A new study of roughly 80,000 bachelor’s degree recipients from a large urban public college system finds that characteristics of a graduate’s first job can explain nearly two-thirds of the otherwise-unexplained earnings gap between students from low-income and high-income families five years after graduation.
The research [PDF], published as an NBER working paper by economists at Columbia University, tracked graduates from 2010 to 2017 using administrative education data linked to state unemployment insurance records. Low-income students — defined as those receiving Pell grants throughout their undergraduate enrollment — earned about 12% less than their high-income peers at the five-year mark. A substantial gap of roughly $4,900 persisted even after the researchers controlled for GPA, college attended, major, and other pre-graduation characteristics. That residual gap fell to about $1,700 once first-job variables entered the equation.
Graduates from lower-income families tended to start at employers paying lower average wages and were less likely to have their first job secured before graduation. Just 34% of low-income graduates continued at a pre-graduation employer compared to 40% of their higher-income peers. The firms employing low-income graduates paid average wages that were 18% lower than those employing high-income graduates. The researchers say that while the study cannot establish causation, the patterns suggest that supporting low-income students during their transition from college to the labor market may be a fruitful area for policy intervention.
Read more of this story at Slashdot.
Back in April, District Court Judge Yvonne Gonzalez Rogers delivered a scathing judgment finding that Apple was in “willful violation” of her 2021 injunction intended to open up iOS App Store payments. That contempt of court finding has now been almost entirely upheld by the Ninth Circuit Court of Appeals, a development that Epic Games’ Tim Sweeney tells Ars he hopes will “do a lot of good for developers and start to really change the App Store situation worldwide, I think.”
The ruling, signed by a panel of three appellate court judges, affirmed that Apple’s initial attempts to charge a 27 percent fee to iOS developers using outside payment options “had a prohibitive effect, in violation of the injunction.” Similarly, Apple’s restrictions on how those outside links had to be designed were overly broad; the appeals court suggests that Apple can only ensure that internal and external payment options are presented in a similar fashion.
The appeals court also agreed that Apple acted in “bad faith” by refusing to comply with the injunction, rejecting viable, compliant alternatives in internal discussions. And the appeals court was also not convinced by Apple’s process-focused arguments, saying the district court properly evaluated materials Apple argued were protected by attorney-client privilege.
We may earn a commission from links on this page.
Philips Hue is one of the most well-respected and popular brands in smart lights—but what about its smart security cameras? Parent company Signify has been developing Hue cameras for a couple of years now, with a video doorbell and 2K camera upgrades recently added to the portfolio of devices. (Note: This 2K version hasn’t yet landed in the U.S., but the existing 1080p versions are quite similar.)
I got a chance to test out the new 2K Hue Secure camera, and alongside all the basics of a camera like this, it came with an extra bonus that worked better than I expected: seamless integration with Philips Hue lights. These two product categories actually work better together than you might think.
While you can certainly connect cameras and lights across a variety of smart home platforms, Philips Hue is one of very few manufacturers making both types of device (TP-Link is another). That gives you a simplicity and interoperability you don’t really get elsewhere.
Hue cameras are controlled inside the same Hue app for Android or iOS as the Hue lights. You don’t necessarily need a Hue Bridge to connect the camera, too, as it can link to your wifi directly, but the Bridge is required if you want it to be able to sync with your lights—which is one of the key features here. (If you already have the lights, you’ll already have the Bridge anyway.)
The 2K Hue Secure wired camera I’ve been testing comes with a 2K video resolution (as the name suggests). two-way audio, a built-in siren, infrared night vision, and weatherproofing (so you can use it indoors or out). As well as the wired version I’ve got here, there’s also a battery-powered option, and a model that comes with a desktop stand.
Once configured, the camera lives in the same Home tab inside the mobile app as any Philips Hue lights you’ve got. The main panel doesn’t show the camera feed—instead, it shows the armed status of the camera, which can be configured separately depending on whether you’re at home or not. The idea is that you don’t get disturbed with a flurry of unnecessary notifications when you’re moving around.
The basic functionality is the same as every other security camera: Motion is detected and you get a ping to your phone with details, with a saved clip of the event that stays available for 24 hours. You can also tap into the live feed from the camera at any time, should you want to check in on the pets or the backyard.
As is often the case with security cameras, there is an optional subscription plan that gives you long-term video clip storage, activity zone settings, and AI-powered identification of people, animals, vehicles, and packages. That will set you back from $4 a month, with a discount if you pay for a year at a time.
I started off a little unsure about just how useful it would be to connect up the Hue cameras and Hue lights—it’s not a combination that gets talked about much—but it’s surprisingly useful. If you delve into the camera settings inside the Hue app, there’s a Trigger lights section especially for this.
You get to choose which of your lights are affected—they don’t all have to go on and off together—and there are customizations for color and brightness across certain time schedules. You could have your bulbs glowing red during the night, for example, or turning bright blue during the daytime. The duration the lights stay on for can also be set.
It’s not the most sophisticated system, but it works: If someone is loitering around your property, you can have a selected number of lights turn on to put them off, or to suggest that someone is in fact at home. This is in addition to everything else you can do, including sounding a siren through the camera, and because it works through the Hue Bridge it all happens pretty much instantaneously.
You can also set specific cameras as basic motion sensors for you and your family—lighting up the way to the bathroom late at night, for example. This can work even when the system is disarmed, so there’s no wifi video streaming happening, but the cameras are still watching out for movement and responding accordingly.
There’s one more option worth mentioning in the security settings in the Hue app: “mimic presence.” This can randomly turn your lights on and off at certain points in the day, and the schedule you choose can be controlled by whether or not your Hue security is armed or disarmed (so nothing happens when everyone is at home).
MIPS has signed an agreement to acquire all the shares in Koroyd, the impact-protection brand, for €40 million.
According to MIPS CEO and president Max Strandwitz, Koroyd will continue to operate separately under its leadership. But the acquisition will mean MIPS can expand its range of protection across consumer and professional uses.
“The acquisition of Koroyd marks another important step in MIPS’ long-term strategy to strengthen our position as the leader in helmet safety. This acquisition positions the group uniquely in the market and enables interesting entries into other adjacent product areas,” says Strandwitz.
Brands such as Smith and Endura have used Koroyd in their helmets. In fact, the Smith Optics Ignite MIPS helmet uses both of the ‘ingredient’ technologies.
Koroyd, which is based in Monaco and was founded in 2010, produces a variety of impact-protection technologies that span head, body and hand protection across sport, motorsport, defence, industrial safety and child-restraint systems.
Founder of Koroyd, John Lloyd, says: “From the beginning, Koroyd has been built on a culture of purposeful innovation, solving real problems with solutions that genuinely improve protection whilst putting the user experience first. In MIPS, we have found a partner who shares that ethos entirely.”
MIPS says the purchase of Koroyd amounts to €40m on a cash and debt-free basis, and with the possibility of an additional earn-out of €25m if performance targets are met.
“The acquisition is expected to contribute positively to MIPS’ earnings per share, EBIT margin, and sales growth on a short and long-term basis,” says MIPS.
For years, Sen. Ron Johnson has been spreading conspiracy theories and misinformation about COVID-19 and the safety of vaccines.
He’s promoted disproven treatments for COVID-19 and claimed, without evidence, that athletes are “dropping dead on the field” after getting the COVID-19 vaccination. Now the Wisconsin politician is endorsing a book by a discredited doctor promoting an unproven and dangerous treatment for autism and a host of ailments: chlorine dioxide, a chemical used for disinfecting and bleaching.
The book is “The War on Chlorine Dioxide: The Medicine that Could End Medicine” by Dr. Pierre Kory, a critical care specialist who practiced in Wisconsin hospitals before losing his medical certification for statements advocating using an antiparasite medication to treat COVID-19. The action, he’s said, makes him unemployable, even though he still has a license.
If you find yourself in an emergency or crisis situation, the more information you can give first responders, the better. Android users can now share a live stream of their surroundings with 911, allowing emergency services to assess and provide guidance in real time while you wait for help to arrive onsite.
Your Android already shares some information with first responders via Emergency Location Services (unless you disable this feature). This built-in tool sends an accurate location as well as contextual information, such as language settings, when you call or text an emergency number. Now, that includes live video from your device’s camera.
You don’t need to do anything to set up Emergency Live Video. Once available in your area, responders can send a request during an emergency call or text to securely share your camera’s live video. You’ll see a prompt on your screen to start sharing with one tap.
According to Google, Emergency Live Video is encrypted by default. Users can choose whether to share their video from the prompt as well as stop the share at any time by clicking the onscreen Stop sharing button.
Live video sharing is rolling out now to U.S. users, as well as those in parts of Germany and Mexico, on Android phones running Android 8 and up. Google says they are partnering with public safety organizations to expand the feature to more users.
Emergency Live Video is the latest in Google’s suite of safety features designed to make help more accessible—more quickly—in an emergency. Pixel users in Australia, North America, and several dozen countries across Europe now have access to Satellite SOS, which allows you to call emergency services even without a cellular or wifi connection. Car Crash Detection contacts emergency services and shares your location in the event of severe crash, while Fall Detection and Loss of Pulse Detection will call for help based on Pixel Watch sensor data.
The Washington Post’s top standards editor Thursday decried “frustrating” errors in its new AI-generated personalized podcasts, whose launch has been met with distress by its journalists. From a report: Earlier this week, the Post announced that it was rolling out personalized AI-generated podcasts for users of the paper’s mobile app. In a release, the paper said users will be able to choose preferred topics and AI hosts, and could “shape their own briefing, select their topics, set their lengths, pick their hosts and soon even ask questions using our Ask The Post AI technology.”
But less than 48 hours since the product was released, people within the Post have flagged what four sources described as multiple mistakes in personalized podcasts. The errors have ranged from relatively minor pronunciation gaffes to significant changes to story content, like misattributing or inventing quotes and inserting commentary, such as interpreting a source’s quotes as the paper’s position on an issue.
According to four people familiar with the situation, the errors have alarmed senior newsroom leaders who have acknowledged in an internal Slack channel that the product’s output is not living up to the paper’s standards. In a message to other WaPo staff shared with Semafor, head of standards Karen Pensiero wrote that the errors have been “frustrating for all of us.”
Read more of this story at Slashdot.
A startup established three years ago to churn out a new class of high-power satellites has raised $250 million to ramp up production at its Southern California factory.
The company, named K2, announced the cash infusion on Thursday. K2’s Series C fundraising round was led by Redpoint Ventures, with additional funding from investment firms in the United States, the United Kingdom, and Germany. K2 has now raised more than $400 million since its founding in 2022 and is on track to launch its first major demonstration mission next year, officials said.
K2 aims to take advantage of a coming abundance of heavy- and super-heavy-lift launch capacity, with SpaceX’s Starship expected to begin deploying satellites as soon as next year. Blue Origin’s New Glenn rocket launched twice this year and will fly more in 2026 while engineers develop an even larger New Glenn with additional engines and more lift capability.
Over the past year Intel engineers have worked a lot on Cache Aware Scheduling for the Linux kernel. The yet-to-be-merged functionality allows for the Linux kernel to better aggregate tasks sharing data to the same last level cache (LLC) domain to reduce cache misses and cache bouncing. The Cache Aware Scheduling development was led by Intel but helps other CPU vendors too for processors with multiple cache domains. Back in October I showed some nice performance wins for AMD EPYC Turin with Cache Aware Scheduling while today’s article are some benchmarks of the newest CAS code and looking at the performance benefit on Xeon 6 “Granite Rapids” processors.
I have a love–hate relationship with the smartwatch on my wrist. This relationship is no doubt shaped by the fact that I write about fitness tech for a living, but I know I’m not alone in succumbing to an obsession with numbers from my wearables. Did I hit 10,000 steps? What’s my resting heart rate today? Is my sleep score better than yesterday’s? When did progressive overload turn into screen time overload, too?
The fitness tech boom is showing no signs of slowing down any time soon—and with it, we consume a constant stream of promises that this data will make us healthier, stronger, and faster. With the sheer amount of health insights potentially available to us at any time, it’s easy to get overwhelmed. I’ve watched my least health-anxious friends become consumed by metrics they’d never heard of two years ago. They’re tracking bone density trends, obsessing over cortisol levels, panicking about stress scores that fluctuate for reasons no algorithm can fully explain. I can feel my fitness trackers pull me away from genuine wellness and into a mental health disaster. The good news: When I look up from my screens and start talking to real people, I see I’m not alone in wanting to unplug and push back against the overly quantified self.
When I put out a call on Instagram asking people about their relationship with posting workout data and fitness content, I received hundreds of responses from people exhausted by the performance of fitness. Even if your only audience is your own reflection, simply owning a wearable can create a real barrier between feeling good about your body and your fitness journey. Did I work out enough today? Will my friends see that I skipped a workout? Should I push through injury to maintain my streak?
For these reasons, celebrity trainer Lauren Kleban says she doesn’t like to rely on wearables at all. “Counting steps or calories can quickly spiral into a bit of an obsession,” says Kleban, and that “takes the joy out of movement and away from learning what’s truly best for us.” She says her clients want to focus on their mind and body connection, now more than ever. There’s a real, growing desire to rebuild a sense of intuition that doesn’t depend on feedback from a watch.
Similarly, Marshall Weber, a certified personal trainer and owner of Jack City Fitness, says that he’s “definitely been surprised by the growing push towards unplugged fitness,” but that he “totally gets it.” Weber says he’s had clients express feeling “overwhelmed with their Fitbit or Apple Watch micromanaging their training.” When every workout becomes about numbers and keeping up with an average, it’s all too easy to lose touch with your body. “The anti-tech movement is about taking back that personal connection,” Weber says. After all, when was the last time you finished a workout and didn’t immediately look at your stats, but instead just noticed how you felt?
This is the paradox at the heart of fitness technology. Tools designed to help us understand our bodies have created a new kind of illiteracy. Maybe you can tell me why you’re aiming for Zone 2 workouts, but can’t actually recognize what that effort feels like without a screen telling you. In a sense, you might be outsourcing your own intuition to algorithms.
If nothing else, the data risks are real. (Because if you think you own all your health data, think again.) Every heart rate spike, every missed workout, every late-night stress indicator gets recorded, stored, and potentially shared. Still, for me, the more insidious risk is psychological: the erosion of our ability to know ourselves without consulting a device first.
So what does unplugged fitness actually look like in practice? It’s not about rejecting all technology or pretending GPS watches and heart rate monitors don’t have value—I promise. Look, I crave data and answers as much as—and maybe more than—the average gym-goer. I’m simply not woo-woo enough to ditch my Garmin altogether.
Instead, I argue for re-establishing a hierarchy in which technology serves your training, not the other way around. “Sometimes, the best performance boost is just learning to listen to what your body is saying and feeling,” says Weber. But what does “listening to your body” actually look like?
If you’re like me, and need to rebuild a connection with your body from the ground-up, try these approaches:
Start with tech-free workouts. Designate certain runs, yoga sessions, or strength workouts as completely unplugged. No watch, no phone, no tracking. Notice what changes when there’s no device to check.
Relearn your body’s signals. Can you gauge your effort level without looking at a heart rate monitor? Do you actually know what “recovery pace” feels like for you, or are you just matching a number? Practice assessing fatigue, energy, soreness, and readiness without checking your watch.
Replace metrics with sensory awareness. Instead of tracking pace, notice your breathing pattern. Instead of counting calories burned, pay attention to how your muscles feel. Instead of obsessing over sleep scores, ask yourself a simple question in the morning: how do I actually feel?
Set goals that can’t be gamified. Rather than chasing step counts or streak days, aim for qualitative improvements. Can you hold a plank with better form? Does that hill feel easier than last month? Are you enjoying your workouts more? These are the markers of real progress.
Create tech boundaries. Maybe you use your GPS watch for long runs but leave it home for everything else. Perhaps you track workouts but delete the social features. Find the minimum effective dose of technology that serves your goals without dominating your headspace.
Reconnect with in-person community. The loss of shared gym culture—people actually talking to each other instead of staying plugged into individual screens—represents more than just nostalgia. There’s real value in working out alongside others, in having conversations about training instead of just comparing data, in building knowledge through shared experience rather than algorithm-driven insights.
Unplugging is easier said than done, but you don’t need to go cold turkey. Maybe in the new year, you can set “body literacy” as a worthwhile resolution. At the end of the day, exercise should add to your life, not become another source of performance anxiety. It should be energizing, not exhausting—and I don’t just mean physically. The never-ending irony of modern fitness culture is that in our pursuit of optimal health, we keep inventing new forms of stress and anxiety. When all forms of wellness come with trackable metrics and social pressure, I think we’ve fundamentally missed the point.
The remarkable, reimagined. Introducing Beolab 90
Skygazers in the Northern Hemisphere can look forward the peak of the Geminids shower on the night of the 13th (tomorrow),
pearOS, a Linux distro that aims to look and behave like Apple’s macOS, is once again in active development with a new base, design, installer, and more.
Carl Rinsch, the director behind the 2013 Keanu Reeves film “47 Ronin,” has been found guilty of defrauding Netflix out of $11 million that was meant to fund a science fiction series called “Conquest,” which the streaming company ultimately cancelled in 2021 after Rinsch failed to meet any production milestones. A jury in the Southern District of New York convicted the 48-year-old on seven charges: one count each of wire fraud and money laundering, and five counts of transacting in illicitly obtained property.
Prosecutors alleged that Rinsch funneled the $11 million through multiple bank accounts into a personal brokerage account, lost more than half of it on securities within two months, and then began speculating on cryptocurrency. Court records show he also spent $2.4 million on a Ferrari and five Rolls Royces, $3.3 million on furniture and antiques, and $387,000 on a Swiss watch. Netflix has written off $55 million in total and has not recovered any funds. Rinsch faces up to 90 years in prison and is scheduled for sentencing on April 17, 2026.
Read more of this story at Slashdot.
Reddit has filed a lawsuit in Australia’s High Court aiming to overturn the country’s under-16 social media ban, Reuters reported. The forum platform called the law contrary to Australia’s constitution as it intrudes on free political discourse. It also argued that Reddit shouldn’t have been included in the ban since it isn’t a social media site, based on the law’s definition. The action is likely to set in motion a protracted legal battle, given Reddit’s resources and its popularity in Australia
Australia’s minimum age social media ban, the first of its kind in the world, went into effect on December 10. The ten platforms affected, including Reddit, must bar underage users or face a fine of up to A$49.5 million ($33 million). Platforms are using a variety of means to determine age, including age inference based on activity and selfies.
However Reddit argued that the law comes with some “serious privacy and political expression issues” for users. “Australian citizens under the age of 16 will, within years if not months, become electors. The choices to be made by those citizens will be informed by political communication in which they engage prior to the age of 18,” it wrote in the filing.
The government disagreed, noting that Reddit filed the lawsuit to protect is profits, not children’s right to free expression. “It is action we saw time and time again by Big Tobacco against tobacco control and we are seeing it now by some social media or big tech giants,” said Health Minister Mark Butler.
With a market capitalization of $44 billion, Reddit certainly has the means to sustain a long fight. It would be motivated to do so as well, given that Australia is its fourth-largest market after Canada, the UK and the United States.
This article originally appeared on Engadget at https://www.engadget.com/social-media/reddit-sues-australia-over-underage-social-media-ban-143018208.html?src=rss
We may earn a commission from links on this page. Deal pricing and availability subject to change after time of publication.
The PlayStation 5 Pro’s price has finally eased up a bit, dropping to $648.99 from $749.99, which is the lowest it has gone on Amazon so far according to price trackers. For anyone who has been eyeing a console upgrade but didn’t want to spend full price, this discount makes it easier to consider.
The Pro looks almost identical to the regular PS5, but the changes inside matter a lot more than the shell. Sony bumped up the GPU power, increased memory bandwidth, and doubled storage to a 2TB SSD. In simple terms, you spend less time uninstalling old games to make room for new ones. Sony also added Wi-Fi 7, which won’t matter to everyone, but if your router is up to the task, downloads will be faster and more stable.
In use, the console feels familiar. You still get the fast loading PS5 owners enjoy, and the DualSense controller remains one of the best reasons to stay in the PlayStation ecosystem. The real difference comes when you fire up games that have been optimized for the Pro—titles like Spider-Man 2, God of War Ragnarok, and Horizon Forbidden West. These games run with higher frame rates and more stable performance. Visual details pop a bit more, especially on a capable TV. Other games will still run well, just without the added polish. You don’t lose anything by making the jump; you just get a bit more when the software allows it. PCMag gave the PS5 Pro an “Excellent” rating, largely due to its performance gains over the original.
That said, the PS5 Pro ships without a disc drive. If you’re someone who still buys physical games or has a stack of PS4 discs lying around, you’ll need to pick up the optional external drive for $79. If you’ve already gone all-digital, that extra cost doesn’t apply. Overall, the PS5 Pro isn’t a must-have for casual players, but it’s a smart buy for anyone who wants more visual fidelity now and doesn’t want to wait around for the PlayStation 6. And at this price, it’s a lot easier to justify the jump.
The Blender 3D modeling software could enjoy better macOS support with better cross-platform code paths thanks to in-development work for leveraging KosmicKrisp for Vulkan API usage on macOS via Metal…
At a time when many young people are using AI for personal and learning purposes, schools are trying to figure out what to teach about AI and how (find out more in this summer 2025 data about young people’s usage of AI in the UK). One aspect of this is how technical we should get in explaining how AI works, particularly if we want to debunk naive views of the capabilities of the technology, such as that AI tools ‘think’. In this month’s research seminar, we found out how AI contexts can be added to current classroom maths to make maths more interesting and relevant while teaching the core concepts of AI.
At our computing education research seminar in July, a group of researchers from the CAMMP (Computational and Mathematical Modeling Program) research project shared their work:
They talked about how maths already taught in secondary schools can be used to demystify AI. At first glance, this seems difficult to do, as it is often assumed that school-aged learners will not be able to understand how these systems work. This is especially the case for artificial neural networks, which are usually seen as a black box technology — they may be relatively easy to use, but it’s not as easy to understand how they work. Despite this, the Austrian and German team have developed a clear way to explain some of the fundamental elements of AI using school-based maths.
Sarah Schönbrodt started by challenging us to consider that learning maths is an essential part in developing AI skills, as:
Sarah explained how the CAMMP team have developed a range of teaching and learning materials on AI (and beyond) with an overall goal to “allow students to solve authentic, real and relevant problems using mathematical modeling and computers”.
She reflected that much of school maths is set in contexts that are abstract, and may not be very interesting or relevant to students. Therefore, introducing AI-based contexts, which are having a huge impact on society and students’ lives, is both an opportunity to make maths more engaging and also a way to demystify AI.
Workshops designed and researched by the team include contexts such as privacy in social networks to learn about decision trees, personalised Netflix recommendations to learn about k-nearest neighbour, word predictions to learn about N-Grams, and predicting life expectancy to learn about regression and neural networks.
For the seminar, Sarah walked through the steps to learn about support vector machines. This is an upper secondary workshop for students aged 17 to 18 years old. The context of the lesson is an image problem — specifically, classifying the data representing the colours of a simplified traffic light system (two lights to start with) to work out if a traffic light is red or green.
She walked through each of the steps of the maths workshop:
Throughout the presentation, Sarah pointed out where the maths taught was linked to the Austrian and German mathematics curriculum.
As well as learning about lines, planes, distances, dot product and statistical measures, learners are also engaged in discussing the social and ethical issues of the approach taken. They are encouraged to think about bias, data diversity, privacy, and the impact of errors on people. For example, if the model wrongly predicts a light as green when it is red, then an autonomous car would run through a red traffic light. This would likely be a bigger consequence than stopping at a green traffic light that was mis-predicted as red. So should the best line reduce this kind of error?
To teach the workshops, Sarah explained they have developed interactive Jupyter notebooks, where no programming skills are needed. Students fill in the gaps of example code, explore simulations, and write their ideas for discussion for the whole class. No software needs to be installed, feedback is direct, and there are in-depth tasks and staggered hints.
Stephan went on to introduce artificial neural networks (ANNs), which are the basis of generative AI applications like chatbots and image generation systems. He focused on regression models, such as those used in weather forecasting.
ANNs are very complex. Therefore, to start to understand the fundamentals of this technology, he introduced a ‘toy ANN’ with one input, three nodes, and one output. A function is performed on the input data at each node. With the toy network, the team wants to tackle a major and common misconception: that students think that ANN systems learn, recognise, see, and understand, when really it’s all just maths.
The learning activity starts by looking at one node with one input and one output, and can be described as a mathematical function, with a concatenation of two functions (in this case a linear and activation function). Stephan shared an online simulator that visualises how the toy neural network can be explored as students change two parameters (in this case, weight and bias of the functions). Students then look at the overall network, and the way that the output from the three nodes is combined. Again, they can explore this in the simulator. Students compare simple data about weather prediction to the model, and discover they need more functions — more nodes to better fit the data. The activity helps students learn that ANN systems are just highly adjustable mathematical functions that, by adding nodes, can approximate relationships in a given data set. But the approximation only works in the bounds (intervals) in which data points are given, showing that ANNs do not ‘understand’ or ’know’ — it’s just maths.
Stephen finished by explaining the mutual benefits of AI education and maths education. He suggested maths will enable a deeper understanding of AI, and give students a way to realistically assess the opportunities and risks of AI tools and show them the role that humans have in designing AI systems. He also explained that classroom maths education can benefit from incorporating AI contexts. This approach highlights how maths underpins the design and understanding of everyday systems, supports more effective teaching, and promotes an interdisciplinary way of learning across subjects.
I have been researching the teaching of AI and machine learning for around five years now, since before ChatGPT and other similar tools burst on the scene. Since then, I have seen an increasing number of resources to teach about the social and ethical issues of the topic, and there are a bewildering number of learning activities and tools for students to train simple models. There are frameworks for the data lifecycle, and an emerging set of activities to follow to prepare data, compare model types, and deploy simple applications. However, I felt the need to understand and to teach about, at a very simple level, the basic building blocks of data-driven technologies. When I heard the CAMMP team present their work at the AIDEA conference in February 2025, I was entirely amazed and I asked them to present here at our research seminar series. This was a piece of the puzzle that I had been searching for — a way to explain the ‘bottom of the technical stack of fundamental concepts’. The team is taking very complex ideas and reducing them to such an extent that we can use secondary classroom maths to show that AI is not magic and AI systems do not think. It’s just maths. The maths is still hard, and teachers will still need the skills to carefully guide students step by step so they can build a useful mental model.
I think we can simplify these ideas further, and create unplugged activities, simulations, and ways for students to explore these basic building blocks of data representation, as well as classification and representing approximations of complex patterns and prediction. I can sense the beginnings of new ideas in computational thinking, though they’re still taking shape. We’re researching these further and will keep you updated.
If you would like to find out more about the CAMMP resources, you can watch the seminar recording, look at the CAMMP website or try out their online materials. For example, the team shared a link to the jupyter notebooks they use to teach the workshops they demonstrated (and others). You can use these with a username of ‘cammp_YOURPSEUDONYM’, where you can set ‘YOURPSEUDONYM’ to any letters, and you can choose any password. They also shared their toy ANN simulation.
The CAMMP team are not the only researchers who are investigating how to teach about AI in maths lessons. You can find a set of other research papers here.
In our current seminar series, we’re exploring teaching about AI and data science. Join us at our last seminar of the series on Tuesday, 27 January 2026 from 17:00 to 18:30 GMT to hear Salomey Afua Addo talk about using unplugged approaches to teach about neural networks.
To sign up and take part, click the button below. We’ll then send you information about joining. We hope to see you there.
The schedule of our upcoming seminars is online. You can catch up on past seminars on our previous seminars page.
The post Secondary school maths showing that AI systems don’t think appeared first on Raspberry Pi Foundation.
