Golden eagles in England? Here’s the ecological case for bringing them back

Source: The Conversation – UK – By Esther Kettel, Senior Lecturer in Ecology and Conservation, Nottingham Trent University

England’s last recorded pair of golden eagles lived in the Lake District. After the female died in 2004, the male was left alone for 12 years before his death in 2016.

This marked the end of golden eagles across English skies. Though they have lived on in Scotland, the birds were largely wiped out across England about 150 years ago, with only a few nesting attempts during that time.

Annotated map of England
The eight ‘recovery zones’ are shaded. Sites where golden eagles were once found are marked with stars.
Forestry England /, CC BY-SA

However, the UK government recently announced it will support reintroducing the species and has identified eight potential “recovery zones” across northern England and the south-west. This is good news for lots of reasons.

Reintroducing lost species aligns with the government’s 25-year environment plan to restore and recover nature. Golden eagles also have an important heritage as symbols of wilderness, freedom and power. We may even have a moral duty to return them to the landscape, since humans were largely responsible for their loss.

Reintroducing golden eagles would also benefit England’s natural environment, helping return it to a healthier and more dynamic state.

Restoring balance to the food chain

Golden eagles are apex predators, occupying the top of the food chain with no natural predators. The removal of a species like this can cause major shifts in ecosystems, as they exert top-down control.

When apex predators are missing from ecosystems, the middle predators of food chains – or “meso-predators” – become dominant. With its native bears, lynx and wolves long gone, England has a high number of meso-predators. These include badgers, red foxes and other birds of prey. These predators, in turn, can limit some populations of prey like seabirds, waders and gamebirds.

Buzzard in England moorland
In England, buzzards often sit at the top of the food web. Elsewhere in the world, the have to be wary of bigger birds.
Serenity Images23 / shutterstock

Meso-predators typically avoid areas where apex predators are due to fear of competition or being eaten themselves. So, if golden eagles return then the predation pressure from smaller birds might be altered. For example, on the Isle of Mull in Scotland, meso-predators like kestrels and buzzards tend to steer clear of areas where golden eagles are.

Controlling prey numbers

Golden eagles also have an important role in the ecosystem by regulating their prey species. They hunt various prey, mostly medium-sized birds and mammals like rabbits, hares and occasionally, young deer.

When not controlled by predators, prey populations can boom. This can lead to greater competition for resources and a higher risk of disease spread among these prey species. Prey populations may also overuse resources, which can negatively affect plant growth.

Because apex predators are absent in England, humans must take up the role of controllers. Deer are shot where they are preventing woodland regeneration and rabbits are widely controlled in agricultural landscapes, costing £5 million a year. Although golden eagles are unlikely to reduce deer and rabbit numbers substantially, they may bring some balance back.

Keeping the environment clean

In addition to being excellent predators, golden eagles also scavenge carcasses – the remains of dead animals. Researchers in Spain found that 90% of the golden eagles in their study fed on carcasses.

Carcasses can quickly become disease and toxin reservoirs that may enter the wider environment if left uneaten. This can have consequences for other species, including humans. So scavengers have a crucial role in maintaining a healthy ecosystem.

If reintroduced back to England, golden eagles would join the cleaning crew, which also includes species like red kites, crows and red foxes.

Indicators of a healthy ecosystem

If a pollutant is in an environment, this could affect top predators through a process called biomagnification, where the concentration of the pollutant increases the further up the food chain. If in high concentrations, the pollutant may become toxic and the predator may fail to reproduce, become unwell, or die.

In the 1960s, birds of prey played a pivotal role in making the environmental dangers of certain agricultural pesticides clear in the UK and globally, leading to the widescale ban. Golden eagles could do something similar today.

A complex picture

If golden eagles are successfully reintroduced in England, they could restore balance to food chains, control prey numbers, scavenge carcasses and act as indicators of environmental dangers.

They will join other birds of prey that have been successfully reintroduced to England, such as red kites, ospreys and white-tailed eagles, all of which have been deemed a success.

However, ecological systems are not straightforward and predicting the consequences of the return of golden eagles is complex. As indicated by the risk assessment conducted by Forestry England, at worst the impacts on biodiversity of golden eagles will be neutral. At best it will be beneficial.

The Conversation

Esther Kettel does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

ref. Golden eagles in England? Here’s the ecological case for bringing them back – https://theconversation.com/golden-eagles-in-england-heres-the-ecological-case-for-bringing-them-back-281040

Eating fruit is linked to lung cancer? Here’s what you need to know about that new study

Source: The Conversation – UK – By Justin Stebbing, Professor of Biomedical Sciences, Anglia Ruskin University

Deadly? Not likely. New Africa/Shutterstock.com

The idea that fruit and vegetables might cause cancer sounds bizarre. For decades, studies have shown that people who eat more plants tend to live longer, healthier lives, with lower rates of heart disease, stroke and several common cancers.

Lung cancer is no exception: in many large studies, higher intakes of fruits and vegetables are linked with lower risks, especially in smokers.

Against that backdrop, a new suggestion that fruit and veg might be driving lung cancer in young adults is surprising.

The story behind this latest wave of anxiety doesn’t come from a definitive, landmark trial. It comes from a brief presentation at a scientific conference, based on 187 people with early‑onset lung cancer.

Most had never smoked. When researchers asked about their diets, a lot of them reported eating plenty of fruits, vegetables and whole grains – the sort of pattern most of us would call “healthy”.

Instead of measuring pesticides in their food or blood, the team estimated probable pesticide exposure using average residue levels from other sources. From there, they speculated that pesticides on otherwise healthy foods might help explain why some young non‑smokers develop lung cancer.

That is a very long way from proving that fruit and vegetables themselves are harmful. Studies like this are meant to raise questions – “could pesticides be part of the story in young lung cancer?” – not to rewrite dietary advice on their own.

Crucially, this particular study looks backwards from people who already have cancer, rather than following healthy people forwards over time, so it cannot tell us whether their diet played any role in causing the disease. Nor does it show that these patients had higher pesticide exposures than comparable people without cancer. It only shows that they ate foods that, on average, can carry residues.

The bigger picture

When you zoom out from this single, tiny study to the broader body of evidence, the picture changes from alarming to reassuringly familiar. Large studies have followed tens or hundreds of thousands of people over many years, asked them what they ate, then waited to see who develops lung cancer. Time and again, those eating more fruit and vegetables either do better or, at very worst, no differently from those eating less.

Meta‑analyses that combine data from multiple studies find reductions in lung cancer risk with higher fruit intake and benefits from vegetables, too. These are the studies that inform official guidelines. They are not perfect – no nutrition study is – but they are far more informative than a single unpublished study of 187 patients.

So why do small studies like this latest one sometimes seem to say something different? One reason is simple statistical noise.

With small numbers, chance plays a huge role. If, for whatever reason, the particular group of young adults who turned up to that clinic happened to be unusually health-conscious, then fruit and vegetable intake will look high among people with lung cancer, even if diet has nothing to do with their disease.

Another issue is what scientists call “confounding”. People who eat more plants often differ in many other ways. They may exercise more, drink less, have different jobs, live in different neighbourhoods, or be more on the ball about seeking medical help.

When you start from patients and look backwards, it is very hard to disentangle these overlapping factors. That is why we place more weight on large, prospective studies that follow people forward in time and can better account for these differences.

Pesticides

Then there is the question of pesticides – the part of the story that understandably unnerves people. It is true that many conventionally grown fruits and vegetables carry measurable pesticide residues, and that people who eat a lot of produce tend to have higher levels of some pesticide breakdown products in their urine.

It is also true that farm workers who handle pesticides regularly and at high doses have higher rates of certain cancers, including some lung cancers. That tells us pesticides are not benign. But what it does not tell us is that eating sprayed apples or lettuce at normal dietary levels causes lung cancer in the general population.

A farm worker spraying pesticides on a crop.
Farm workers who are exposed to high doses of pesticides do have higher rates of certain cancers.
Kuro1982/Shutterstock.com

That doesn’t mean we should be complacent: there is an ongoing discussion about cocktails of many different chemicals, about vulnerable groups such as children and pregnant women, and about longer‑term hormone or brain effects that might not show up in crude cancer rates. However, these are arguments for improving how we farm and regulate pesticides, not arguments for abandoning fruit and vegetables.

If you are still uneasy about pesticides, there are practical, proportionate things you can do that don’t involve swapping an orange for a packet of crisps. Washing produce under running water helps remove surface residues and soil, and varying the types of fruit and veg you eat means you are not relying heavily on any one item that tends to carry higher residues.

If your budget allows, choosing organic versions of a few “high‑residue” foods can make sense. But the key point is that these are tweaks at the margins. They don’t change the central message that a diet rich in plant foods is overwhelmingly associated with better health.

Perhaps the most important lesson from this episode is about how to read nutrition headlines. Whenever you see “X food causes cancer” or “Y ingredient is the next miracle cure”, it helps to ask a couple of simple questions. How big was the study? Was it in healthy people followed over time, or patients looked at after the fact? Did the researchers actually measure what they are claiming (like pesticide levels)? And how do the new findings sit alongside decades of existing research?

In the case of the early-onset lung cancer study, the answers are sobering: it was small, it was retrospective, it used indirect exposure estimates, and its suggestion that fruit and vegetables might be harmful sits awkwardly with a much larger body of work pointing the other way.

None of this means we should ignore the possibility that pesticides contribute in some way to cancers in non‑smokers, or that diet is irrelevant to lung health. But we should be wary of turning one provocative conference talk into a reason to fear the very foods that consistently show up as markers of better health.

The Conversation

Justin Stebbing does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

ref. Eating fruit is linked to lung cancer? Here’s what you need to know about that new study – https://theconversation.com/eating-fruit-is-linked-to-lung-cancer-heres-what-you-need-to-know-about-that-new-study-281003

Israel’s onslaught against Lebanon may strengthen Hezbollah – just when it’s at its weakest

Source: The Conversation – UK – By John Nagle, Professor in Sociology, Queen’s University Belfast

As the tentative ceasefire in Lebanon holds, people are returning to their homes in the south to find widespread destruction. Whole villages laid waste, roads and bridges ruined, hospitals and other civic infrastructure flattened. And the Israeli army still very much in evidence in many areas.

The most recent conflict between Israel and Lebanon has killed more than 2,100 people and displaced more than a million more. Israel’s stated aim is to destroy Hezbollah, which it describes as an Iranian proxy. But this is a misleading framing of the situation. And trying to destroy Hezbollah by attacking and occupying Lebanon is a dangerous misreading of the situation.

Hezbollah, the so-called “Party of God”, is not the same thing as Lebanon. Yet the party is deeply embedded in Lebanese politics. The group emerged during the Lebanese civil war and in the aftermath of Israel’s 1982 invasion. It grew rapidly by combining armed resistance with political representation and services for Shia communities that had long been neglected by the Lebanese state.

In the southern suburbs of Beirut, known as Dahiyeh, and across the south, it became a provider of services. Hezbollah built schools, clinics and welfare networks that helped it convert resistance into social legitimacy. That presence built loyalty and dependence that outlasted its original resistance role.

Lebanon’s postwar political system is built on sectarian power sharing. Hezbollah entered parliament in the 1990s and built alliances well beyond its core Shia base, which enabled it to join coalition governments.

But unlike other major Lebanese factions, it retained its weapons after the civil war. This allowed it to combine formal political participation with an armed capacity that was outside the control of the state. Its alliance with Christian groups, most significantly Free Patriotic Movement, Lebanon’s largest Christian party, gave it cross-sectarian legitimacy and protection against isolation.

Hezbollah’s ability to shape Lebanese politics has often rested less on governing than on stopping other groups from governing. The clearest illustration was the presidency. After Michel Aoun completed his term in October 2022, Lebanon went without a president for more than two years. Hezbollah blocked every candidate that threatened its interests. Parliament failed to elect a successor 13 times.

Lebanon drifted without a head of state through the 2024 war with Israel. Its caretaker government could not take major decisions. Desperately needed economic assistance was withheld by international donors. It was Hezbollah’s blocking power made visible. Lebanon’s caretaker government could not take major decisions or enact the reforms international donors required. Desperately needed economic assistance was withheld as a result.

Hezbollah’s political weakness

This current conflict has caught Hezbollah in a weaker political position than it once enjoyed. The anti-government protests of 2019, economic collapse and the Beirut port explosion has deepened public anger at Lebanon’s ruling class — and at Hezbollah as part of it. Hezbollah’s attempts to obstruct the judicial investigation into the explosion deepened that anger further.

The 2022 elections confirmed the shift. Hezbollah and its allies lost the majority they had held since 2018. Independents and reformists who emerged out of the protests took seats in a more fragmented legislature.

The Arab Barometer’s 2024 survey found that just 30% of Lebanese expressed significant trust in Hezbollah, with 55% saying they had no trust at all. Hezbollah’s claim to speak for Lebanon — or even for all Lebanese Shia — is now more contested than at any point in its modern history.

The 2024 war, with the devastating pager attacks of September 17 and 18, substantially degraded Hezbollah’s military and further weakened its political standing. Assad’s fall in Syria in December removed a key source of regional support.

In January 2025, the Lebanese parliament finally elected Joseph Aoun as president — something that would have been unthinkable when Hezbollah was at its peak and was able to use its influence to exclude him. Aoun, a former army commander, has always insisted it was the army – not Hezbollah – that should be the defender of Lebanon’s sovereignty.

Operation: destroy Hezbollah

Israel’s stated objective for many years has been to create a more durable security order along its northern border by weakening or dismantling Hezbollah. But, at the same time, Israeli strikes have inflicted devastation far beyond Hezbollah itself, hitting civilians, infrastructure and communities across the country.

The destruction of places such as Dahiyeh reflects a broader logic of warfare in which dense urban space is treated as part of the battlefield. UN experts have argued that the destruction of homes and mass displacement amount to collective punishment in violation of international law.

The argument echoes broader legal debates about Israel’s conduct in Gaza, where UN experts have made similar findings.

That is also why the simple frame of “Israel versus Hezbollah” erases so much. Many of those driven from their homes in the south or in Dahiyeh had grown critical of Hezbollah, or had not chosen this war at all. Yet they found themselves bombed out of neighbourhoods that had been designated as legitimate targets, because of an assumed association with Hezbollah. The civilians killed and displaced are not bystanders to somebody else’s conflict. They are among its principal victims.

A ceasefire was announced on April 17, and – while Hezbollah has not formally endorsed it – the group appears to be observing it for now. Yet the truce leaves the central political question unresolved. Israeli officials have made clear they do not regard it as settling the question of southern Lebanon’s demilitarisation.

Expecting the Lebanese army to dismantle Hezbollah by force is unrealistic. If Hezbollah resisted — and it would — the result could be open civil conflict. It would fracture the army, deepen sectarian tensions, and drive Shia communities back behind the very organisation whose grip had begun to loosen, leaving it politically stronger than it was before the latest round of hostilities.

Any lasting settlement will have to reckon with the reality this war has exposed: Hezbollah is not Lebanon. But at the moment it’s Lebanon which is being punished.

The Conversation

John Nagle does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

ref. Israel’s onslaught against Lebanon may strengthen Hezbollah – just when it’s at its weakest – https://theconversation.com/israels-onslaught-against-lebanon-may-strengthen-hezbollah-just-when-its-at-its-weakest-281031

We designed the turf for soccer’s biggest World Cup ever – here’s how we created the same playing experience across 3 countries

Source: The Conversation – USA (2) – By John N. Trey Rogers, Professor of Turfgrass Research, Michigan State University

World Cup pitches take a beating. AP Photo/Bernat Armangue

With 104 matches in 16 stadiums across Canada, the United States and Mexico, the 2026 FIFA World Cup will be soccer’s biggest event ever.

It’s our job as turfgrass researchers hired by FIFA, the game’s governing body, to make sure those pitches feel the same for players and that the grass thrives.

That’s not so simple. In fact, it seemed like an impossible challenge at first.

Picking the right turf

The scale of this job was unprecedented: three distinct climatic zones, over 3,100 miles between the farthest stadiums, and venues ranging from stadiums open to the heat of Mexico City and Miami to enclosed NFL stadiums in Dallas and Atlanta, to the cooler climates of Boston and Toronto.

Despite the unique situations of each stadium, FIFA has a long list of rules for how the fields must be built. The grass has to be real but reinforced so it can handle a lot of games and ceremonies. Each field needs an automatic irrigation system, good drainage, built-in vacuum and vents to keep the grass and soil aerated, and artificial grow lights to keep the grass healthy.

Each host city is responsible for figuring out how to meet these requirements.

Right now, eight of the 2026 host stadiums normally use artificial turf – how will they temporarily switch to real grass for the World Cup?

Even trickier, five of the stadiums have domes, which means the grass gets less sunlight. How can they keep the grass alive for eight weeks?

How can we make sure that a player competing in Philadelphia has the same on‑field experience as a player competing in Guadalajara or Seattle?

The new turfgrass goes down in New England’s Gillette Stadium near Boston. WCBV.

Our team at the University of Tennessee and Michigan State University has spent the past five years researching these questions to provide guidance to the host cities. Here, we’ll explore some of the most important questions we faced: which grass to grow, how it’s grown, how we plan to make it even stronger, and how to move it safely to each stadium.

Growing the grass

Typically, sod is grown on native soil. When harvested, the roots are cut, which shocks the plant and can delay root reestablishment for several weeks.

That wouldn’t work for the World Cup because games may take place within just 10 days of installation. If the roots can’t become established fast enough, the grass will be weaker and more prone to damage.

To address this, we decided to use sod grown on plastic with sand as a base.

Think of it like growing grass in a plastic tray, but on a much larger scale. When the roots reach the plastic, they spread sideways and intertwine, forming a dense rooting system. Because the roots stay intact during harvest, the sod experiences minimal stress and can be ready to play almost immediately after installation.

Sod for sports fields is typically grown in a base of sand to provide quick drainage and prevent the grass from getting compacted as the roots become established.

The problem is that growing grass in 2 inches of sand on a plastic sheet comes with risks. Because of the plastic, a single heavy rainfall while the grass is becoming established can wash the exposed sand away.

For warm‑season sod farmers – those that grow grass that thrives in high temperatures – sand washing away is less of a concern because the Bermudagrass they grow establishes quickly. On the other hand, cool‑season sod farmers usually grow Kentucky bluegrass, which germinates slowly compared to other turfgrass species, increasing the risk of washouts.

We decided to mix a faster‑germinating species – perennial ryegrass – with Kentucky bluegrass grown on plastic and then tested various seeding ratios. We found that an 84% Kentucky bluegrass and 16% perennial ryegrass mixture produced a stronger sod than pure Kentucky bluegrass alone four months after seeding. Since 2025, these findings have been used on sod farms across North America, beyond those growing grass for the World Cup.

Stabilizing the surface

“One World Cup game is equal to a Super Bowl,” FIFA officials like to remind us. Since each field will host a lot of games and ceremonies, including up to nine games over six weeks, the fields need to be extremely strong.

To make them tougher, we mix plastic fibers into the natural grass, which creates a hybrid turfgrass system. As the grass grows, its roots wrap around these plastic fibers, which helps to keep the surface stable and firm. These fibers are also colored to match the natural grass, so even if the real grass wears down, they help the field stay green.

Hybrid turfgrass systems can be created in two ways: by stitching plastic fibers into an existing grass field or by laying down a carpet of plastic fibers that is then filled with sand and seeded to grow new grass.

Stitched systems have been used in World Cup games for a long time, but carpet systems are still fairly new to the tournament – they have been used only in the 2023 Women’s World Cup.

We tested eight carpet systems to see how they performed and found that all could be successfully grown on plastic. All the surface performance tests – ball bounce, rotational resistance and surface hardness – on these eight carpets also met FIFA standards.

One type of carpet was chosen by three host cities for their stadiums: Vancouver, Los Angeles, and Philadelphia.

Getting the sod from farm to stadium

Most of the stadiums – 14 of them – will have sod that is grown on plastic, then rolled up and shipped to the venue during spring 2026. Some of the grasses won’t have to travel far, but some will be shipped in refrigerated trucks across the country. Since the sod remains fully intact after harvest, it can withstand long travel times.

Five of those stadiums don’t get enough sunlight, so they will use cool-season grasses that require less light than warm-season grasses.

While the open-air stadium in Miami will use Bermudagrass, the domed stadium in Houston, despite being at a similar latitude, will use the Kentucky bluegrass and perennial ryegrass mix. That means cross-country trips from cool-season sod farms in Denver and Washington to domed stadiums in the southern regions is essential.

It’s wild to think that this is all necessary, but the length of the tournament and unique stadium environments call for innovation.

The Conversation

The University of Tennessee was the prime awardee of the FIFA grant. Michigan State University is a subawardee. John N. Trey Rogers is the principal investigator for the Michigan State work.

Jackie Lyn A. Guevara is affiliated with Michigan State University. She received compensation through a FIFA grant awarded to Michigan State University.

John Sorochan is the principal investigator for the FIFA grant at the University of Tennessee.

Ryan Bearss works for Michigan State University

ref. We designed the turf for soccer’s biggest World Cup ever – here’s how we created the same playing experience across 3 countries – https://theconversation.com/we-designed-the-turf-for-soccers-biggest-world-cup-ever-heres-how-we-created-the-same-playing-experience-across-3-countries-279536

Cities helping cities rebuild: How local partnerships are shaping Ukraine’s recovery

Source: The Conversation – Canada – By Tamara Krawchenko, Associate Professor, School of Public Administration, University of Victoria

The Congress of Local and Regional Authorities of the Council of Europe recently called for local and national authorities to work together to help Ukraine recover and rebuild four years after Russia’s full-scale invasion of the country.

The message is clear: cities and regions must lead, and their counterparts around the world should help them do it. The congress also calls on Russia to pay for the damage it has caused, pointing to frozen Russian assets worldwide as one source for those funds — an acknowledgment that recovery cannot wait for the war to end, since communities are already rebuilding under fire.

As a lead author of the report underpinning the congress’s call to action, I want to explain why it matters and why Canada, in particular, has both the track record and the responsibility to step up.

The scale of what needs to be done

The numbers are almost impossible to absorb. By the end of 2024, direct damage to Ukraine’s buildings and infrastructure had reached approximately US$176 billion, with more than 2.5 million households destroyed or damaged.

Total aggregate economic losses from the invasion are estimated at more than US$1.1 trillion. Nearly 6.8 million Ukrainians have also sought refuge abroad, the largest displacement of people in Europe since the Second World War.

Behind every statistic is a community struggling to survive — a mayor trying to keep schools open under missile attacks; a municipal council managing hundreds of thousands of internally displaced persons with dwindling resources; a city engineer repairing the same water system for the third time after it was bombed yet again.

Local and regional authorities across Ukraine face these situations every day. And it is precisely because the challenges are so local — tied to specific communities and capacities — that the response must also be local.

Ukrainian decentralization reforms since 2014 have expanded the fiscal capacity of the country’s municipalities, enabling them to respond to the unprecedented shocks of war far more effectively than before. In fact, local budget revenues quintupled between 2014 and 2021.

Russia’s invasion disrupted these reforms.

Why city-to-city networks are different

The call to action by the congress asks local and regional authorities in Council of Europe member states to use “existing co-operation platforms and bilateral partnerships to offer practical support to their Ukrainian counterparts.”

It’s an appeal for cities that have solved difficult problems — managing mass displacement, rebuilding after disaster, reforming service delivery — to share what they know with Ukrainian cities doing the same under fire.

City-to-city partnerships are fundamentally different from top-down aid. They are peer relationships built on what scholars call horizontal assistance — the exchange of practical knowledge and structural social capital between cities navigating similar challenges.

Research on municipal technical exchanges, including a study of Seattle’s city-to-city delegations, shows these networks generate direct benefits: lower costs of accessing policy information, facilitation of collective action and long-term institutional ties that outlast any individual project cycle.

As the researchers who conducted the Seattle research point out, the exchanges “disseminate information, and through the personal relations they initiate, have a potential for influencing future resource decisions among cities and countries.”

This matters enormously for Ukraine. The most effective city networks are those oriented toward concrete policy transfer such as sharing regulatory frameworks, governance tools and public administration practices.

Ukrainian cities need exactly this: working models of how to manage housing allocation for displaced persons, how to deliver trauma-informed social services and how to rebuild energy infrastructure with built-in resilience.

Canadian municipal engineers who might advise a Kharkiv counterpart on water system resilience wouldn’t just be delivering aid — they’d be sharing hard-won professional knowledge among equals. That knowledge sticks in ways that consultant reports rarely do.

The lessons of past reconstruction

History offers clear guidance on what works. Comparative analysis of post-war and post-disaster reconstruction experiences identifies local community engagement and bottom-up leadership as the single most consistent factor separating successful from failed reconstruction.

Top-down donor interventions that bypassed local institutions, as in Iraq after 2003, produced waste, duplication and projects misaligned with community priorities. By contrast, programs that genuinely incorporated recipient input — like the post-Second World War Marshall Plan — achieved lasting results.

The review by the Organization for Economic Co-operation and Development (OECD) of Ukraine’s recovery architecture echoes this: Ukraine’s reconstruction ecosystem remains fragmented, with co-ordination gaps among federal government departments, international donors and local authorities.

City-to-city networks can help fill that gap at the most practical level by channelling directly applicable knowledge to the local officials who most need it.




Read more:
Engineering hope: how I made it my mission to help rebuild Ukraine’s critical infrastructure


Canada’s proven record, and its moment

Canada has been here before. Beginning in 2010, the Federation of Canadian Municipalities (FCM), financed by the Canadian government through Global Affairs Canada, built exactly this kind of peer network in Ukraine through the Partnership for Local Economic Development and Democratic Governance.

The $19.5-million, six-year initiative worked directly with 16 Ukrainian cities to strengthen local democracy, support small and medium-sized businesses and advance decentralization.

FCM’s municipal experts worked alongside counterparts in cities like Lviv and Dnipro, co-publishing Ukraine’s first municipal guide to local economic development and helping local governments design collaborative regional projects. A key partner throughout was the Association of Ukrainian Cities, a key municipal advocacy organization.

That program ended, but the relationships it built did not. And the decentralization reforms it supported are now widely credited — by the congress’s call to action itself, the OECD and scholars of Ukrainian resilience — with giving Ukrainian local authorities the capacity to respond as effectively as they have to the shocks of war.

The case for reinvestment

The congress explicitly notes in its call to action that decentralization reforms “have played a crucial role in Ukraine’s wartime resilience.” That is, in part, a legacy of Canada’s investment.

A new, scaled-up commitment through FCM building on existing relationships with the Association of Ukrainian Cities — drawing on FCM’s international programs expertise and connecting Canadian municipal professionals to their Ukrainian peers across the priority domains the resolution identifies (like in housing, social and mental health supports, economic recovery, emergency management, community energy and citizen participation) — would represent a return to a proven formula.

The congress’s call to action urges deeper, more focused work on local recovery and reconstruction. City-to-city partnerships stand out as one of the most cost-effective and sustainable tools: they share practical knowledge that endures, strengthen institutions over the long term and recognize Ukrainian cities as active participants in their own reconstruction.

Canada helped build local government capacity in Ukraine before the war. The Council of Europe’s Congress has now called on the world to do so again. Canada should answer that call.

The Conversation

The author served as a contributing expert to the Council of Europe Congress Report CG(2026)50-10prov, adopted at the 50th Session of the Congress of Local and Regional Authorities in March 2026.

ref. Cities helping cities rebuild: How local partnerships are shaping Ukraine’s recovery – https://theconversation.com/cities-helping-cities-rebuild-how-local-partnerships-are-shaping-ukraines-recovery-280205

Exploring conversational AI and poetry but not as we know it

Source: The Conversation – France – By Thierry Poibeau, DR CNRS, École normale supérieure (ENS) – PSL

What do the large language models behind human-like conversational AI really know and what does it mean to live alongside them?

A new book Understanding Conversational AI by Thierry Poibeau offers a critical and interdisciplinary exploration of large language models (LLMs), examining how they reshape our understanding of language, cognition and society. Drawing on philosophy of language, linguistics, cognitive science and AI ethics, it investigates how LLMs generate meaning, simulate reasoning and perform tasks that once seemed uniquely human from translation to moral judgement and literary creation.

The book explores the limitations of these models, their embedded biases and their role in processes of automation, misinformation and platform enclosure while reflecting on how they prompt us to revisit fundamental questions: What is understanding? What is creativity? How do we ascribe agency or trust in a world of synthetic language?

Recent findings are suggesting that non-expert poetry readers prefer AI works to poems written by humans. Thierry Poibeau reminds us how “aesthetic judgment in poetry often involves elements that are not easily codifiable, such as originality, emotional resonance, metaphorical depth, and cultural embeddedness”. Poibeau sheds light on why AI-generated poems can be appealing, by highlighting how literary value is “shaped by evolving norms within particular reading communities, making judgments of poetic merit historically contingent and socially negotiated”.

AI Poetry in motion

Excerpts from Understanding Conversational AI

“The arrival of AI-generated poetry raises fundamental questions about whether current evaluation criteria, often rooted in human experience, intentional expression, and historical context are adequate for assessing texts produced through large-scale statistical recombination.

As large language models generate increasingly fluent and plausible verse, their outputs may challenge existing notions of aesthetic authenticity, precisely because they blur the boundary between craft, imitation, and genuine creative insight.

These systems are capable of reproducing recognizable poetic forms, stylistic conventions, and affective tones, raising the question of how such texts should be evaluated when traditional assumptions about authorship and intention no longer clearly apply.

However, despite this formal competence, recent studies reveal recurrent stylistic tendencies characteristic of large language model–generated poetry. GPT-4, for instance, shows a strong preference for quatrain structures, a frequent reliance on iambic meter and end rhyme, and a tendency to employ recurring lexical choices such as heart, whisper, or dream.

These outputs often reflect a flattening of emotional and metaphorical complexity, favoring literal formulations and conventional poetic tropes over ambiguity, innovation, or semantic depth.

In comparison to human poetry, LLM-generated verse appears more homogeneous and less nuanced, and is less capable of producing the kinds of conceptual tension or unexpected imagery often found in more original human compositions.

In addition to these stylistic patterns, recent reader evaluation studies suggest that some AI-generated poems are judged comparably to human-authored texts when assessed blind.

However, once authorship is disclosed, evaluations tend to decline, indicating a persistent skepticism toward machine-generated poetry. This reception dynamic highlights the ongoing tension between the formal fluency of AI-generated verse and readers’ perceptions of authenticity and creative agency.

Interestingly, readers often find AI-generated poems easier to interpret. They can more readily grasp the images, themes, and emotions, which are typically presented in a more accessible and transparent manner than in the often denser and more ambiguous work of human poets.

As a result, readers may develop a preference for these texts and mistakenly interpret their own ease of understanding as evidence of human authorship. When the machine origin of the poem is revealed, this interpretation shifts, and the same textual features that previously facilitated comprehension may be reinterpreted as signs of superficiality or lack of depth.

Such findings should be interpreted with caution. They do not demonstrate that AI-generated poetry is superior in literary quality, but rather that it tends to conform to familiar forms and accessible conventions. In contrast, human-authored poetry frequently relies on layers of complexity, allusion, and ambiguity that demand substantial background knowledge in literature, history, and poetic traditions. Interpreting such works is a cognitively demanding task, and it is therefore problematic to ask lay readers to evaluate poetry in a meaningful way without accounting for differences in expertise and interpretive competence.

Beyond questions of evaluation, the reception of AI-generated poetry raises broader issues of attribution, disclosure, and creative ownership. Literary institutions, publishers, and competition organizers increasingly require authors to disclose the use of AI tools, reflecting uncertainty about how machine-generated texts should be situated within existing frameworks of creativity and value.

The discomfort that emerges when readers discover that a poem was generated by a machine is not merely a reaction to deception but an indication that aesthetic judgment remains deeply entangled with assumptions about intention, experience, and authorship.”


Thierry Poibeau is Director of Research at the Centre national de la recherche scientifique (CNRS), Paris and head of the LaTTiCE laboratory in Paris, France.

He is a member of PRAIRIE-PSAI (Paris AI Research Institute – Paris School of Artificial Intelligence) and is also an affiliated lecturer at the Department of Theoretical and Applied Linguistics (DTAL) of the University of Cambridge. He works on natural language processing (NLP), particularly focusing on information extraction, question answering, semantic zoning, knowledge acquisition from text, and named entity tagging.


A weekly e-mail in English featuring expertise from scholars and researchers. It provides an introduction to the diversity of research coming out of the continent and considers some of the key issues facing European countries. Get the newsletter!


The Conversation

Thierry Poibeau is a member of PRAIRIE-PSAI (Paris AI Research Institute — Paris School of Artificial Intelligence) and has received funding in this capacity.

ref. Exploring conversational AI and poetry but not as we know it – https://theconversation.com/exploring-conversational-ai-and-poetry-but-not-as-we-know-it-280284

Women working in Uganda’s pig sector: how challenging prejudices can unlock opportunities – research

Source: The Conversation – Africa – By Esther Leah Achandi, Post Doctoral Fellow- Gender, International Livestock Research Institute

In some communities in Uganda, women aren’t supposed to work with pigs. This stems from restrictive social and gender norms, some of which are rooted in culture and religious beliefs.

Until recently, eating pork was associated with drunkards because the meat was typically served alongside home-brewed alcohol in local bars. That’s changing, as “pork joints” become popular everyday eating places. What’s more, pigs are unfairly thought of as dirty and therefore some people think the people who work with them must be dirty too. Women, in particular, according to prevailing social norms, are meant to keep themselves clean.

The pig sector is growing rapidly in east Africa on the back of rising demand. Uganda is one of three top pork producers in Africa, after Nigeria and Malawi. The country also has the highest per capita consumption of pork in the region, estimated at 3.4 kilograms per person per year. This has led to job opportunities in pig farming, trading, butcheries, food stalls, artificial insemination, and feed and veterinary supply shops.

Across Africa, social and gender norms determine whether a woman can work, what kind of work she can do, where she can work, with which animals, and how much she gets paid. This is the case in Uganda. In some parts of central Uganda, while the management – and cleanliness – of piggeries have improved, resulting in better perceptions about pig hygiene, lingering prejudices have meant women working in the pig industry have little bargaining power and lower incomes, and may feel pressured to work covertly. All this results in missed opportunity for women to develop professional skills and support their families, and reduced food safety for everyone.

In 2022/2023 we conducted a study in two districts to understand how local gender norms affected women in the pig farming sector. The findings revealed that women faced restrictions in conducting artificial insemination, castrating animals, taking sows to boars for mating, and transporting pigs on motorcycles. Additionally, certain activities – including slaughtering, trading livestock, producing feed, and owning large farms – were deemed inappropriate for women.

We also found systemic barriers such as lower wages, lack of control over income, restricted physical mobility, and exclusion from influential networks blocked them from fully reaping the benefits of the sector.

These findings led us to launch a range of interventions in the districts. Working with the international NGO Ripple Effect, my team at the International Livestock Research Institute and I trialled a range of interventions in Uganda’s Masaka and Mukono districts.

The results, evaluated a year later in December 2025, showed that social norms can be both accommodated and transformed for the benefit of all. For example, radio shows and conversations challenged widely held sentiments and sought to normalise roles that were taboo for women – such as providing pig insemination services to other farmers and contributing to a growing pig sector.

Our findings have lessons that are of value across many industries and in many places.

Doing things differently

We worked with pig farmers, business people, regulators and community members in five different communities to address the restrictive norms that prevented women from engaging in pig businesses. The work was carried out in Masaka district (south-west of Kampala) and Mukono district (east of the capital).

The interventions we put in place included:

  • providing women farmers with weigh-bands to estimate live pig weights and make sure they weren’t being cheated

  • offering training for women farmers to help them negotiate better prices and animal services

  • providing branded lab coats and badges to certified professionals to help combat the lack of respect for women in technical roles like artificial insemination

  • providing aprons, head wraps and boots to women working in slaughterhouses and butcher shops, so they would not be seen wearing dirty clothes.

These interventions provided solutions to accommodate existing norms without directly challenging them.

We also trialled some interventions aimed at transforming gender norms. We organised broadcasts on local radio talk shows, featuring a panel discussion between gender officers from Ripple Effect, community leaders and local men who explained why they supported their wives and daughters to work in the pig industry.

For instance, in one broadcast, one local leader shared his family’s story:

My wife rears pigs in large numbers, and I help her look for markets. When I travel, I bring her feeds for them. A home without money is unhappy. Piggery projects are family enterprises … When a woman earns an income, her husband is relieved financially; an empowered woman is a responsible woman.

We also held large community meetings, and used recordings from these shows to spark dialogue about these issues.

The changes

Over a year we observed changes.

Women butchers, farmers and artificial insemination agents felt more confident and accepted, and their services were sought after, especially by other women.

They were able to negotiate higher prices for their pigs. They invested their savings in their piggeries; some were able to use the profits to buy their own land and build houses.

There has been movement towards policy changes, too. Traditionally, pigs have had to be killed in official slaughterhouses – male-dominated spaces. Women did not feel welcome there, and men felt women would not be able to cope with the practical act of slaughter.

After our work in the sector, including inspection officials, authorities are now allowing some women to slaughter their pigs at home.

Lessons

Norms are powerful. Any efforts to improve livelihoods, boost community health, or grow a particular industry will be shaped by these norms. Ignoring them is a recipe for failure, while understanding them – and, where appropriate, moving beyond them – can benefit a whole community.

To transform restrictive norms, both men and women must be included in dialogues that encourage critical curiosity about their impacts. Religious, political and community leaders – people who often enforce these unwritten rules – must also be part of the conversations and solutions.

Radio talk shows and social media can showcase women successfully performing traditionally masculine tasks and supportive men, to normalise new behaviours and reduce shaming. And something as simple as professional clothing can send a signal that women are competent – and clean.

Gender norms can change, and these social changes can have practical and economic effects. Livestock development, as we have seen in Uganda’s pig industry, can be an entry point to promote gender equality.

At the same time, removing barriers to women’s participation can boost families’ incomes, bolster rural industries and alleviate poverty.

Challenge norms, empower women, and everyone benefits.

The Conversation

Esther Leah Achandi received funding from the CGIAR Sustainable Animal and Aquatic Foods (SAAF) Science Program, with funding from CGIAR Trust Fund donors. We would like to thank all funders who support this research through their contributions to the CGIAR Trust Fund:
www.cgiar.org/funders.

ref. Women working in Uganda’s pig sector: how challenging prejudices can unlock opportunities – research – https://theconversation.com/women-working-in-ugandas-pig-sector-how-challenging-prejudices-can-unlock-opportunities-research-277751

Political violence in South Africa is driven by a power elite trying to establish dominance – new research

Source: The Conversation – Africa (2) – By Ivor Chipkin, Associate lecturer, University of Pretoria

For much of the past two decades, South Africa’s recurring waves of protest have been interpreted through a dominant lens: the failure of the post-apartheid state to deliver services to its poorest citizens. Rising unemployment, corroding infrastructure and inadequate housing are the familiar explanations offered.

We are political scientists who have been analysing protests and protest data for years. In a recent article we propose that the overall pattern of protest activity in South Africa cannot be explained by socio-economic conditions alone. It tracks the internal power struggles of the ruling party, the African National Congress (ANC).

This has led us to a new reading of state capture.

As we set out in a paper in 2025, state capture in South Africa is often reduced to a phenomenon of large-scale corruption. The focus has been on the way that private businesses, working with politicians, repurposed legislative and administrative processes to serve their interests and disable the criminal justice system to avoid consequences.

The conventional understanding casts state capture as looting: the opportunistic and organised theft of public resources by politically connected networks and enabled by a compromised presidency.

We do not contest the reality of this pillaging. But we argue that it was also something more structurally purposeful. State capture, in our account, was the mechanism by which former president Jacob Zuma sought to forge a “power elite” in the ANC.

This is a term we borrow from the sociological tradition of C. Wright Mills to refer to a small cohesive group that is able to make decisions with national consequences in political, military and economic institutions. In contrast a politically connected network may have influence but is too diffuse to exercise power as such.

The power elite matters because it explains who really makes the biggest decisions in society and why democratic institutions do not always fully control those decisions.

The argument we’re presenting has consequences for how the country understands what state capture is, and the trajectory of South African democracy itself.

Protests as a barometer

Drawing on data from the South African Police Service, the Armed Conflict Location and Event Data Project, and the Institute for Security Studies, we identify a striking pattern. Protest events rose sharply from around 2006, reaching what some researchers called “insurrectionary proportions” by 2011.

Then they stabilised and began to decline between roughly 2013 and 2017. This period coincided with the consolidation of Zuma’s hold on power and the height of state capture.

After 2018, protests surged again to unprecedented levels. In 2021, the country
experienced its worst civil revolt since the end of apartheid.

The socio-economic conditions typically cited to explain protest – unemployment, inequality, poor service delivery – do not follow this same pattern. They did not improve during the 2013-2017 lull. If anything, they worsened. As our paper records, municipal audit outcomes deteriorated sharply by the end of the period.

Inequality, measured by Gini coefficients across South Africa’s major
cities, remained essentially unchanged. The exception was Cape Town, where inequality seems to have declined.

The stabilisation of protest activity, we conclude, cannot be attributed to improvements in the living conditions of poor South Africans.

Something else was suppressing the mobilisation of discontent.

Our answer draws on political sociology and on comparative work on elite formation in Africa and beyond. We conclude that protests are instruments of elite competition. This includes the tactical deployments of professional agitators by local politicians and their networks contesting for control of resources, positions and patronage within the ANC.

When these competitions are acute and unresolved, they spill outward as protests. When they are contained, protest subsides.

The how

By repurposing state-owned enterprises away from their public mandates, the Zuma network generated enormous rents that were then used for private enrichment and to finance factional political activity. This included paying for party rallies, sustaining provincial and regional networks, creating sympathetic media infrastructure, and distributing cash and contracts to potential opponents in exchange for loyalty or silence.

The result was a temporary stabilisation of what had been a fractured and contested elite terrain.

Between roughly 2013 and 2017, a group of politically aligned operators was able to discipline internal competition, in part by allocating positions in government, state-owned enterprises and the party apparatus.

Those who would not be bought were expelled, marginalised, or subjected to violence. We note that political assassinations rose sharply during Zuma’s second term. Evidence before the Zondo Commission into state capture pointed to the deployment of armed units under presidential operational control.

The relative “stability” observable in protest data between 2013 and 2017 was the successful suppression of elite competition through corruption, patronage and coercion. The modest improvement in municipal spending was the result of elite power exercised over administrative systems.

The unravelling under Ramaphosa

If Zuma’s presidency saw the construction of a power elite, Cyril Ramaphosa’s has seen its unravelling.

The consequences have been severe.

At the ANC’s 54th national conference in December 2017, Ramaphosa narrowly defeated Nkosazana Dlamini-Zuma for the party presidency. Zuma’s internal compact then began to fracture. The spike in protest activity that followed was almost immediate.

Ramaphosa was not prepared to deploy corruption and violence as political solutions. But without an alternative basis for managing elite competition, the ANC’s internal fissures deepened.

There were symptoms of this disintegration in 2023:

Gatekeeping became decentralised and unregulated. Elite contestation began migrating out of the party system altogether.

A sobering conclusion, and hint of hope

We conclude that some of it will be pushed towards organised crime. Mafia-type networks, we suggest, should be expected to grow.

There is, however, a more hopeful possibility. The reason the ANC has functioned as the primary arena for elite competition is that it has controlled access to the “gate” – the allocation of positions in the state, the civil service and state-owned enterprises.

Remove that control, and the character of elite competition changes. This is precisely what is at stake in the amendments to the Public Service Act of 1994. Signed into law by Ramaphosa on 26 March 2026, it was gazetted on 1 April 2026.

The legislation aims to:

  • reduce executive discretion over appointments in the public service

  • insulate civil service recruitment and operations from party-political interference.

If implemented, political parties will be compelled to compete for support through policy and performance rather than patronage. Elite competition will shift to the public administration system itself. Ideally, this will be governed by merit, transparency and professional standards.

We are cautious about the prospects for this reform. History is not encouraging and the political conditions are challenging.

But if it can end gatekeeping, new legislation like the Public Service Amendment Act will change the elite social terrain in South Africa.

The Conversation

Ivor Chipkin and Jelena Vidojević are affiliated with New South Institute

Jelena Vidojević does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

ref. Political violence in South Africa is driven by a power elite trying to establish dominance – new research – https://theconversation.com/political-violence-in-south-africa-is-driven-by-a-power-elite-trying-to-establish-dominance-new-research-280504

L’IA générative, première révolution cognitive de l’histoire du travail

Source: The Conversation – in French – By Hugo Spring-Ragain, Doctorant en économie / économie mathématique, Centre d’études diplomatiques et stratégiques (CEDS)

L’intelligence artificielle ne détruit pas tant des emplois qu’elle modifie profondément les compétences nécessaires pour les accomplir. De cette confusion entre emploi et compétences risquent de naître des erreurs dans les politiques d’accompagnement des mutations en cours.


Chaque grande vague technologique a produit son lot de prédictions contradictoires sur l’emploi. L’intelligence artificielle (IA) ne fait pas exception. Mais avant de savoir combien d’emplois l’IA va créer ou détruire, il faudrait s’accorder sur ce qu’elle automatise réellement. La réponse oblige à distinguer trois notions que le débat public confond régulièrement : l’emploi, la compétence et la tâche.

Les grandes vagues d’automatisation ont suivi une logique remarquablement stable en deux siècles : vapeur, électricité, robotique industrielle ont déplacé les tâches physiques répétitives et épargné le travail cognitif non routinier. Cette régularité empirique a été formalisée par Autor, Levy et Murnane dès 2003 sous le nom d’« hypothèse de polarisation des tâches ».

Une illusion persistante

L’automatisation ronge les emplois intermédiaires, ceux des cols bleus qualifiés et employés de bureau exécutant des tâches routinières, mais épargne les deux extrémités. D’un côté, les tâches manuelles non routinières, comme la plomberie ou les soins, de l’autre, les tâches cognitives non routinières, comme l’analyse, le conseil ou la rédaction experte. Ces dernières constituaient le cœur des professions du tertiaire qualifié, et la conviction s’était solidement installée qu’elles resteraient hors d’atteinte.




À lire aussi :
Pourquoi l’IA oblige les entreprises à repenser la valeur du travail


Cette conviction reposait sur une confusion conceptuelle qu’il faut dissiper avant tout. Ce n’est pas l’emploi de juriste ou d’analyste financier qui était protégé, c’est un ensemble de tâches précises qui composaient cet emploi et qui résistaient jusqu’ici à l’automatisation. La distinction entre ces trois niveaux est fondamentale.

Un emploi désigne un poste occupé dans une organisation, avec un contrat, un salaire, une fiche de poste. Une compétence est une capacité cognitive ou technique mobilisable dans plusieurs contextes professionnels. Une tâche est une action précise, délimitable, dont on peut évaluer si elle est ou non automatisable à un coût donné. C’est à ce troisième niveau que se joue réellement la transformation en cours, et c’est précisément ce niveau que le débat public ignore.

Rupture dans la longue histoire du capitalisme industriel

L’IA générative constitue une rupture dans cette longue histoire. Pour la première fois depuis l’industrialisation, les tâches cognitives qualifiées, rédaction, analyse documentaire, synthèse, production de premiers jets, se retrouvent directement exposées. Eloundou, Manning, Mishkin et Rock estiment qu’environ 80 % de la population active états-unienne pourrait voir au moins 10 % de ses tâches affectées par les grands modèles de langage, et que cette exposition croît avec le niveau de salaire. C’est l’exact inverse du schéma observé lors de toutes les vagues précédentes.

Le cadre analytique développé par Acemoglu et Restrepo permet d’aller plus loin. Leur modèle distingue deux effets opposés produits par toute vague d’automatisation :

  • l’effet de déplacement, d’abord : des travailleurs perdent des tâches au bénéfice de la machine, ce qui réduit mécaniquement la demande de travail et pèse sur les salaires des groupes affectés ;

  • l’effet de réintégration, ensuite : l’automatisation produit de nouvelles tâches où la valeur humaine est décisive, générant une demande compensatrice.

L’histoire longue du capitalisme industriel peut se lire comme une succession de ces deux effets, le second finissant généralement par compenser le premier.

Le cas de la traduction permet de voir très concrètement comment déplacement et réintégration se combinent, l’IA générative peut produire en quelques secondes un premier jet dans une autre langue, ce qui déplace une partie du travail auparavant effectué par des traducteurs humains vers la machine. Mais cette automatisation réintègre simultanément d’autres tâches ou renforce leur importance, telles que la vérification des contresens, l’adaptation au contexte culturel, l’harmonisation de la terminologie, le contrôle de la qualité et la validation finale.

Potentiel déséquilibre

Ce qui est préoccupant avec l’IA générative, c’est le déséquilibre potentiel entre ces deux dynamiques. Le déplacement s’opère à une vitesse que les marchés du travail et les institutions de formation peinent à absorber, tandis que la réintégration reste encore largement à construire.

Cependant, le phénomène le plus important n’est pas sectoriel, mais il est interne aux métiers eux-mêmes. Dans ses « Perspectives de l’emploi », l’OCDE met en évidence que les professions les plus exposées à l’IA générative sont précisément celles à forte densité cognitive : finance, droit, conseil, enseignement supérieur. Contrairement aux vagues précédentes qui frappaient les zones rurales et les bassins industriels, l’exposition est désormais plus forte dans les grandes métropoles et chez les travailleurs hautement qualifiés, un renversement géographique et social inédit.

Redistribuer les tâches

Ce renversement se joue concrètement au niveau de la tâche.

Dans un même poste d’analyste financier ou de juriste, certaines tâches migrent vers l’IA (produire un résumé exécutif, générer une première analyse de contrat, synthétiser une revue de littérature), tandis que d’autres se revalorisent mécaniquement : définir le cadre d’analyse pertinent, évaluer la qualité d’un raisonnement automatisé, détecter une erreur factuelle dans un output, assumer la responsabilité juridique ou éthique d’une décision. Ce ne sont pas des emplois qui disparaissent. Ce sont des bouquets de tâches qui se redistribuent entre humains et machines, transformant de l’intérieur ce qu’un employeur attend d’un salarié qualifié.

Cette redistribution des tâches a une implication directe sur les compétences qui seront réellement valorisées dans les années à venir, et elle renverse une partie des évidences habituelles sur la formation professionnelle.

Former les travailleurs à utiliser l’IA au sens instrumental, maîtriser un outil, rédiger des prompts efficaces, s’approprier une interface, est utile à court terme, mais c’est insuffisant si la compétence réellement demandée demain n’est pas de produire avec l’IA, mais de superviser et de critiquer ce qu’elle produit.

Un enjeu de formation

Or, superviser efficacement un output d’IA requiert exactement ce que les formations courtes et techniques peinent à développer : une culture générale solide permettant de détecter une erreur de fond, une capacité argumentative pour évaluer la cohérence d’un raisonnement, une connaissance des biais cognitifs pour identifier les angles morts d’une analyse automatisée. Ce sont des compétences que les sciences de l’éducation regroupent sous le terme de métacompétences : apprendre à apprendre, à exercer un jugement critique, à mobiliser des savoirs dans des situations inédites.

Arte, 2025.

Le paradoxe devient alors le suivant. À mesure que l’IA automatise les tâches routinières de la connaissance, elle valorise précisément ce que les formations généralistes et les cursus de sciences humaines cultivent de longue date et que les débats sur l’employabilité ont eu tendance à déconsidérer au profit de compétences techniques plus immédiatement mesurables.

Non par nostalgie des humanités, mais par logique économique pure. Si la machine produit le texte, l’analyse et la synthèse, la valeur marginale de l’humain réside dans sa capacité à juger si ce texte dit vrai, si cette analyse est pertinente au regard du contexte réel, si cette synthèse sert l’objectif poursuivi.

The Conversation

Hugo Spring-Ragain ne travaille pas, ne conseille pas, ne possède pas de parts, ne reçoit pas de fonds d’une organisation qui pourrait tirer profit de cet article, et n’a déclaré aucune autre affiliation que son organisme de recherche.

ref. L’IA générative, première révolution cognitive de l’histoire du travail – https://theconversation.com/lia-generative-premiere-revolution-cognitive-de-lhistoire-du-travail-279911

« StravaLeaks » : quand les traces numériques deviennent un enjeu de sécurité

Source: The Conversation – in French – By Fabrice Lollia, Docteur en sciences de l’information et de la communication, chercheur associé laboratoire DICEN Ile de France, Université Gustave Eiffel

L’affaire « StravaLeaks » montre que, dans un monde saturé d’objets connectés et de données de localisation, les traces numériques ordinaires sont devenues un enjeu central de sécurité pour les environnements sensibles. De simples données de déplacement issues d’un footing, enregistrées et partagées par une application publique, ont pu être utilisées pour localiser des navires ou des bases militaires.


Un footing, en apparence, n’a rien de sensible. Pourtant, en mars 2026, une activité enregistrée sur Strava par un militaire français a permis de localiser en temps quasi réel le porte‑avions Charles-de-Gaulle en Méditerranée orientale. Dès 2018, la carte de chaleur mondiale de Strava – une visualisation agrégée des activités publiques enregistrées par ses utilisateurs – avait déjà révélé des bases militaires et des sites sensibles, et des enquêtes plus récentes ont montré que les pratiques sportives de gardes du corps pouvaient trahir des habitudes de déplacement de chefs d’État.

Le problème ne vient pas d’un piratage sophistiqué, mais d’un usage banal de montre connectée, compte public et trace GPS accessible en ligne. Ce cas illustre comment la sécurité d’aujourd’hui ne se limite plus à la protection physique, mais inclut aussi la maîtrise des traces numériques produites par nos comportements les plus ordinaires.

Quand une application déborde de son usage initial

Strava est une application conçue pour suivre et partager des performances sportives. Son usage premier relève du loisir, de la sociabilité numérique et du suivi de soi, non de la documentation d’activités sensibles. C’est pourtant là toute l’ambivalence de ce type d’outils car sans avoir été pensés pour la sécurité, ils peuvent produire des effets très concrets sur elle.

À mesure que les technologies de traçabilité s’installent dans les usages quotidiens, elles cessent d’apparaître comme des dispositifs de contrôle. Elles deviennent des outils familiers, associés au confort ou à l’optimisation des pratiques. Dès lors, une course, un itinéraire répété, un point de départ ou d’arrivée ou une activité enregistrée en mer peuvent révéler bien davantage qu’une simple pratique sportive. Une donnée de performance peut devenir un indice sur une routine, une présence ou une habitude de déplacement.

Le cas Strava n’est d’ailleurs pas isolé. À l’aéroport d’Heathrow (Londres), en 2014, des toilettes connectées ont été testées pour mesurer anonymement leur fréquentation, améliorer le nettoyage et mieux répartir les moyens de maintenance. L’exemple peut sembler éloigné, mais il montre que, au-delà des outils explicitement sécuritaires, des dispositifs connectés collectent eux aussi discrètement des traces numériques sur les comportements des usagers. En ce sens, la vulnérabilité ne naît plus seulement d’une attaque ou d’une fuite volontaire, mais aussi d’usages ordinaires dont les effets de visibilité sont souvent sous-estimés.

La sécurité ne se joue plus seulement sur le terrain

Longtemps, la sécurité a été pensée selon un modèle essentiellement physique. Il fallait protéger une personne, sécuriser un déplacement, contrôler un périmètre, anticiper une menace. Cette logique est toujours d’actualité mais, à l’ère numérique, elle ne suffit plus.

Dans un environnement saturé d’objets connectés, de plateformes et de données de localisation, la vulnérabilité peut désormais naître à la périphérie du dispositif de protection. Elle ne résulte plus forcément d’une intrusion ou d’une action malveillante. Elle peut venir d’un usage mal paramétré, d’une routine numérique non interrogée ou d’un outil utilisé sans conscience de ses effets de visibilité.

La sécurité d’un responsable politique, d’un chef d’entreprise, d’un diplomate ou d’un site sensible dépend donc aussi des traces numériques produites par son environnement humain et technique : assistants, chauffeurs, escortes, collaborateurs, militaires, objets connectés, applications de suivi ou réseaux de partage. Protéger une « personne sensible », une personnalité, aujourd’hui, ce n’est plus seulement protéger son corps ou son itinéraire. C’est aussi protéger l’écosystème informationnel qui l’entoure.

Cette évolution renvoie à une sécurité de plus en plus renforcée par la technologie via les capteurs, les données et les outils de suivi. Mais l’ajout de technologie ne supprime pas la vulnérabilité. C’est justement le problème d’une lecture technosolutionniste qui surestime la complémentarité humain-machine. Elle rappelle au contraire qu’une technologie n’est efficace qu’à condition d’être articulée à l’analyse humaine, à l’expérience de terrain et à une compréhension fine du contexte. Certes, la technologie renforce donc la vigilance, mais elle ne remplace ni le jugement, ni la formation, ni la culture du risque.

La vulnérabilité observée est également organisationnelle, culturelle et humaine. Elle naît d’une forme de non-concordance entre la banalité des usages numériques (courir avec une montre connectée, par exemple) et la sensibilité des environnements dans lesquels elles s’inscrivent (être dans un endroit classifié secret-défense). Un même outil peut être perçu comme un outil de confort ou de performance tout en produisant des effets d’exposition importants.

Former devient donc aussi important qu’équiper dans la mesure où il ne s’agit pas seulement d’interdire certains usages, mais plutôt de faire comprendre comment une trace numérique par définition invisible peut, par agrégation et recoupement, devenir une information sensible. La sécurité ne se joue donc plus dans le contrôle des outils, mais dans l’intelligence des pratiques.

Réintégrer l’humain au centre de la doctrine de sécurité

L’un des principaux enseignements de ces affaires est qu’aucune technologie ne protège à elle seule. Une application, une montre connectée ou un dispositif de géolocalisation ne sont ni bons ni mauvais en soi. Comme le montre la recherche, tout dépend du cadre dans lequel ils sont utilisés, des règles qui les entourent et de la capacité des acteurs à en comprendre les effets. Raison pour laquelle la réponse ne peut pas être juste « technique ».

Elle suppose aussi une doctrine d’usage, une formation adaptée et une culture sécuritaire partagée. À l’inverse la traçabilité peut aussi renforcer la protection, mais elle ne remplace ni l’analyse humaine, ni l’appréciation du contexte, ni les méthodes classiques de sécurité.

Autrement dit, la sécurité des environnements sensibles repose sur une complémentarité entre l’outil et l’humain. Il ne suffit pas de déployer des dispositifs ; encore faut-il que les utilisateurs comprennent ce qu’ils produisent, ce qu’ils exposent et les conséquences possibles de leurs usages entre une possible surveillance et une souveillance, c’est-à-dire une forme plus discrète de captation des traces intégrées aux gestes ordinaires et parfois à peine perçue par ceux qui y participent.

Dans le cas Strava, l’enjeu n’est donc pas seulement de mieux paramétrer une application. Il est de construire une culture du risque numérique, capable d’intégrer les gestes les plus ordinaires à la réflexion sécuritaire.

Ce que la recherche nous enseigne en lien avec ces cas est que la vraie leçon de ces affaires est peut-être là : dans un monde connecté, la menace ne réside pas seulement dans ce que l’on cherche à cacher, mais aussi dans ce que l’on produit sans y penser.

Ces affaires dites « StravaLeaks » montrent que la traçabilité numérique, loin d’être un simple confort d’usage, peut devenir un enjeu de sûreté dès lors qu’elle s’inscrit dans un environnement sensible. Protéger, aujourd’hui, ce n’est plus seulement verrouiller un périmètre ou escorter une personnalité. C’est aussi apprendre à gouverner les traces que produisent les usages les plus ordinaires.

The Conversation

Fabrice Lollia ne travaille pas, ne conseille pas, ne possède pas de parts, ne reçoit pas de fonds d’une organisation qui pourrait tirer profit de cet article, et n’a déclaré aucune autre affiliation que son organisme de recherche.

ref. « StravaLeaks » : quand les traces numériques deviennent un enjeu de sécurité – https://theconversation.com/stravaleaks-quand-les-traces-numeriques-deviennent-un-enjeu-de-securite-279942