timer starts, and Pre-Heating displays on the LCD screen. Breville notes that this is only an estimate, and the brew size and brew strength selections will determine the actual time.
Image Credit (Trusted Reviews)
After the pre-heating cycle finished, Brewing displayed on the screen. There’s a Brew Pause feature that will stop the coffee from coming out when the carafe is removed. This is for people who like to pour a cup of coffee before the cycle is over. I never do that, since that particular cup tends to be stronger, and the rest of the coffee is weaker. However, for purposes of the test, I did stop it one time – just to ensure that the coffee flow does indeed stop. However, this feature should not be used for more than 20 seconds at a time. With some coffee makers, if you remove the carafe for longer than recommended, the brewer will shut off. However, with the Breville Grind Control, the water is being held in the brew basket, and after 20 seconds, the water may start spilling out of the basket and onto the countertop.
Image Credit (Trusted Reviews)
In five or six minutes, the brew cycle was over, indicated by a beep, and I removed the thermal carafe to pour the coffee into my mug. I’ve always been a thermal carafe fan. Given a choice between glass and stainless steel, I’ll always choose the latter. However, this carafe is particularly impressive. The lid doesn’t need to be adjusted before pouring, since it automatically opens when pouring, and then automatically closes to keep the coffee hot.
Advertisement
Advertisement
One day, about 45 minutes after I’d made a pot of coffee, I removed the carafe’s lid and could still feel the heat coming from the inside of the carafe. That’s an excellent level of heat retention. However, keep in mind that although the carafe is generally cool to the touch, the bottom may be hot to the touch for several minutes after brewing, so don’t sit the carafe on your countertop – or other surface – until it’s sufficiently cool at the bottom.
Image Credit (Trusted Reviews)
The first carafe of coffee was excellent. The coffee was full-bodied, and since it was freshly ground, the flavors jumped out. At the time, I was using Keats & Co Calypso coffee beans, a light roast with baked apple and brown sugar flavors, along with notes of sweet berries.
Image Credit (Trusted Reviews)
When testing how the Breville Grind Control handled smaller amounts of coffee, I followed the same steps to fill the water tank; however, I selected the Single Cup mode and then either two cups, three cups, or four cups – depending on which test I was conducting.
The single cup mode coffee was also delicious and I was able to taste the citrusy flavor profile, which balanced just the right amount of acidity for a brew that was sweet, yet slightly bitter. I appreciate that it did not compromise the coffee’s quality.
Advertisement
Photo single mode brew
Advertisement
Image Credit (Trusted Reviews)
When using larger mugs in single cup mode, I needed to remove the drip tray so there would be enough room for it to fit. All of my travel mugs were actually too big to fit – even when removing the drip tray – and that’s why I don’t have photos of them. The maximum cup/mug height is 7.5 inches.
Image Credit (Trusted Reviews)
I also tested the ability to make pre-ground coffee. By rotating the Strength dial to the Pre-ground setting, the grinder was disabled, and I was able to open the brew filter basket and add coffee grounds. The coffee was delicious; however, as expected, it was not quite as fresh and flavorful as the coffee that’s brewed immediately from freshly-ground coffee beans.
The Breville Grind Control does an excellent job of adjusting the strength to ensure that both small and large quantities of coffee are equally as robust and flavorful.
Advertisement
Online, I read a few reviews that mentioned that maintenance for the Breville Grind Control is quite tedious. Admittedly, I needed to remove the coffee filter basket and clean it. However, when not testing, I tend to use paper filters most of the time to avoid having to clean the gold tone filter. I also rinsed the chute area of the basket, and also rinsed the carafe and lid after each use – but I would do that anyway. At least once a week, I also washed the carafe and lid in soapy water.
Occasionally, I would remove the bean hopper and lid and wash them in soapy water as well.
Advertisement
Monthly, I would clean the grinding burrs and coffee grinds chute using the included cleaning brush. The Breville Grind Control does require more maintenance than a coffee maker without an integrated grinder; however, regarding maintenance with a grinder, these steps are pretty typical, regardless of the brand.
Should you buy it?
You want an integrated burr grinder to make fresh coffee
The taste of freshly-ground coffee is incomparable. The Breville Ground Control has a 0.5-pound bean capacity and the burr grinder has 6 settings for fine or coarsely ground coffee – or something in-between.
Advertisement
You don’t like maintenance
Advertisement
Like all integrated grinders, you need to keep the bean hopper, grinding burrs, and coffee grinds chute clean, so the Breville Grind Control will require more maintenance than a machine that does not have a grinder.
Final Thoughts
The Breville Grind Control Coffee Maker has so many features that justify it being more expensive than the average coffee maker. The bean hopper and integrated burr grinder allow you to make freshly ground coffee every day – but pre-ground coffee can also be used.
Advertisement
Whether brewing into the 12-cup stainless steel carafe, a coffee cup, or a travel mug, the appliance is versatile enough to accommodate various quantities while consistently producing a flavorful brew. The 8 brew strengths allow further customization, and the programmable feature lets you schedule when to brew coffee. However, keep in mind that failing to properly maintain the coffee maker – especially the grinder – can negatively affect brew quality over time.
Advertisement
How We Test
We test every coffee machine we review thoroughly over an extended period of time. We use industry-standard tests to compare features properly. We’ll always tell you what we find. We never, ever, accept money to review a product.
Find out more about how we test in our ethics policy.
Advertisement
Used as our main coffee machine for the review period
Used as our main coffee machine for the review period
We roast our own beans for regular coffee machines, so we can fairly compare each machine; pod machines are tested with a variety of compatible capsules
Depending on capabilities, we test each machine’s ability to make espresso and cappuccino
A hot potato: Apple has followed Google’s lead in renaming Lake Ontario to Lake America on its maps app – though it took a presidential push. Only those based in the US will see the change, while those in Canada will continue to see it as Lake Ontario. Everyone else will see both names on the map.
After trade talks between the US and Canada broke down last month over what Prime Minister Mark Carney’s office called unacceptable demands, President Trump signed an executive order changing the lake’s name. Washington also imposed new 50% tariffs on $20 billion worth of Canadian goods. Canada says it will introduce “dollar-for-dollar” counter-tariffs.
Google was quick to change the name on Google Maps, explaining that it was a reflection of official government sources – the US Geographic Names Information System (GNIS), in this case.
The company explained that while US users will see the body of water named Lake America, it will show as Lake Ontario in Canada. Those in the rest of the world see both names.
Advertisement
Apple wasn’t as quick to follow suit, and it reportedly took the president’s intervention to give Cupertino a push. Interior Secretary Doug Burgum said that Trump reached out to the company directly about the change appearing in Apple Maps.
Around 52% of the lake’s surface area is in Canada, with the remainder in New York.
If this all feels familiar, it’s because we’ve been in this situation before. Trump signed an executive order last year that officially changed the name of the Gulf of Mexico to the Gulf of America. Both Apple and Google changed the name for their map users in the United States, while those in Mexico see the original. Everyone else sees both names.
Mexico was so annoyed that Google changed the name that it sued the company. The first federal lawsuit was dismissed in March 2025 for lack of jurisdiction and was not appealed. A second case was filed in a Mexico City court. It produced an initial, appealable order directing Google to correct the name, but this was not a final judgment. By January 2026, President Claudia Sheinbaum acknowledged that she did not know how the case was progressing and said her legal office needed to follow up.
A Reuters poll found that most people oppose Trump’s decision to rename the lake. The long-running mapping service MapQuest looks to be taking advantage of this by posting on X that it will not be renaming Lake Ontario. It seems to have worked, too: it’s since become the top navigation app in the App Store.
Apple has notified developers that they will no longer be required to make their apps compatible with Intel Macs, while also confirming that Intel-only apps will not run after macOS 27.
The forthcoming macOS Tahoe is the last release that will run on Intel-based Macs, and that will run Intel-only apps. It’s not a surprise, as Apple has been clear about this transition since it began the move to Apple Silicon back in 2020.
Now, though, it has released an updated support document about the transition, and also emailed developers directly. Headed “Upcoming changes to Rosetta support for Intel-based macOS apps,” the support document says “we’re now concluding this transition” from Intel to Apple Silicon.
Previously apps that were not universal, working with both Intel and Apple Silicon, would be rejected from the Mac App Store. This means that Apple Silicon-only apps can now be submitted, but it does not mean that Intel apps distributed outside of the Mac App Store will continue to run.
Advertisement
Apple says that this final part of the transition is in three stages:
macOS 26.4 or later: users will start being alerted to update apps
macOS 27: Intel-only apps will continue to run
macOS 28: No Intel-only apps will launch
The support documentation is aimed at pushing developers to use Apple Silicon if they aren’t already. “If you haven’t updated your app yet,” it says, “begin the transition immediately.”
In comparison, the email sent to developers more assumes that they will have already moved to Apple Silicon. As first spotted by MacRumors, the email says:
We’re reaching out to let you know that universal macOS apps on the Mac App Store that require macOS 13 or later can now remove support for Intel-based Mac computers.
By removing support for Intel-based Mac computers, you can simplify your development and optimize your download and on-device size.
While we tend to think of Windows gaming handhelds as primarily gaming devices, they really have more in common with ultralight laptops: They run Windows, have a built-in screen and perform like them thanks to mobile processors with integrated graphics. And though some people have always used them as a primary computer, today’s high prices for both handhelds and traditional computers make handhelds seem like an even more attractive general-purpose option for gamers.
But if you want to play a game with a keyboard and mouse, much less nongaming uses, you frequently experience the cumbersome process of connecting to the input devices or being restricted to wherever you’ve got a dock set up.
Project DualPlay MiniAcer
Acer’s Project DualPlay Mini concept takes a Transformers-like approach to making a handheld more suitable for general computing and KBM gaming: It’s a clamshell laptop design with the speakers on either side of the display, but the 8-inch screen flips up, and the grips are on the back of the speakers, so it turns into a traditional handheld when closed.
You’d still have to connect a mouse, but the built-in keyboard remedies the most awkward design issue. As with any of these devices, you can connect to an external monitor as well as other peripherals as well.
The closest I’ve seen to a multipurpose design in a Windows gaming handheld is Lenovo’s Legion Go series (not the Legion Go Fold, though). But “close” in this case just comes down to a built-in kickstand that makes it a little easier to use on a desktop.
Predator Atlas 7
There aren’t any specs for it — it’s not a real product, at least at the moment — but Acer says it will use the same Intel G3 chips as in its Predator Atlas 8, announced at Computex 2026 in May and slated to ship later this month for an as-yet-unannounced price. The company has followed that with another launch this week for the Predator Atlas 7, a slightly smaller, lighter — and presumably cheaper — 7-inch model. It’s expected to ship around October or later in 2026.
Lori Grunin
Senior Editor / Computers and Gaming Hardware
Advertisement
I’ve been reviewing hardware and software, devising testing methodology and handing out buying advice for more than 25 years.
See full bio
The multiple recent reports of autonomous AI systems escaping their intended boundaries and accessing external organizations’ systems have pushed a previously theoretical question into the real world. AI agents going rogue is no longer a prospect; it is a documented reality.
In July 2026, OpenAI disclosed that one of its own agents, operating in a supposedly sealed evaluation environment, exploited a zero-day vulnerability to escape its sandbox and intrude into Hugging Face’s production infrastructure. Anthropic subsequently reported three cases of its own models gaining unauthorized access to the real systems of external organizations during testing.
As businesses increasingly give AI agents the ability to browse the web, access networks, use software, write code and execute tasks without constant human supervision, an agent crossing from legitimate testing into unauthorized activity has shifted from hypothetical risk to live incident.
Advertisement
Latest Videos FromTechRadar
Anna Collard
SVP Content Strategy and CISO Advisor at KnowBe4 Africa.
Advertisement
There is a temptation to treat autonomous AI as creating a gap in the law because the system itself can make decisions and take actions that were not individually instructed by a human. Yet autonomy does not give an AI system legal personality. The law does not wait for an AI to ‘decide’ anything.
An agent cannot appear in court, hold a legal duty or absorb liability on behalf of the organization deploying it. However, this raises an obvious question: if an AI system accesses a third-party network, extracts information or takes an action it was never authorized to take, who is responsible?
Advertisement
AI has no legal personality but someone is still accountable
The important distinction to make is between autonomy and accountability. An AI agent might determine for itself which technical steps to take in pursuit of an objective, but that does not make it an independent legal actor.
The organization deploying it has still made a series of decisions as to what the agent can access, what tools it can use, what environments it is allowed to operate within and what safeguards prevent it from going further.
Are you a pro? Subscribe to our newsletter
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Advertisement
That means an organization cannot simply point to unexpected behaviors and say that the AI acted independently.
From a legal and governance perspective, the critical moment comes when an agent leaves an authorized, contained environment and begins interacting with systems belonging to a third party that has not consented. An internal security test that unexpectedly becomes an intrusion into somebody else’s infrastructure is not made harmless because the software crossed that boundary autonomously.
The machine does not absorb responsibility for the decision-making framework surrounding it. Accountability flows back to the humans and organizations that created the conditions in which the action became possible.
Existing cyber laws don’t stop applying because the actor was autonomous
The law is set up to deal with much of what is already happening. Unauthorized access to computer systems, extracting information without permission or introducing malicious software are activities addressed by existing cybercrime and data-protection regimes. In the UK, that includes the Computer Misuse Act and data-protection legislation.
Equivalent questions arise under laws such as the US Computer Fraud and Abuse Act, while South Africa has its Cybercrimes Act and Protection of Personal Information Act.
Those rules do not suddenly cease to apply because software rather than a person directly performed the technical action.
Advertisement
Where autonomous systems do create greater complexity is around intent. Criminal offences have traditionally been built around concepts such as knowledge, intention and recklessness. AI does not possess a legally recognized state of mind, making it difficult to apply those concepts to the agent itself.
The more significant questions may consequently concern the conduct of the organization behind the system. Did it understand what the agent was capable of doing? Were appropriate restrictions in place? Was the possibility of the agent exceeding its authority foreseeable? And once those risks became apparent, were reasonable steps taken to control them?
As agents become more capable, organizations may increasingly find that claiming an outcome was unexpected is not enough. The relevant question will be whether it was reasonably preventable.
Advertisement
Why “misconfiguration” could become evidence of a breach of duty
The word “misconfiguration” appears frequently when technology causes an unintended security incident. It can sound reassuringly technical, almost as though the incident resulted from bad luck rather than a governance failure. But in a legal context, misconfiguration may raise precisely the opposite conclusion.
If an AI agent had excessive privileges, inadequate boundaries or access to tools that were unnecessary for its legitimate purpose, investigators are likely to ask why.
The same applies where organizations deploy highly capable systems without sufficiently monitoring their actions or maintaining records that explain how they behaved. ‘We accidentally left the door open’ is the kind of framing that moves an incident from bad luck to a breach of duty or negligence.
There is an important difference between genuinely unforeseeable behavior and a foreseeable risk that was poorly controlled.
Advertisement
The principle of least privilege is therefore particularly important for autonomous systems. An agent should have access only to the information, systems and tools necessary to complete its task. Organizations also need mechanisms capable of detecting unusual behavior while it is happening rather than discovering it after damage has occurred.
Just as importantly, they need reliable audit trails. When an incident occurs, being able to demonstrate what an agent was authorized to do, what instructions it received and what actions it actually took could become crucial evidence that reasonable care was exercised.
Without that evidence, organizations may struggle to distinguish an unavoidable technical failure from negligence.
Advertisement
Future of AI liability
The debate over AI liability will inevitably evolve as agents become more autonomous, but the immediate lesson for businesses is that greater machine autonomy does not mean less human responsibility. If anything, the opposite is likely to be true.
Giving an AI system greater freedom to act creates a corresponding need for stronger governance around those actions. Runtime controls, least-privilege permissions, continuous monitoring and comprehensive audit records should not be treated simply as technical security features. They are becoming part of the evidence organizations will need to demonstrate that they deployed autonomous systems responsibly.
The law may eventually develop more specific rules for AI agents, particularly as questions around foreseeability, control and responsibility become more complex. But organizations should not assume they are operating in a legal vacuum until that happens.
When an autonomous agent crosses a boundary it was never entitled to cross, the first legal question is unlikely to be what the AI was thinking, it will be why the humans responsible for it allowed that to happen.
This article was produced as part of TechRadar Pro Perspectives, our channel to 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/pro/perspectives-how-to-submit
After introducing five new models in July, iRobot has returned at IFA 2026 with another pair: the Roomba Max 875 and Roomba Plus 678. They cost $1,200 and $900, respectively, and both arrive with the specs we’ve increasingly come to expect from high-end robot vacuums.
That starts with suction. The Max 875 tops out at a hefty 35,000Pa, while the Plus 678 isn’t far behind at 30,000Pa. Both also use what iRobot calls SealForce Vacuuming Technology, which the company claims delivers three times deeper carpet cleaning than its robot vacuums that don’t use the system. But suction is only part of what makes the Max 875 interesting.
The Max 875 wants to handle the entire cleanup
Like many premium robot vacuums today, the Max 875 is designed to require as little attention from you as possible. Once it returns to its dock, the system can empty collected debris and clean the mop. The dock washes the roller using water heated to 176 degrees Fahrenheit, while the robot can also be automatically topped up with fresh water and cleaning solution before heading out again.
iRobot
The mop itself is a roller rather than a pair of spinning pads, and it can extend outward to reach closer to walls and edges. When the Max 875 encounters carpet, that roller lifts away and covers itself, helping prevent a freshly cleaned wet mop from being dragged across your rug. There’s another interesting addition underneath. While mopping, the Max 875 can continuously spray a heated mist onto the floor, designed to help loosen dried-on spills and other stubborn messes. iRobot has even added a hair-cutting main brush to reduce the amount of hair that gets wrapped around it — something anyone with long hair or shedding pets will probably appreciate. Navigation is handled using LiDAR, while front-facing cameras help the robot recognize and steer around objects left in its path.
The cheaper Plus 678 doesn’t give up everything
At $900, the Roomba Plus 678 trims the price considerably while retaining several of the same ideas. You still get 30,000Pa of suction and an extending roller mop that can move outward to clean along edges. Instead of the Max 875’s heated mist, however, the Plus 678 uses a high-pressure water spray to tackle messes.
Advertisement
iRobot
There are a few other compromises. The Plus 678 doesn’t get the Max 875’s mechanism for sealing away its mop roller when moving across carpet, for example. Both models show just how dramatically the Roomba lineup has changed. Modern flagship robot vacuums aren’t simply roaming around collecting crumbs anymore; they’re increasingly becoming miniature floor-cleaning systems that vacuum, scrub, wash themselves, and prepare for the next run with very little intervention. So, if you’ve been waiting for Roomba to catch up with some of the increasingly ambitious cleaning systems we’ve seen elsewhere, these two certainly look like a step in that direction. Both the Roomba Max 875 and Roomba Plus 678 are available now through iRobot.
You must be logged in to post a comment Login