Category: MIL-OSI

Source: The Conversation – UK – By Alex Dryden, PhD Candidate in Economics, SOAS, University of London

The global investment frenzy around AI has seen companies valued at trillions of dollars and eye-watering projections of how it will boost economic productivity.

But in recent weeks the mood has begun to shift. Investors and CEOs are now openly questioning whether the enormous costs of building and running AI systems can really be justified by future revenues.

Google’s CEO, Sundar Pichai, has spoken of “irrationality” in AI’s growth, while others have said some projects are proving to be more complex and expensive than expected.

Meanwhile, global stock markets have declined, with tech shares taking a particular hit, and the value of cryptocurrencies has dipped as investors appear increasingly nervous.

So how should we view the health of the AI sector?

Well, bubbles in technology are not new. There have been great rises and great falls in the dot-com world, and surges in popularity for certain tech platforms (during COVID for example) which have then flattened out.

Each of these technological shifts was real, but they became bubbles when excitement about their potential ran far ahead of companies’ ability to turn popularity into lasting profits.

The surge in AI enthusiasm has a similar feel to it. Today’s systems are genuinely impressive, and it’s easy to imagine them generating significant economic value. The bigger challenge comes with how much of that value companies can actually keep hold of.

Investors are assuming rapid and widespread AI adoption along with high-margin revenue. Yet the business models needed to deliver that outcome are still uncertain and often very expensive to operate.

This creates a familiar gap between what the technology could do in theory, and what firms can profitably deliver in practice. Previous booms show how quickly things wobble when those ideas don’t work out as planned.

AI may well reshape entire sectors, but if the dazzling potential doesn’t translate quickly into steady, profitable demand, the excitement can slip away surprisingly fast.

Fit to burst?

Investment bubbles rarely deflate on their own. They are usually popped by outside forces, which often involve the US Federal Reserve (the US’s central bank) making moves to slow the economy by raising interest rates or limiting the supply of money, or a wider economic downturn suddenly draining confidence.

For much of the 20th century, these were the classic triggers that ended long stretches of rising markets.

But financial markets today are larger, more complex, and less tightly tied to any single lever such as interest rates. The current AI boom has unfolded despite the US keeping rates at their highest level in decades, suggesting that external pressures alone may not be enough to halt it.

Instead, this cycle is more likely to end from within. A disappointment at one of the big AI players – such as weaker than expected earnings at Nvidia or Intel – could puncture the sense that growth is guaranteed.

Alternatively, a mismatch between chip supply and demand could lead to falling prices. Or investors’ expectations could quickly shift if progress in training ever larger models begins to slow, or if new AI models offer only modest improvements.

Overall then, perhaps the most plausible end to this bubble is not a traditional external shock, but a realisation that the underlying economics are no longer keeping up with the hype, prompting a sharp revaluation across related stocks.

Artificial maturity

If the bubble did burst, the most visible shift would be a sharp correction in the valuations of chipmakers and the large cloud companies driving the current boom.

These firms have been priced as if AI demand will rise almost without limit. So any sign that the market is smaller or slower than expected would hit financial markets hard.

This kind of correction wouldn’t mean AI disappears, but it would almost certainly push the industry into a more cautious, less speculative phase.

The deepest consequence would be on investment. Goldman Sachs estimates that global spending on AI-related infrastructure could reach US$4 trillion by 2030. In 2025 alone, Microsoft, Amazon, Meta and Google’s owner Alphabet have poured almost US$350 billion into data centres, hardware and model development. If confidence faltered, much of this planned expansion could be scaled back or delayed.

That would ripple through the wider economy, slowing construction, dampening demand for specialised equipment, and dragging on growth at a time when inflation remains high.

But a bursting AI bubble would not erase the technology’s long-term importance. Instead, it would force a shift away from the “build it now, profits will follow” mindset which is driving much of the current exuberance.

Companies would focus more on practical uses that genuinely save money or raise productivity, rather than speculative bets on transformative breakthroughs. The sector would mature. But it would probably do so only after a painful period of adjustment for investors, suppliers and governments who have tied their growth expectations to an uninterrupted AI boom.

Alex Dryden 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. Is the AI bubble about to burst? What to watch for as the markets wobble – https://theconversation.com/is-the-ai-bubble-about-to-burst-what-to-watch-for-as-the-markets-wobble-270113

Source: The Conversation – UK – By Vincent Charles, Reader in AI for Business and Management Science, Queen’s University Belfast

On a sunny morning on October 19 2025, four men allegedly walked into the world’s most-visited museum and left, minutes later, with crown jewels worth €88 million (£76 million). The theft from Paris’s Louvre Museum – one of the world’s most surveilled cultural institutions – took just under eight minutes.

Visitors kept browsing. Security didn’t react (until alarms were triggered). The men disappeared into the city’s traffic before anyone realised what had happened.

Investigators later revealed that the thieves wore hi-vis vests, disguising themselves as construction workers. They arrived with a furniture lift, a common sight in Paris’s narrow streets, and used it to reach a balcony overlooking the Seine. Dressed as workers, they looked as if they belonged.

This strategy worked because we don’t see the world objectively. We see it through categories – through what we expect to see. The thieves understood the social categories that we perceive as “normal” and exploited them to avoid suspicion. Many artificial intelligence (AI) systems work in the same way and are vulnerable to the same kinds of mistakes as a result.

The sociologist Erving Goffman would describe what happened at the Louvre using his concept of the presentation of self: people “perform” social roles by adopting the cues others expect. Here, the performance of normality became the perfect camouflage.

The sociology of sight

Humans carry out mental categorisation all the time to make sense of people and places. When something fits the category of “ordinary”, it slips from notice.

AI systems used for tasks such as facial recognition and detecting suspicious activity in a public area operate in a similar way. For humans, categorisation is cultural. For AI, it is mathematical.

But both systems rely on learned patterns rather than objective reality. Because AI learns from data about who looks “normal” and who looks “suspicious”, it absorbs the categories embedded in its training data. And this makes it susceptible to bias.

The Louvre robbers weren’t seen as dangerous because they fit a trusted category. In AI, the same process can have the opposite effect: people who don’t fit the statistical norm become more visible and over-scrutinised.

It can mean a facial recognition system disproportionately flags certain racial or gendered groups as potential threats while letting others pass unnoticed.

A sociological lens helps us see that these aren’t separate issues. AI doesn’t invent its categories; it learns ours. When a computer vision system is trained on security footage where “normal” is defined by particular bodies, clothing or behaviour, it reproduces those assumptions.

Just as the museum’s guards looked past the thieves because they appeared to belong, AI can look past certain patterns while overreacting to others.

Categorisation, whether human or algorithmic, is a double-edged sword. It helps us process information quickly, but it also encodes our cultural assumptions. Both people and machines rely on pattern recognition, which is an efficient but imperfect strategy.

A sociological view of AI treats algorithms as mirrors: they reflect back our social categories and hierarchies. In the Louvre case, the mirror is turned toward us. The robbers succeeded not because they were invisible, but because they were seen through the lens of normality. In AI terms, they passed the classification test.

From museum halls to machine learning

This link between perception and categorisation reveals something important about our increasingly algorithmic world. Whether it’s a guard deciding who looks suspicious or an AI deciding who looks like a “shoplifter”, the underlying process is the same: assigning people to categories based on cues that feel objective but are culturally learned.

When an AI system is described as “biased”, this often means that it reflects those social categories too faithfully. The Louvre heist reminds us that these categories don’t just shape our attitudes, they shape what gets noticed at all.

After the theft, France’s culture minister promised new cameras and tighter security. But no matter how advanced those systems become, they will still rely on categorisation. Someone, or something, must decide what counts as “suspicious behaviour”. If that decision rests on assumptions, the same blind spots will persist.

The Louvre robbery will be remembered as one of Europe’s most spectacular museum thefts. The thieves succeeded because they mastered the sociology of appearance: they understood the categories of normality and used them as tools.

And in doing so, they showed how both people and machines can mistake conformity for safety. Their success in broad daylight wasn’t only a triumph of planning. It was a triumph of categorical thinking, the same logic that underlies both human perception and artificial intelligence.

The lesson is clear: before we teach machines to see better, we must first learn to question how we see.

The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.

– ref. How the Louvre thieves exploited human psychology to avoid suspicion – and what it reveals about AI – https://theconversation.com/how-the-louvre-thieves-exploited-human-psychology-to-avoid-suspicion-and-what-it-reveals-about-ai-269842

Source: The Conversation – UK – By Andrea Rigon, Professor, Politecnico di Milano, and, UCL

The UK has announced much harsher rules for asylum seekers including the prospect of more deportations for those whose applications fail. The US is trebling the size of its deportation force. The EU is doubling its border budgets. And in the coming decades, hundreds of millions of people might be displaced by ecological changes.

In the face of this challenge, those countries which are most responsible for climate change have two options. Either they can share resources more equitably, and fund adaptation plans on a massive scale. Or they can prevent others from accessing resources and liveable land through physical and regulatory walls, enforced through mass deportation.

Recent events show that, faced with this choice, many governments are choosing not to share resources to anywhere near the extend needed, and are instead building higher walls.

Climate change is already making life unliveable in some parts of the world. According to a 2020 report from thinktank the Institute for Economics and Peace (IEP), 2.6 billion people face high or extreme water stress. By 2040, this may jump to 5.4 billion. Droughts, heatwaves, floods, cyclones, food shortages and related conflicts will force millions from their homes.

The IEP warns that up to 1.2 billion people globally might be displaced by 2050, while even the more-cautious World Bank predicts 216 million climate migrants.

Most of these people will move internally within nations, but this too is likely to mean more walls and borders. In very unequal countries, internal migration has already triggered security-driven responses, with a rise in gated communities and other segregated living arrangements to keep the poorer away from the wealthy.

Many other climate migrants will be pushed to travel internationally. It’s likely their motivation will be characterised by many as economic rather than due to climate change. But it’s misleading to separate “economic” from “climate” migrants. When drought kills crops in Somalia or floods wash away farmland in Pakistan, the loss of income is inseparable from the climate shocks that caused it.

Even before the worst impacts hit, climate change is already woven into the economic pressures that push people to move – shrinking harvests, emptying wells and ruining livelihoods. The most severe climate-driven displacement is still ahead, but it has already begun.

Importantly, these pressures come with inequalities in causing climate change and bearing the costs. The richest 1% of the world’s population produces as much carbon as the poorest two-thirds, according to a study of global emissions in 2019 by Oxfam. Northern Europe and the US alone account for 92% of historical emissions.

Those who have contributed the least to climate change are the worst affected and often have the fewest resources to adapt, forcing many people to migrate.

More walls, more deportations

In this context, governments of wealthier countries are massively increasing spending on migration policing. In the US, proposed funding levels are extraordinary.

Recent legislation allocates nearly US$30 billion (£22 billion) to the Immigration and Customs Enforcement agency (Ice) for enforcement and deportation operations – roughly three times its current budget.

The US has also authorised US$45 billion for new detention centres – a 265% increase, more than the entire defence budget of Italy – and US$46.6 billion for additional border walls. Under this plan, Ice would become the largest US law enforcement agency, three times the size of the FBI.

Donald Trump’s policies can be easily labelled as the excess of one would-be autocrat, but this is a global trend across the political spectrum, albeit implemented with more acceptable language by the centre-left.

Introducing the UK Labour government’s new asylum and returns policy, Prime Minister Keir Starmer said: “We need an approach with a stronger deterrent effect and rules that are robustly enforced.” But previously-supportive MPs from his own party have warned this will mean “Ice-style raids” to deport asylum seekers.

The European Commission’s 2028–34 budget proposal earmarks €25.2 billion (£21.7 billion) for border management and €12 billion for migration, plus €11.9 billion for the Frontex border agency – more than double its current resources.

All this effectively triples current migration and border spending. In 2024, the EU ordered 453,000 non-EU nationals to leave, and actually deported 110,000 of them.

This is part of a much wider pattern, with borders today being far more militarised than at the end of the cold war. After decades of globalisation, states are now reterritorialising, building armoured fortifications against unwanted flows.

In the past two decades, more than 70 new international barriers have gone up, including Poland’s barbed-wire fence with Belarus, Greece’s steel wall on the Turkish border, Turkey’s stone wall on its Iranian border, and the new sections of the infamous wall between the US and Mexico.

Israel has built an “iron wall” around Gaza and border fences through much of the West Bank. Supposedly built to prevent Palestinians moving into Israel, these barriers have become a clear example of migration control tied to power grabs for land and resources.

A crossroads for human rights

Resource-driven migration pressures are rising just as the world is hardening its borders. In July 2025, the International Court of Justice declared that countries have a legal responsibility to address and compensate for climate change – and can be held accountable for their emissions. It is another signal that as humanity, we are at a crossroads.

The world can either prioritise universal human rights by sharing resources. Or it can attempt to protect a small, wealthy minority through walls, mass deportations and border violence on an unprecedented scale.

Andrea Rigon has received funding from UKRI and now receives funding from Fondation Botnar.

– ref. How the rich world is fortifying itself against climate migration – https://theconversation.com/how-the-rich-world-is-fortifying-itself-against-climate-migration-262936

Source: The Conversation – UK – By Catherine Norton, Associate Professor Sport & Exercise Nutrition, University of Limerick

Once the clocks have gone back and darkness falls before many of us even leave work, the rhythms of winter can feel heavier — shorter days, darker evenings, and often, later dinners. But shifting when we eat during the winter could make these months a little easier on our bodies and minds.

Our bodies operate on circadian rhythms – internal 24-hour clocks that regulate sleep, metabolism, digestion and hormone cycles. These rhythms are naturally synchronised with light and dark, so when daylight fades earlier, our metabolism also begins to wind down.

This connection between metabolism and daylight may help explain why a growing body of research from the field of chrononutrition suggests that when we eat may be nearly as important as what we eat. Chrononutrition examines how meal timing interacts with out internal body clock, and what affect short days might have on mood, metabolism and health.

For instance, one study found that healthy adults who ate dinner at 10pm experienced 20% higher blood sugar peaks and burned 10% less fat compared to those who ate dinner at 6pm. This was despite both groups eating identical meals and having similar bedtimes.

Broader analyses support the same trends, with a meta-analysis of 29 trials reporting that earlier eating windows, fewer meals and eating the bulk of one’s calories earlier in the day were linked to greater weight loss and improved metabolic markers (such as better blood pressure and lower blood sugar and cholesterol levels).

Other research links consistent late-night eating – especially close to bedtime – with poorer health outcomes and a greater risk of obesity and metabolic disorders such as type 2 diabetes.

Earlier dinners may better align with the body’s natural metabolic rhythms, particularly when the last meal occurs well before the body enters its “rest” phase. This might explain why eating earlier has health benefits.

Many chronobiologists conclude that aligning food intake with circadian biology represents a promising, low-cost method of improving metabolic outcomes – especially when combined with other lifestyle factors such as physical activity and healthy eating.

Eating with intent

In winter, especially in northern latitudes, shorter days and longer nights can disrupt circadian rhythms.

Reduced sunlight can lower serotonin levels, contributing to low mood or seasonal affective disorder (SAD). When paired with longer evenings indoors, it’s common for people to snack more often or delay eating dinner until later at night.

But digestion, hormone release (including those that help with sleep and digestion) and even the amount of calories you burn throughout the day all follow circadian rhythms. When meals are pushed too close to sleep, these processes overlap in ways that can affect both metabolism and rest – potentially increasing risks of poor sleep and metabolic ill health.

While light and dark have the biggest influence on circadian rhythms, food intake, stress, physical activity and temperature also affect them.

So, should you eat dinner earlier in winter?

For some people, yes — at least a little earlier. There are three main reasons why.

The first has to do with metabolic alignment. Eating when your metabolism is still active supports better blood sugar control, energy use and fat burning.

The second has to do with digestion. Leaving a few hours between dinner and bedtime allows digestion to wind down before sleep, which may improve sleep quality and recovery.

The third reason has to do with supporting mood and circadian rhythms. A consistent eating window and earlier dinner can help anchor daily routines – especially helpful when other time cues (such as daylight) are weaker.

But here’s the caveat: this isn’t a one-size-fits-all solution. Many different factors – such as how active you are, if you have any chronic conditions and your schedule – need to be taken into account.

An elite athlete training in the evening may need a later meal to support performance and recovery. But someone less active might benefit more from an earlier, lighter dinner.

So rather than rigid rules, think of meal timing as a flexible tool in your nutrition toolkit. The real focus should be on eating with intent.

This means taking into account your goals (such as whether you want to lose weight or boost athletic performance), how often you exercise, how close to bedtime you normally eat, how you feel depending on the time of day you eat dinner and what’s realistic given your schedule.

If you’re eating after 9pm most nights and waking up sluggish or find sleep less restful, experimenting with earlier meals may be worthwhile. But if you’re training late or eating socially, that’s fine too — focus on quality over timing, choosing lighter, balanced meals and allowing at least two to three hours before bed.

Some other mealtime tips you can try during the darker months include:

• finishing dinner earlier, ideally between 5.30pm–7.00pm, or at least two to three hours before bedtime
• front loading your calories by making breakfast and lunch more substantial while there’s more daylight and your metabolism is more active
• planing around activity, so if you exercise late, have your main meal earlier and a small recovery snack afterwards
• keeping a consistent eating window, finishing eating by around 8pm most nights to support circadian alignment
• reflecting and adjusting by noting how meal timing affects your energy, sleep quality and mood for a week or two then changing as needed
• staying flexible by remembering perfection isn’t required – a regular schedule and awareness of what you need is what counts.

As winter settles in, paying attention to when you eat may be just as important as what you eat. Aligning mealtimes with your body’s natural rhythms can help steady energy, mood and sleep through the darker months.

But the real key is intentionality: making choices that serve your health, not rigid rules that create stress. The healthiest rhythm is the one that harmonises with both your biology and your lifestyle.

Catherine Norton 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. Should we eat dinner earlier in winter? Why timing might matter more than you think – https://theconversation.com/should-we-eat-dinner-earlier-in-winter-why-timing-might-matter-more-than-you-think-269559