Increasingly, pop culture shapes how science is imagined, discussed, and in some cases, how it is developed.

From Jurassic Park to The Last of Us and cutting-edge debates about the safety of artificial intelligence (AI), fictional narratives do more than entertain.

They shape the frameworks through which audiences – including scientists, policymakers and funders – make sense of complex scientific ideas and of science itself. In doing so, they influence what seems possible and plausible, as well as what we want and fear.

From Jurassic Park to reality

Your scientists were so preoccupied with whether they could, they didn’t stop to think if they should.

This famous line, delivered by fictional mathematician Ian Malcolm in Jurassic Park, has become a touchstone in debates about emerging technologies.

Take de-extinction. When biotechnology company Colossal Biosciences announced plans in 2021 to revive bygone species such as the woolly mammoth, the comparison was immediate: Jurassic Park. The film has become a cultural shorthand for the promises and pitfalls of bringing extinct species back to life.

Scientists and commentators alike invoke its famous ethical warning – that the question of whether we should do something is separate from whether we can. These references are not merely rhetorical. They shape how research is communicated, debated and understood.

By framing de-extinction through a familiar narrative, Jurassic Park has influenced public expectations, ethical anxieties and media discourse. We see projects described as “real-life Jurassic Park”, debates about whether such technologies should be pursued citing the film, and journalists using it as a shorthand when covering emerging biotechnologies.

Assimilating aliens and fungal zombies

The influence of science fiction can extend to scientific practice itself. Researchers named DNA elements which incorporate foreign genetic material “Borgs”, for example, after the assimilating aliens from Star Trek.

A similar dynamic can be seen in responses to HBO’s The Last of Us, which imagines a global pandemic caused by a parasitic fungus that transforms humans into zombie-like creatures. Following the show’s release, scientists reported renewed public interest in fungal pathogens.

Indeed, the “worst-case scenario” presented in the series prompted immunologists and mycologists to examine the biological plausibility of a fungal leap to humans.

While the temperature of the human body is inhospitable to most kinds of fungus, and we need not fear the aggressive biting depicted in fiction, experts warn that climate change and agricultural fungicide overuse are accelerating fungal adaptation to higher temperatures. This makes The Last of Us a sobering alarm for real-world problems.

In both cases, pop culture does not simply reflect scientific knowledge. It shapes how that knowledge is encountered, interpreted and imagined.

Killer superintelligence

One of the most compelling examples of this feedback loop today is AI. Popular culture has long been fascinated with intelligent machines, often imagining them as existential threats. We see this from deceptive superintelligences to human extinction, as portrayed in Ex Machina, The Matrix and The Terminator. These narratives have left a deep imprint on public consciousness.

Today, similar themes appear in real-world debates about AI safety. Prominent figures in AI debates, such as Nick Bostrom, Eliezer Yudkowsky and Geoffrey Hinton, have warned about the potential risks of advanced AI. The warnings include scenarios that echo earlier fictional imaginings.

While these arguments are grounded in technical and philosophical work, they resonate so widely in part because they align with familiar cultural narratives.

This does not mean concerns about AI are simply fictional. Rather, it shows how deeply intertwined scientific thinking and cultural imagination can be.

Understanding the feedback loop

Pop culture helps establish the language, metaphors, and expectations through which emerging technologies are understood. It shapes how scientific ideas, ideas about science, and images of scientists circulate beyond laboratories and institutions – and, in turn, how science is understood, valued and positioned in society.

At the same time, science continues to feed back into pop culture. Advances in genetics, epidemiology and AI provide new material for storytellers, shaping the kinds of futures that are imagined on screen. The result is a dynamic feedback loop: science inspires stories, and those stories in turn influence how science develops.

Despite this, the role of pop culture is rarely acknowledged in how we think about science policy and funding. Discussions tend to focus on infrastructure and technical capability, while overlooking the cultural forces that shape public imagination.

Yet these forces play a crucial role in determining which scientific futures feel worth pursuing. This matters because public perception influences everything from research funding to regulatory priorities.

If certain technologies are seen as exciting, frightening or inevitable, this affects how they are supported, scrutinised or resisted. Pop culture is one of the key arenas in which these perceptions are formed.

• Anna-Sophie Jürgens is a Senior Lecturer in Science Communication at the Australian National Centre for the Public Awareness of Science, and Founder of Popsicule, ANU’s Science in Popular Culture and Entertainment Hub, Australian National University
• Shao-Jie Jhou is a PhD Candidate, Centre for the Public Awareness of Science, Australian National University
• This article first appeared in The Conversation

The post From Jurassic Park to dreams of AI doom, pop culture shapes science more than we like to admit appeared first on Stuff South Africa.

Hachette Book Group was approached with what The New York Times claimed was evidence that Shy Girl by Mia Ballard was allegedly AI-generated. Following this, the publisher said its imprint Orbit was removing the book from publication in the US and UK.

The novel follows Gia, a young woman who is “lonely, broke and depressed with a serious case of OCD”. She encounters a mysterious and rich man who, in exchange for her living as his devoted pet, promises to erase all her debts. The novel follows her time in captivity as she becomes increasingly animalistic in nature.

In an email to The New York Times, Ballard said the controversy “has changed my life in many ways and my mental health is at an all-time low”. Ballard has denied personally using AI to write the novel. But she has said that an acquaintance she hired to work on an earlier self-published version incorporated AI tools.

Many people disagree with the use of AI for a host of reasons, from environmental to ethical concerns. But cultivating a climate of distrust around writing and authors is also not necessarily productive, and further pushes AI use into secrecy.

The author now faces a challenging situation, as Hachette withdrawing the book will appear to some to validate the accusations, even if it simply reflects uncertainty.

What happened?

The book was initially self-published in February 2025 before it was bought by Orbit Books, following a growing industry trend to traditionally publish successful self-published or fan-fiction works.

Issues started to arise regarding the novel’s provenance in mid-2025 on Reddit when one user, who claimed they were a book editor, made a post which pointed out several issues with the novel that suggested it was AI-generated.

Their main claim was based on the novel’s repetitive style, something also pointed out by other critical readers. Specifically, they highlighted that almost every noun is preceded by an adjective, actions are frequently described with similes, descriptions come in lists of three, and certain words are overused.

The discussion spread to other platforms such as the BookTok community (TikTok users dedicated to discussing books and publishing), Instagram and YouTube.

There is still no final consensus about how Shy Girl was written and Ballard has removed herself from the public eye and taken her social media accounts offline following the scandal. Hachette told The Independent that they “remain committed to protecting original creative expression and storytelling”. They have made no definitive statement on the claims but did tell the NYT that they conducted a thorough and lengthy review of the text.

How should readers and publishers respond?

Readers and publishers have spent years debating the impact of AI in the abstract but 2026 is the year these debates have become reality.

Stories like Shy Girl and The New York Times’ profile of AI romance author Coral Hart, who boasted of using AI to write and self-publish 200 hundred books across 21 pen names in a recent profile by The New York Times, demonstrate that theoretical disputes did not prepare us to be confronted with the reality of AI.

It’s clear that even the suggestion of AI writing inspires immense disgust in many readers. This means that regardless of the truth (if we ever find it out) Shy Girl and Ballard will likely be tainted by this scandal. Therefore, we must ask whether it is possible for publishing and reading to survive not just AI’s increasing normalisation but also the hostile and suspicious environment its use is creating for writers.

As a researcher of contemporary and digital reading culture, I believe we should cultivate an openness around the use of AI in writing by lobbying publishers to provide this information openly and clearly. This is already starting to happen. The Society of Authors, which is the UK’s largest writers’ trade union, has launched a logo to be used to identify “human authored” books – a step toward empowering consumers to know what they are choosing to support with their money.

Copyright law also needs to reflect AI’s reshaping of the creative field. A work requires a human author to be covered under copyright law in the US, and any doubts about this are potentially a big part of Hachette pulling Shy Girl from publication due to the publisher’s inability to copyright.

This creates a difficult position for the novel and author. The book’s cancellation looks like confirmation of guilt, whereas it may just be doubt. However, UK copyright law does offer protection for computer-generated works. This creates a murky area where AI-generated or assisted works can receive certain legal protections, but not necessarily the same rights as human-authored works.

Under UK law, computer-generated works can qualify for copyright, with authorship attributed to the person who made the necessary arrangements for the work’s creation. However, these works do not benefit from the full range of protections afforded to human authors, particularly moral rights, such as the right to be identified as the author or to object to derogatory treatment of the work.

This framework may change following a recent consultation led by the UK government on copyright and artificial intelligence. The consultation has now closed and the government has not yet implemented definitive legislative changes. However, its stated priorities suggest any reforms will aim to balance protecting creators’ rights with supporting innovation, investment and growth in the AI sector.

It’s an undeniably fraught situation, which is continually developing. In the near future we may unfortunately see more authors like Ballard made examples of while, behind the scenes, many more may be using AI undetected.

• Natalie Wall is a PhD in English Literature, University of Liverpool
• This article first appeared in The Conversation

The post A popular horror novel was pulled over AI concerns – here’s what it means for publishing appeared first on Stuff South Africa.

The extraordinary complexity of the brain’s billions of neurons makes this a very difficult task, of course, even in the era of AI and big data. Until now, whole-brain models have struggled to capture what makes each brain unique.

People’s brains are all wired slightly differently, so everyone has a unique network of neural connections that represents a kind of “brain fingerprint”.

However, most so-called brain twins are currently more like distant cousins. Their performance is barely any closer to the real thing than if the model were using the wiring diagram of a random stranger.

This matters because digital twins are increasingly proposed as tools for testing treatments by computer simulation, before applying them to real people. If these models fail to capture fundamental principles of each patient’s unique brain organisation, their predictions won’t be personalised – and in worst cases could be misleading.

In our latest study, published in Nature Neuroscience, we show that realistic digital brain twins require something that many existing models overlook: competition between the brain’s different systems.

Our findings suggest that without competition, digital twins risk being overly generic, missing out on what makes you “you”.

Excess of cooperation

The human brain is never static. The ebb and flow of its activity can be mapped non-invasively using neuroimaging methods such as functional MRI. A computer model can be built from this, specific to that person and simulating how the regions of their brain interact. This is the idea of the digital twin.

The brain is often described as a highly cooperative system. Yet everyday experiences such as focusing attention or switching between tasks tells us intuitively that brain systems compete for limited resources. Our brains cannot do everything at once, and not all regions can be active together all the time.

Despite this, the vast majority of brain simulations over the past 20 years have not taken these competitive interactions between regions into account. Rather, they have “forced” neighbouring regions to cooperate. This can push the simulated brain into overly synchronised states that are rarely seen in real brains.

In a large comparative study of humans, macaque monkeys and mice, our international team of researchers used non-invasive brain activity recordings to show that the most realistic whole-brain models not only require cooperative interactions within specialised brain circuits, but long-range competitive interactions between different circuits.

To achieve this, we compared two types of brain model: one in which all interactions between brain regions were cooperative, and another in which regions could either excite or suppress each other’s activity. In humans, monkeys and mice, the models that included competitive interactions consistently outperformed cooperative-only models.

Using a large-scale analysis of over 14,000 neuroimaging studies, we found that spontaneous activity in the competitive models more faithfully reflected known cognitive circuits, such as those involved in attention or memory. This suggests competition is crucial for enabling the brain to flexibly activate appropriate combinations of regions – a hallmark of intelligent behaviour.

Visual summary of our study:

We concluded that competitive interactions act as a stabilising force, allowing different brain systems to take turns in shaping the direction of the brain’s ebbs and flows without interference or distraction. This ability to avoid runaway activity may also contribute to the remarkable energy-efficiency of the mammalian brain, which is many orders of magnitude more efficient than modern AI systems.

Crucially, models with competitive interactions were not only more accurate but also more individual-specific. This means they were better at capturing the unique brain fingerprint that distinguishes one person’s brain from another’s.

No longer lost in translation?

The fact that our findings hold across humans and other mammals suggests they reflect fundamental principles of how intelligent systems work. In each case, we found models with competitive interactions generated brain activity patterns that closely resembled those associated with real cognitive processes.

This could have major implications for translational neuroscience. Animal models are routinely used to test treatments before human trials, yet differences between species often limit how well these results translate. Around 90% of treatments for neuropsychiatric disorders are “lost in translation”, failing in human clinical trials after showing promise in animal trials.

Combining brain imaging data from human patients with whole-brain modelling could radically change this. A framework that works across species would provide a powerful bridge between basic research and clinical application.

If someone needs intervention in the brain, for example due to epilepsy or a tumour, their digital twin could be used to explore how the patient’s brain activity would change when stimulated with different levels of drugs or electrical impulses. This might significantly improve on existing trial-and-error approaches with real patients, and thus provide better treatments.

The general principles of brain organisation across species also offer a path for understanding how to shape the next generation of artificial intelligence. In the not-too-distant future, we may be able to construct digital twins that are more faithful in reproducing the salient features of the human brain – and potentially, AI models that are more faithful to the human mind.

• Andrea Luppi is a Senior Research Associate, Department of Psychiatry, University of Oxford
• Gustavo Deco is a Professor of Computational Neuroscience, Center for Brain and Cognition, Universitat Pompeu Fabra
• Morten L. Kringelbach is a Professor of Neuroscience, University of Oxford
• This article first appeared in The Conversation

The post How to build a digital ‘twin’ of the human brain – what existing models overlook appeared first on Stuff South Africa.

NASA has just reset its Artemis program, marking a clear strategic shift: Space exploration is moving away from a race to achieve milestones and toward a system built on repeated operations, a sustained presence and lunar infrastructure that could become part of the technology networks we rely on here on Earth.

That shift is reflected in newly announced plans to invest billions of dollars in building a long-term lunar base, with habitats, power systems and surface infrastructure designed to support ongoing human activity. The message? Humans have already normalised travel to space. The next step is normalising living beyond Earth.

Artemis is NASA’s plan to return people to the Moon with the goal of staying. Unlike the short Apollo missions of the 1960s and 1970s, it consists of increasingly complex missions: flying around the Moon, landing on its surface and eventually establishing a base near the lunar south pole. The program aims to create a reliable way for humans to live and work there, develop technologies useful on Earth and prepare for the journey to Mars.

Rather than moving straight from the upcoming Artemis II crewed lunar flyby to a surface landing, the new road map adds an intermediate mission in 2027. Astronauts will test docking, life-support systems and communications with commercial lunar landers from SpaceX and Blue Origin, but in low Earth orbit, the region roughly 100 to 1,200 miles (160 to 2,000 kilometres above Earth’s surface, where rescue remains possible.

The first landing near the lunar south pole is now targeted for 2028. This timeline may sound delayed, but in reality, it has been deliberately reset to prioritise building reliable systems that can operate long into the future over speed.

As a professor of air and space law, I’ve been watching these developments closely. The United States is still in a race – particularly with China – but it is choosing to compete on its own terms. Rather than chasing the fastest possible landing, NASA is focused on building a system that can support repeated missions and a lasting human presence.

From sprint to system

The original Artemis plan aimed to leap quickly from test flights to a crewed landing while simultaneously developing new rockets, spacecraft and landing systems. That approach carried risk. Artemis I, an uncrewed mission, flew successfully in 2022. After a few delays, Artemis II is now nearing launch, with windows planned for early April 2026. But the further jump to a safe and reliable landing remains significant.

NASA’s new road map slows the transition deliberately. Instead of stand-alone milestones, NASA is now building a sequence of repeatable steps to gain hands-on experience.

This change includes a substantial new investment, with a multiphase plan for a lunar base with habitats, power systems and the surface infrastructure needed for a long-term human presence on the Moon. Consistent launch cadence and repeatable operations are how teams develop the expertise needed for safe, reliable spaceflight and eventually for travelling to Mars.

This shift is reflected in the decision to pause the planned lunar Gateway station, a small space station intended to orbit the Moon, and prioritise infrastructure on the lunar surface itself, where astronauts will live, work and build over time.

The new changes also emphasise a shifting role for commercial companies. SpaceX’s and Blue Origin’s lunar landers are integrated into the mission architecture.

The 2027 test mission, for example, will practice docking between crewed spacecraft and new commercial lunar landers in low Earth orbit. NASA is coordinating a network of public and private partners rather than running a single government-run Apollo-like program.

This method spreads risk across partners, lowers costs and speeds development, though success now depends on multiple players working reliably together.

Law follows activity

NASA’s road map is not just about lowering technical risk. It is also about shaping the future environment of lunar activity.

International space law, including the 1967 Outer Space Treaty, sets out broad principles to guide space activities, like avoiding harmful interference with others’ activities. But those rules only gain real meaning through repeated, coordinated activity, especially on the lunar surface, where desirable landing sites are limited.

Countries and companies that maintain a sustained presence on the Moon will shape the practical expectations everyone will share while living and working on the Moon. One-off demonstrations, like lunar landings, don’t shape lunar activity like continued operations would.

Why this matters – even if you never go to space

It would be easy to see these changes as purely technical, but they are not. The structure of a space program shapes what technologies are developed, how industries grow and which countries influence how space is used. Technologies developed for sustained lunar activity, including life-support systems, energy storage and advanced communications, have found applications on Earth, from medicine to disaster response.

There are economic effects as well. The Artemis program supports jobs across the United States and among its international partners. It helps build industries that extend far beyond NASA itself.

And there is a strategic dimension. As more countries and companies operate in space, the question is no longer just who arrives first, but who helps define how activity is carried out. Over time, that presence will likely become part of the infrastructure that supports daily life on Earth.

Communications, navigation, supply chains and scientific data already depend on space-based systems. As activity expands to the Moon, facilities there, from energy systems to communications relay systems that transmit data and signals back to Earth, will become integrated into those networks. What is built on the Moon will not sit apart from life on Earth, but increasingly function as an extension of it.

The Moon is becoming a place where infrastructure, industry and rules and expectations for how humans operate there are already beginning to take shape. NASA’s updated plan signals that the United States intends to be present there consistently.

The updates to the Artemis program are a statement about how the United States intends to engage in the next phase of space exploration. Rather than pursuing a single dramatic landing, the U.S. is committing to the steady, repeatable work of building a lasting foothold on the Moon, and redefining humanity’s relationship with space itself.

• Michelle L.D. Hanlon is a Professor of Air and Space Law, University of Mississippi
• This article first appeared in The Conversation

The post NASA wants to build a base on the Moon by the 2030s – how and why it plans to build up to a long‑term lunar presence appeared first on Stuff South Africa.

But as the UK’s chief medical officer recently warned, that may not be wise when it comes to medical decisions. In a recent study, colleagues and I tested how well large language model (LLM) chatbots help the public deal with common health problems. The results were striking.

The chatbots we tested were not ready to act as doctors. A common response to studies like this is that AI moves faster than academic publishing. By the time a paper appears, the models tested may already have been updated. But studies using newer versions of these systems for patient triage suggest the same problems remain.

We gave participants brief descriptions of common medical situations. They were randomly assigned either to use one of three widely available chatbots or to rely on whatever sources they would normally use at home. After interacting with the chatbot, we asked two questions: What condition might explain the symptoms? And where should they seek help?

People who used chatbots were less likely to identify the correct condition than those who didn’t. They were also no better at determining the right place to seek care than the control group. In other words, interacting with a chatbot did not help people make better health decisions.

Strong knowledge, weak outcomes

This does not mean the models lack medical knowledge because LLMs can pass medical licensing exams with ease. When we removed the human element and gave the same scenarios directly to the chatbots, their performance improved dramatically. Without human involvement, the models identified relevant conditions in the vast majority of cases and often suggested appropriate levels of care.

So why did the results deteriorate when people actually used the systems? When we looked at the conversations, the problems emerged. Chatbots frequently mentioned the relevant diagnosis somewhere in the conversation, yet participants did not always notice or remember it when summarising their final answer.

In other cases, users provided incomplete information or the chatbot misinterpreted key details. The issue was not simply a failure of medical knowledge – it was a failure of communication between human and machine.

The study shows that policymakers need information about the real-world performance of technology before introducing it into high-stakes settings such as frontline healthcare. Our findings highlight an important limitation of many current evaluations of AI in medicine. Language models often perform extremely well on structured exam questions or simulated “model-to-model” interactions.

But real-world use is much messier. Patients describe symptoms in a vague or incomplete way and can misunderstand explanations. They ask questions in unpredictable sequences. A system that performs impressively on benchmarks may behave very differently once real people begin interacting with it.

It also underscores a broader point about clinical care. As a GP, my job involves far more than recalling facts. Medicine is often described as an art rather than a science. A consultation isn’t simply about identifying the correct diagnosis. It involves interpreting a patient’s story, exploring uncertainty and negotiating decisions.

Medical educators have long recognised this complexity. For decades, future doctors were taught using the Calgary–Cambridge model. This meant building a rapport with the patient, gathering information through careful questioning, understanding the patient’s concerns and expectations, explaining findings clearly and agreeing a shared plan for management.

All these processes rely on human connection, tailored communication, clarification, gentle probing, judgement shaped by context and trust. These qualities cannot easily be reduced to pattern recognition.

A different role for AI

Yet the lesson from our study is not that AI has no place in healthcare. Far from it. The key is understanding what these systems are currently good at and where their limitations lie.

One useful way to think about today’s chatbots is that they function more like secretaries than physicians. They are remarkably effective at organising information, summarising text and structuring complex documents. These are the kinds of tasks where language models are already proving useful within healthcare systems, for example in drafting clinical notes, summarising patient records or generating referral letters.

The promise of AI in medicine remains real, but its role is likely to be more supportive than revolutionary in the near term. Chatbots should not be expected to act as the front door to healthcare. They are not ready to diagnose conditions or direct patients to the right level of care.

Artificial intelligence may be able to pass medical exams. But just as passing a theory test doesn’t make you a competent driver, practising medicine involves far more than answering questions correctly. It requires judgement, empathy and the ability to navigate the complexity that sits behind every clinical encounter. For now, at least, that requires people rather than bots.

• Rebecca Payne is a Clinical Senior Lecturer, Bangor University; University of Oxford
• This article first appeared in The Conversation

The post Why AI health chatbots won’t make you better at diagnosing yourself – new research appeared first on Stuff South Africa.

Konstantinos Kolokythas, a radio astronomer and postdoctoral research fellow at Rhodes University and the South African Radio Astronomy Observatory (SARAO), has led research into what these radio emissions reveal about our cosmic history. His findings provide a glimpse of what powerful instruments like MeerKAT and the upcoming Square Kilometre Array (SKA) will discover as they explore the “invisible” radio universe.

What has MeerKAT found, thanks to its sensitivity?

Think of a galaxy cluster not as a collection of thousands of galaxies, but as a bustling city. While telescopes usually see the “bright lights” of individual galaxies, MeerKAT has enabled us to detect the faint “smog” or “mist” filling the streets between them. Our search has been for this extremely faint “diffuse radio emission”. It is spread over millions of light-years, like a thin, glowing fog.

In the vast spaces between galaxies lies the Intracluster Medium – an incredibly hot, thin gas that fills the cluster. While the gas itself is usually seen by X-ray telescopes, it also contains magnetic fields and electrons travelling at nearly the speed of light.

When galaxy clusters merge, it is like a cosmic dance: the electrons encountering a magnetic field are compelled to spiral along the magnetic field lines, emitting energy as radio waves. This is the radio emission we see at 1.28 GHz with MeerKAT. It reveals the places of shock accelerations (the aftermath of cosmic collisions).

Our research within the MeerKAT Galaxy Cluster Legacy Survey (MGCLS), a programme led by the South African Radio Astronomy Observatory, used this capability to map 115 of these “cosmic cities”. We identified 103 diffuse sources, including 60 structures that were completely invisible to previous generations of telescopes. The legacy survey also produced its own overview.

We have essentially moved from having a blurry map of the neighbourhood to a high-definition atlas, revealing that the “empty” space between galaxies is actually teeming with energy. By combining this radio data with X-ray and optical observations, we can calculate the “energy budget” – essentially a full accounting of all the power, heat and magnetic energy moving through these massive structures.

How does this clarify or add to what was known before?

Before this work, we mainly observed only the brightest, most violent merger events. With our new catalogue, we can see the broader picture of cosmic evolution, detecting the faintest structures arising from galaxy cluster collisions. By identifying these features in over half (54%) of the surveyed clusters, we can study how energy is processed on a cosmic scale.

These radio signatures are the “scars” left by cluster mergers – colossal, slow-motion collisions where gravity draws two massive collections of galaxies together. This process generates turbulence and shockwaves that “kick” particles to extreme speeds.

Our findings demonstrate that these high-energy events are a fundamental part of a cluster’s life cycle and the universe’s evolution. Clusters that appear “quiet” or “relaxed” in X-ray light often conceal a history of radio activity. We are mapping the secret structures of magnetic fields over billions of years. In radio astronomy, the universe is never truly silent.

What direction does this point to for future research?

This catalogue serves as a high-resolution “baseline” for the coming decade. With MeerKAT, we have pushed the limits further, allowing us to observe more “ultra-steep spectrum” sources – faint emissions from the oldest, most “tired” particles in the universe. These are vital for understanding the long-term lifecycle of cosmic energy.

Looking forward, this research paves the way for the Square Kilometre Array (SKA) observatory, the world’s largest and most sensitive radio telescope, which is expected to be fully operational by 2030. If MeerKAT can detect 60 new structures in a small patch of the sky, the SKA will likely find thousands.

Why does this matter?

Because these structures forming in clusters are the largest “natural laboratories” in the universe. By studying them, we aren’t just looking at pretty pictures; we are learning how gravity, magnetism and matter behave on a scale that is otherwise impossible to recreate and the human mind can barely conceive.

This research proves that South Africa is at the forefront of this discovery, using homegrown technology to answer the deepest questions about the fabric of our universe, where our universe came from and how it evolves.

• Konstantinos Kolokythas is a Postdoctoral research fellow, Rhodes University
• This article first appeared in The Conversation

The post South Africa’s MeerKAT telescope is mapping previously invisible spaces between galaxies – and it’s found 60 new cosmic structures appeared first on Stuff South Africa.

Molly Russell, at just 14, took her own life. Her parents blame the apps on her phone for exposing her to graphic and disturbing content that took control of her mindset.

These stories are not unique. Data from thousands of people shows that social media increases loneliness, depression, anxiety and suicidal thoughts. Last week, a jury in California found that Meta and YouTube were liable for causing a teenager’s addiction to social media. The idea that social media causes harm is no longer in dispute.

The proposed response – in Australia, now proposed in the UK and elsewhere – is to ban social media for under-16s. It is an understandable impulse. But there are good reasons to think it won’t work – despite politicians claiming a successful start to the ban.

Teenagers have always found ways around rules. Getting an older sibling to buy alcohol is a time-honoured tradition. When it comes to social media, teenagers are more tech-savvy than the adults trying to restrict them, and evidence is emerging that many are working around the age verification systems put in place to enforce bans, such as by using VPNs (virtual private networks).

Rules will exist, but compliance will be patchy and hard to enforce. Those most determined to access social media may also be the most resourceful in getting around restrictions. This means that the teenagers most at risk may also be the least affected by a ban. Evidence from other areas shows that when certain activities are driven underground, they often become more harmful.

Not neutral tools

Even if the bans worked perfectly, they would address only part of the problem. It is difficult to disentangle the harms of social media from the devices that deliver it.

Smartphones are not neutral tools: they are engineered to hold attention through constant notifications, “frictionless” access to content, and rewards for regular interactions. Research links smartphone use – not just social media – to disrupted sleep, impaired attention and cognition, mental health problems, physical ailments such as chronic back pain and addiction.

Social media is one component of a broader “smartphone ecosystem”, and targeting one app while leaving the ecosystem intact is unlikely to solve much.

If social media is blocked, teenagers are not going to put their phones down. They will migrate to mobile games, group chats and endless web browsing – activities that rely on the same design features driving their social media use: notifications, streaks (features that track consecutive days of use and reward consistency), infinite scroll. The problem is not any single app but a pattern of behaviour that will find new outlets.

Nor is this only a problem for teenagers. Adults struggle with excessive smartphone use, too. Heavy use is associated with poorer sleep, reduced attention and higher stress – and in some respects, the adult consequences are more severe. Distracted driving, often fuelled by phone use, kills thousands of people every year.

This matters for teenagers because behaviour is learned by watching others. Children who see parents, teachers and other adults checking their phones absorb that as the norm. A policy that targets only young people does nothing to change the culture they are growing up inside.

And opting out is becoming harder for everyone. Primary school children are expected to use smartphones for homework – on apps that share more than a passing resemblance to addictive games. Online banking has become more difficult without one. Workplaces assume employees are reachable via multiple WhatsApp groups at all hours.

When opting out means opting out of modern life, restricting access to one category of app starts to look less like a solution and more like a gesture.

If the goal is to reduce harm, the focus needs to widen. The deeper issue is the central role smartphones now play in everyday life – for all of us, not just teenagers. That points towards different kinds of intervention: delaying smartphone adoption among younger children, encouraging simpler devices, redesigning compulsive features across all apps, and ensuring that essential services such as banking, education and travel stop assuming everyone is glued to a screen.

Banning social media for teenagers may feel like decisive action. But until the broader dependency is addressed, it will not deliver the change its advocates are hoping for.

• Jeremy Howick is a Professor and Director of the Stoneygate Centre for Excellence in Empathic Healthcare, University of Leicester
• This article first appeared in The Conversation

The post Why a social media ban for teenagers misses the point appeared first on Stuff South Africa.

The science behind this sounds intimidating – DNA sequencing, mRNA vaccines, “neoantigens” – but at its core, it is about reading the instructions inside a tumour and then writing a new set of instructions to help the immune system see it.

Rosie is an eight-year-old rescue Staffordshire bull terrier cross that developed aggressive mast cell cancer, a common skin cancer in dogs. She had surgery and chemotherapy, but the disease kept coming back, and she ended up with large, ugly tumours on her leg.

Vets told her owner, Paul Conyngham, that she probably had only months to live. Instead of accepting that, he decided to use the tools he knew from his day job in tech – data analysis, AI and coding – and apply them to his dog’s cancer.

Decoding the tumour

The first step was to understand what made Rosie’s tumour different from her healthy cells.

Every cell in the body carries DNA – a long, chemical molecule that acts like a biological instruction manual. You can think of DNA as a very long string of letters written in a four-letter alphabet. Cancer happens when enough of those letters change, by chance or through damage, so that some cells start to grow and divide out of control.

Sequencing a tumour’s or normal cell’s DNA is essentially reading through that long string of letters and comparing it to the “normal” version to see where it has gone wrong. A lot of my own research has focused on this. Conyngham paid a university lab to sequence the DNA from Rosie’s tumour. That produced a huge file listing the mutations – the spelling mistakes in the cancer’s instruction manual – that set her tumour apart from her healthy tissues.

On their own, those files are just data. The question is what to do with them. This is where he turned to an AI chatbot. He asked it how scientists design personalised cancer vaccines and how he might go from a list of mutations to specific targets for a vaccine for Rosie.

A cancer vaccine in this context is different from the childhood vaccines we are used to. Traditional vaccines prevent infections: you give someone a harmless version or fragment of a virus or bacterium so their immune system can “learn” to recognise it in advance. A cancer vaccine, by contrast, is usually therapeutic rather than preventive. It is given to someone who already has cancer, with the aim of training their immune system to spot markers on the cancer cells that it has previously ignored and then attack them.

This is where mRNA comes in. If DNA is the master instruction book, mRNA (messenger RNA) is more like a photocopied page that gets sent to the cell’s protein-making machinery – think of it as a short piece of code that carries a single command: “make this protein”.

Some of the COVID vaccines use mRNA: they deliver a strand of mRNA that tells our cells to make the spike protein from the coronavirus, so the immune system can practise on it. The body then breaks down the mRNA; it does not change your DNA.

For a personalised cancer vaccine, scientists choose small parts of proteins that are unique to a particular tumour – so-called neoantigens – and encode them in an mRNA sequence.

When this mRNA is injected, cells take it up and briefly make those tumour-linked protein fragments. The immune system can then see these fragments and, ideally, begin to treat any cell displaying them as abnormal and dangerous. In effect, it is using mRNA to give the immune system a “most wanted” poster for that individual cancer.

With help from AI tools, Conyngham sifted through Rosie’s tumour mutations to pick out candidates that might make good neoantigens. He also used protein structure prediction software to model how some of these mutated proteins would look, trying to guess which ones would be visible to her immune system.

Crucially, he did not manufacture a vaccine in his garage. Once he had a shortlist of targets, he approached researchers at the University of New South Wales, including experts in RNA technology, who reviewed the data and designed an mRNA construct based on it. Their team turned this digital design into a physical mRNA vaccine in the lab.

It was a one-off product, made just for Rosie, encoding several of the mutations in her tumour. She then received this experimental vaccine at a veterinary research centre, with booster doses over the following months.

A lot of people have been asking if this can be done for their dogs and for people. I’m speaking with everyone involved to see what is possible here.

If you would like to be involved, please complete the following Google form:https://t.co/qs9WwDNgBH pic.twitter.com/ANvxF9LX47

— Paul S. Conyngham (@paul_conyngham) March 14, 2026

Reports from her vets and owner suggest that several tumours shrank markedly, her overall tumour burden fell, and her energy and behaviour improved. One resistant tumour has prompted a second round of analysis and a follow-on vaccine targeting a different set of mutations.

Promising, but not a cure

It should be noted that this is a single dog, not a controlled study, and mast cell tumours can behave unpredictably. We cannot be sure how much of Rosie’s improvement is due to the vaccine, how long it will last, or whether the same approach would help other dogs, let alone humans.

The AI did not “cure cancer” by itself. It acted as an always-available guide and assistant, but qualified scientists still had to check its work and do the hard parts in the lab.

Even so, this case is a vivid example of several ideas coming together. DNA sequencing allows you to read the specific mutations in an individual cancer. mRNA technology lets you quickly write a custom set of instructions to show those mutations to the immune system.

AI systems make the complex biology more navigable for non-experts, suggesting possible targets and explaining concepts – though their outputs still require expert scrutiny. Put those together, and something that would once have required a major pharmaceutical programme – a bespoke cancer vaccine – can now be attempted, at least experimentally, for a single animal.

For the informed public, perhaps the most important point is not that AI has magically solved cancer, but that the basic ingredients of high-end personalised medicine are becoming more accessible. A motivated dog owner can now order tumour DNA sequencing, ask an AI to help interpret it, and partner with an academic lab to turn that interpretation into an mRNA vaccine.

A significant scientific and ethical challenge ahead is to develop methods for testing such approaches properly, protect patients and animals from false hope and unsafe experiments, and determine who should have access if they prove to be effective.

• Justin Stebbing is a Professor of Biomedical Sciences, Anglia Ruskin University
• This article first appeared in The Conversation

The post A man used AI to help make a cancer vaccine for his dog – an oncologist urges caution appeared first on Stuff South Africa.

Over the past decade, scientists have come up with a variety of frameworks to guide their search for life in the universe. While it’s most convenient to start looking for life using the knowledge that biologists have about life on Earth, scientists have also begun integrating broader conceptions of life, including life that perhaps evolved in different chemical environments.

To expand on the idea that life out in space might look nothing like life on Earth, here are five articles The Conversation U.S. rounded up from our archives, and written by astronomers and astrobiologists.

1. Why base the search on life on Earth?

Astronomers participating in the Search for Extraterrestrial Intelligence typically start by identifying potentially habitable planets. And to do that, they look for what sustains life on Earth: water.

Planets that are close enough to their Sun that liquid water wouldn’t freeze, but far enough away that it wouldn’t evaporate, fall into what’s called the Goldilocks Zone. But why base the search on water, which complex life on Earth uses to survive, if an extraterrestrial life-form might use different chemistry?

Cole Mathis, a physicist and astrobiologist at Arizona State University who studies complex adaptive systems, explained that out of convenience, astronomers start by looking for signals similar to those produced by life on Earth.

Detecting chemical signatures using the instruments on telescopes is tricky – it’s like playing hide-and-seek, but you’re outside the house and can only peer in through the window. You might as well start by ruling out the easy and more obvious hiding spots.

Missions to Mars have looked for signs of photosynthesis – the process by which plants take in energy – and telescopes peering deep into space look for oxygen, which organisms on Earth release into the atmosphere.

“Most astronomers and astrobiologists know that if we only look for life that’s like Earth life, we might miss the signs of aliens that are really different,” Mathis wrote. “But honestly, we’ve never detected aliens before, so it’s hard to know where to start. When you don’t know what to do, starting somewhere is usually better than nowhere.”

2. Finding patterns of purpose

Sometimes, scientists find chemical ingredients that make up life on Earth out in space, but they can’t assume that these ingredients on their own indicate life. Geological and environmental processes on planets may produce these chemical signatures without any living organisms involved.

The key difference, to Amirali Aghazadeh, a computational scientist at the Georgia Institute of Technology, is purpose. Life grows, adapts and changes over time to better fit its environment.

His research team came up with a framework that, instead of looking for a specific type of life-form, looks at patterns in collections of chemicals and evaluates whether they could have been produced by processes like metabolism and evolution.

“If we assume that alien life uses the same chemistry, we risk missing biology that is similar – but not identical – to our own, or misidentifying nonliving chemistry as a sign of life,” wrote Aghazadeh.

3. Lessons from complex, evolving systems

Like Aghazadeh, many astrobiologists are starting to look more broadly at how complexity emerges, rather than searching for a specific type of molecule that could indicate the presence of extraterrestrial life. Other forms of life may be made up of entirely different chemical ingredients to humans, but to be considered life, they would still have to adapt and evolve over time.

Chris Impey, an astronomer from the University of Arizona, attended a workshop where scientists across disciplines came together to try to understand how and why systems in the universe – from organisms to languages and information – change or grow more complex over time.

Figuring out these underlying drivers of complexity, or finding signals that indicate the presence of a complex system, could help scientists search for unique forms of life in the universe.

“As astrobiologists try to detect life off Earth, they’ll need to be creative,” Impey wrote. “One strategy is to measure mineral signatures on the rocky surfaces of exoplanets, since mineral diversity tracks terrestrial biological evolution. As life evolved on Earth, it used and created minerals for exoskeletons and habitats.”

4. Beyond biology: Looking for ‘technosignatures’

Another option for searching for life has nothing to do with biology. Some scientists, wrote astronomers Macy Huston and Jason Wright from Penn State University, look for “technosignatures:”: signals that would come from technology originating beyond Earth.

Human technology – from TV towers to satellite and spacecraft communications – emits enough radio waves to create faint but detectable signals travelling through space. Scientists use this idea to search for artificial signals that could potentially come from an extraterrestrial civilisation.

Other technosignatures could include chemical pollution, artificial heat or light from industry, or signals from a large number of satellites.

“While many astronomers have thought a lot about what might make for a good signal, ultimately, nobody knows what extraterrestrial technology might look like and what signals are out there in the universe,” wrote Huston and Wright.

5. Evaluating extraordinary claims

Detecting extraterrestrial life in any form would be a momentous occasion, so, as Impey wrote, making a declaration might not be cut-and-dried. In “Project Hail Mary,” the fictional scientists sample and study the “space dots” they find extensively before drawing a conclusion.

Scientists must first rule out any possible non-biological explanations for a discovery, meaning the discovery would have to be unexplained by any chemical or geological processes. If scientists ever found a potential life-form very different from all life on Earth, it might take extensive research before they could rule out all other possibilities and determine that it’s a living organism. But setting this bar so high protects scientists from making a claim they would later need to walk back.

“A detection of life would be a remarkable development,” Impey wrote. “On scales large and small, astronomers try to set a high bar of evidence before claiming a discovery.”

This story is a roundup of articles from The Conversation’s archives.

• Mary Magnuson is an Associate Science Editor

Interviewed:

• Amirali Aghazadeh is an Assistant Professor of Electrical and Computer Engineering, Georgia Institute of Technology
• Chris Impey is a University Distinguished Professor of Astronomy, University of Arizona
• Cole Mathis is an Assistant Professor of Complex Adaptive Systems, Arizona State University
• Jason Wright is a Professor of Astronomy and Astrophysics, Penn State
• Macy Huston is a Ph.D. Candidate in Astronomy and Astrophysics, Penn State
• This article first appeared in The Conversation

The post ‘Project Hail Mary’ explores unique forms of life in space – 5 essential reads on searching for aliens that look nothing like life on Earth appeared first on Stuff South Africa.

While these types of studies provide interesting snapshots of reported practices, their methodologies may hide something important: how writing actually unfolds while students are composing with the assistance of AI.

A pilot study I led of undergraduate writers at Kennesaw State University takes a different approach. Using think-aloud protocols – a method in which participants verbalise their thoughts while performing – our research captures how students interact with generative AI tools during the writing process. This method helps us understand decision-making processes as they occur.

Our preliminary findings suggest a more complex reality than the common narrative that students are simply having AI write their assignments. Instead, many students appear to be negotiating when and how AI belongs in their writing.

Looking inside the writing process

In our study, 20 undergraduate students completed a 20-minute writing session responding to the following prompt:

People spend a lot of time trying to achieve perfection in their personal or professional lives. People often demand perfection from others, creating expectations that may be challenging to live up to. In contrast, some people think perfection is not attainable or desirable.

The assignment was to draft a thesis and evidence-based paragraphs that argue their position on the value of striving for perfection. Students were told they were not expected to complete them but to work through their writing process toward finishing. Students were told there were no right or wrong ways to use AI and were asked to use generative AI exactly as they normally would while writing.

Instead of direct observation, the study relied on post-session screen recordings and analysis of students describing their process. Collecting this data – their actions on the computer and transcripts of the voice recordings – allowed researchers to analyse the writing process without interrupting it. To reduce the possibility that students might alter their behaviour if they felt observed, researchers set a timer and left the room during the writing session. The goal was to minimise the Hawthorne Effect, a phenomenon in which people change their behaviour because they know they are being watched.

What we found

Across the transcripts, a few qualitative patterns consistently emerged in how students collaborated with AI while writing.

First, many participants turned to AI at the beginning of the writing process to help generate ideas or draft a thesis.

What we see in this practice is the student using AI-generated output to spark and shape their own ideas. One student explained the strategy this way: “After [generating a few ideas,] I usually just use that [output] as a prompt.”

In these moments, AI functioned less as a final answer and more as a brainstorming tool that helped students move past the blank page.

However, students frequently continued drafting independently after generating initial ideas. Many transcripts include statements such as “I think my thesis should be …” or “Let me write this part,” suggesting that some students retained control over their argument.

Editing the bot

Another strong pattern across transcripts is that students rarely accept AI text without editing it. Instead, they actively revise the generated language. As one student described the process, the AI “rewrites” their initial prompts and then the student rewrites the AI’s output. This allows the student to claim “authorship and ownership” of the final draft.

Another participant redirected the AI response when it did not align with the assignment: “AI is not following the prompt … try again.”

These moments show students evaluating AI output critically and treating it almost as a sparring partner, rather than simply copying it.

We also found that some students rejected AI’s suggestions altogether.

In several writing sessions, participants explicitly decided not to use the AI responses. One student reflected on this decision while composing: “I don’t really use AI for my research.”

Other transcripts show students switching back to their own writing when AI responses felt too generic or disconnected from their argument. These moments indicate that students are not only collaborating with AI, but they are also drawing boundaries around where it belongs in their writing process.

Finally, several transcripts showed students turning to AI during moments of uncertainty or when they felt stuck.

As one participant explained, “I used a lot of AI because I was struggling.”

Even in those cases, students often used AI as support as they drafted their essays, rather than directly copying and pasting its responses.

What this says about AI and writing

Our analysis suggests that generative AI is entering student writing not as a wholesale replacement for human authorship, but as part of a negotiated collaboration. The results suggest that AI most often enters the composing process during idea generation, revision and moments of writer’s block, while students maintain control over argument choice, voice and final phrasing.

Understanding how decisions to use AI unfold during the writing process, and not just what appears in the final essay, may help educators design assignments and policies that keep the human writer firmly at the helm.

Because our current findings come from a pilot cohort of 20 undergraduate writers, the results should be interpreted cautiously. To test whether these patterns hold at a larger scale, the research team is expanding the study to 100 undergraduate participants. The expanded study will also examine how neurodivergent writers interact with generative AI during composing, an area that remains largely unexplored in current research.

Undergraduate student researchers at Kennesaw State contributed to the preliminary analysis described in this article: Kylee Johnson, Vara Nath, Ruth Sikhamani and Kaylee Ward.

• Jeanne Beatrix Law is a Professor of English, Kennesaw State University
• This article first appeared in The Conversation

The post University students are writing with AI – but a pilot study finds they’re not simply letting it write for them appeared first on Stuff South Africa.