Australia boosts shark-spotting drone coverage at Sydney beaches

AFP
June 27, 2026

Australian lifesavers are expanding shark-spotting drone coverage in New South Wales – Copyright AFP/File DAVID GRAY



Australia will expand shark-spotting drone coverage year-round at beaches across Sydney and beyond from July 1, authorities said Sunday, following a rise in attacks and sightings.

The New South Wales state government said it would invest an additional Aus$34 million ($23 million) in drones, harnessing artificial intelligence and emerging technologies in a “major scale-up” of coverage.

The decision will boost investment in “shark mitigation” in the state to Aus$120 million over the next two years, the government said in a statement.

“While no one can ever promise no shark interactions, this investment is about putting more eyes in the sky so we can spot sharks earlier and give people a clear heads-up when they’re in the water,” said the state’s premier, Chris Minns.

“More drones in the air means we’re getting a better picture of what’s happening offshore and it means we’ll get better at seeing them.”

The enlarged drone monitoring deployment comes after a string of incidents in the state.

A local teacher swimming at Sydney’s popular Coogee beach was mauled by a shark on June 13.

She was left in intensive care in hospital where her family says she has faced multiple surgeries including the amputation of her arm.

In the past week, sightings of a great white shark at Bondi Beach led to beach closures for three days in a row.

A 12-year-old boy died after he was bitten by a shark while playing in Sydney Harbour in January and a bull shark killed a woman swimming at a remote beach north of Sydney in November.

– Shark siren –

In other parts of the country, three divers were fatally mauled in separate incidents between May and June this year — two in Western Australia and the third in Queensland.

Under the expanded drone watch run by the state’s lifesavers, about 70 beaches in New South Wales — including 38 in Sydney — will be monitored every day.

“Surf Life Saving NSW will prioritise beaches with high numbers of swimmers, surfers and paddlers including in Sydney and the North Coast, where shark incidents have become more frequent,” the state government said.

Drone flight hours will be extended from dawn to dusk, and coverage will include popular beaches that are not patrolled by lifesavers.

Australian scientists believe rising ocean temperatures are shifting sharks’ migratory patterns, which may be contributing to an uptick in attacks.

There have been nearly 1,300 shark incidents around Australia since 1791, of which more than 260 resulted in death, according to a database of shark encounters with humans.

Australia has a multi-layered approach to protection against sharks, including old-fashioned nets which have come under criticism for trapping other marine species and their questionable effectiveness.

Lifeguards and drones can also spot the marine predators and alert beachgoers to the danger, often with a siren.

In New South Wales, sharks are lured by bait to “SMART drumlines” where they are then tagged with devices that can be detected when they swim past one of dozens of listening stations along the coast.

That sets off an alarm on a SharkSmart app, giving beachgoers an instant notification on their mobiles and smart watches.

In Idaho, the next generation of US nuclear reactors nears reality

AFP
June 28, 2026

Startup Antares became the first company to run a new-design nuclear reactor in the US in nearly 50 years – Copyright AFP Josh Edelson

A new generation of small nuclear reactors is up and running — or nearly so — in the United States, in what backers are calling a turning point for the industry.

The milestone, made possible by billions in private and government funding, was on display in the middle of the Idaho desert, where a cluster of drab hangars might otherwise go unnoticed.

But the presence of heavily armed soldiers, security checkpoints, and signs warning of radioactivity are anything but ordinary.

It was here, at the Idaho National Laboratory (INL), that startup Antares on June 4 became the first company to run a new-design nuclear reactor in the US in nearly 50 years.

“This is the first real moment in this new nuclear renaissance,” said Jordan Bramble, CEO of Antares.

Aalo Atomics, another participant in the program launched in 2025 under President Donald Trump, is set to do the same in the coming days — also here in Idaho, just hours before a presidential target date: July 4 and the nation’s 250th anniversary.

Meanwhile, on June 18, another startup, Valar Atomics, hit the same milestone in Utah, reaching what is known as criticality — the point at which a reactor can sustain its own nuclear chain reaction.

After developing more than 50 reactor prototypes — including the world’s first to feed electricity into the grid, in 1951 — INL had pressed pause following accidents at Three Mile Island in the US and Chernobyl in current day Ukraine.

Then came the war in Ukraine, followed by the AI boom — putting the energy sector under severe strain and leading both Joe Biden and Donald Trump to revive civilian nuclear power.



– ‘Simpler’ –



Billions of dollars in both private and public funding have already been mobilized to develop these small modular reactors (SMRs) — compact enough that one was transported to the site towed by a pickup truck.

SMRs promise cheaper, faster-to-build nuclear power that can go almost anywhere — from remote military bases to power-hungry data centers. But they have yet to be proven at commercial scale, and some analysts doubt they can compete on cost with wind and solar.

Beyond financial support, the government has put INL’s facilities and staff — who have accumulated nearly 80 years of experience — at the disposal of the selected companies.

The new reactors also use different technology from conventional plants, ruling out the kind of cascading disasters seen at Three Mile Island and Chernobyl, and allowing for far simpler, cheaper construction.

“The whole plant can get simpler. We don’t need to have several feet thick of concrete and steel line containment,” said Yasir Arafat, President and CTO of Aalo Atomics.



– ‘Golden age’ –



Even as the pace has sharply accelerated, Tori Shivanandan, President and COO of Radiant Nuclear, does not want regulatory shortcuts.

The team at the lab, “they hold the line, and we want them to, because ultimately, if we don’t make safe products, we’ll never sell reactors,” she said.

Reaching criticality is not the same as being ready for commercial use.

The reactor designs — whose prototypes operate under a special government waiver — still need to be cleared by the US nuclear regulator, the NRC.

But Energy Secretary Chris Wright, speaking to AFP at a “celebration of the golden age of nuclear energy” in Idaho Falls, was bullish on the timeline.

“We’ll have hundreds by the end of the decade. In fact, our aggressive goal is we will have some of these reactors producing electricity for beneficial use next year,” he said.

If all goes according to plan, Radiant’s first SMRs will go to US military sites, as will Antares’s, while Aalo is targeting data centers.

Nuclear power is also positioning itself as a tool for American influence abroad, with China the only other country operating an SMR.

“Every country I go to asks about the next-generation American nuclear technology. I say…it’s happening right now,” Wright said.

“This will be a massive American export a decade from now,” he added.

Cycling industry bets on smart bikes to boost sales

AFP
June 28, 2026

A man walks past bikes on display at a booth at the EUROBIKE 2026 fair in Frankfurt – Copyright AFP Kirill KUDRYAVTSEV

At the Eurobike trade fair, hopes are high that smart and AI-enabled bicycles can revive an industry that has been dealing with years of flagging sales.

Artificial intelligence, long used in cars and smartphones, is now entering the cycling world in areas ranging from electric motors to safety and services.

At the stand of Avinox, a manufacturer of motors for electric bicycles, the DNA of parent company and Chinese drone specialist DJI has been transposed to cycling.

The motor on display features sensors that continuously monitor the cyclist’s movements and terrain conditions, allowing AI to automatically adjust the motor’s assistance to the pedal drive.

This makes the ride “easier and safer without having to think about it,” Avinox developer Ferdinand Wolf said.

The system even allows a rider to transmit their real-time heart rate so that the e-bike motor modifies the level of assistance as needed.



– Safety alerts –



Elsewhere at the show, there is technology that aims to keep cyclists alive and injury-free.

At Germany’s Canyon, a racing bike equipped with cameras and radars promises to alert cyclists “to the presence of elements they do not necessarily perceive”, company spokesman Ben Hilldson said.

“If a car is parking, the system can anticipate the opening of a door and warn the cyclist,” he said.

The rider would then be alerted via either visual signals on the frame, vibrations in the handlebars or through technology inside their helmet.

Canyon is presenting a helmet fitted with a large visor capable of displaying real-time alerts or receiving an audio signal, depending on the user’s preference.

The products are for now in the prototype stage, Hilldson said.

Canyon is also working with carmaker Volkswagen on a communication system that would allow the bicycle to interact with surrounding cars and other infrastructure, with the launch expected in about three years.

The main obstacle: almost all vehicles currently on the road are not yet equipped to exchange such data.



– Smart networks –



Artificial intelligence is also shaking up services in the cycling industry.

At insurer Linexo “around 90 percent of claims will be handled entirely automatically by the end of the year”, head of the bicycle division Soeren Hirsch said.

Automation handles standard cases, while experts review complex claims and detect fraud, “the only way to keep insurance premiums stable”, he added.

Start-up Wunderfix meanwhile offers repair services linking retailers, customers and shops via an application that allows cyclists to diagnose and, where possible, repair their bicycles themselves.

Some 3,000 service requests have already been recorded this year, the company says.

The rise of AI-enabled and smart bikes has fuelled hopes of a rebound for the business.

The European bicycle market shrunk in 2025 for the third year in a row, with sales dropping four percent to 15.2 million units, according to consultants EY-Parthenon.

“After the boom during the Covid-19 pandemic, the sector has been going through a painful consolidation since 2023: lower sales, high inventories and strong pricing pressure have weighed heavily on many players,” EY-Parthenon analyst Constantin Gall said.

The market is nevertheless expected to stabilise this year before slowly recovering, with revenue forecast at 21.2 billion euros for 2031 — on a par with the record-breaking sales of 2022.

Alongside infrastructure investment, “digital and data-driven offerings” will be a growth-driver, the consultancy said.

Record number of ‘new millionaires’ in 2025, says UBS

AFP
June 30, 2026

Last year, nearly a million people became dollar millionaires – Copyright AFP JUAN MABROMATA


Global personal wealth surged in 2025, with a record number of new millionaires, Swiss bank UBS said Tuesday.

Last year saw nearly one million people worldwide become US dollar millionaires — the equivalent of 2,600 people a day, according to the bank’s estimates.

The United States accounted for almost half of new millionaires in 2025, adding more than 440,000 individuals, followed by China, Japan, Germany, Britain and France, which each count more than two million millionaires in total.

Switzerland’s biggest bank, which is among the world’s largest wealth managers, produces the annual UBS Global Wealth Report, which assesses personal wealth trends.

It covers all financial and non-financial assets, primarily property, minus debts, with asset values converted into dollars.

“The real story is one of continued expansion: more people moving up the wealth ladder,” the report said.

“The gains… point to a world that kept building wealth, deepening its affluent population and extending a long-running upward trend.”

UBS said that in 2025, global personal wealth rose by 10.8 percent in dollar terms, significantly outpacing growth seen in 2024 (4.6 percent) and 2023 (4.2 percent).

Wealth growth was strongest in Europe, the Middle East and Africa, at 17.5 percent — helped by a weaker dollar — followed by the Americas at 8.5 percent. Asia‑Pacific recorded growth of 5.9 percent.

Over half of global personal wealth remains concentrated in the United States and mainland China combined.

Since 2020, South Korea has led growth in real average wealth per adult across the 56 analysed markets, with gains above 50 percent. There have been increases above 25 percent in Croatia, Norway, Latvia, Taiwan and Bulgaria.



– 1.5% are millionaires –



In terms of real average wealth per adult, Switzerland leads the way on $910,382, followed by the United States ($696,277), Luxembourg ($654,732), Hong Kong ($648,267) and Australia ($616,306).

The report said 42 percent of the world population had assets worth less than $10,000; and 41 percent have assets worth $10,000 to $100,000.

At the top end, 15.3 percent have net assets worth $100,000 to $1,000,000, and in the top bracket, 1.5 percent have more than a million dollars.

The report noted that being a millionaire did not mean having a million dollars in the bank, with owner-occupied property representing the single biggest asset for most people up to millionaire level. So rising property values were propelling people into millionaire status.

As for billionaires, in 47 markets covered, UBS counted 3,302 in April — an increase of 383, or 13.1 percent on the last report.

More than 1,000 live in the United States, with 562 in China and 211 in India.

“We count 18 individuals with wealth situated between $50 and $100 billion and a further 19 with assets above $100 billion, 15 of which are based in the United States,” said UBS.

World Bank to phase out lending to China by 2031

AFP
June 30, 2026

World Bank lending to China peaked at $2.42 billion in 2017, but has fallen since then, reducing to $750 million in 2025 – Copyright AFP Eric BARADAT

The World Bank will phase out its lending to China by 2031, according to the organization’s new country partnership framework, a source familiar with the matter told AFP on Tuesday.

The source confirmed an earlier report of the development by the Financial Times.

“China has made significant development advances over the past several decades — progress that the World Bank and others have supported,” said a World Bank official familiar with the matter, speaking on condition of anonymity.

“Now we are reaching a new phase of our relationship, reflecting that reality.”

World Bank lending to China — the world’s second-largest economy — has steadily declined in recent years as the Asian giant saw explosive growth and a reduction in poverty indicators.

In his first term in office, US President Donald Trump demanded that the World Bank stop lending to China entirely, as he adopted a more aggressive approach to Washington’s chief economic rival.

Trump has maintained that tone in his second term, but has not specifically repeated that demand.

World Bank lending to China peaked at $2.42 billion in 2017, but has fallen since then, reducing to $750 million in 2025.

China also contributes funds to the World Bank’s International Development Association (IDA) pool for the world’s least developed countries, with its $1.5 billion under the latest replenishment round making Beijing the fifth-largest donor.

“The World Bank’s role is shifting from lender to knowledge partner, in line with China’s development trajectory,” said the World Bank official.

On June 16, the World Bank announced a similar plan for Poland, planning to reduce loans to zero by 2031 while maintaining technical assistance.

Swedish court orders Google pay nearly $2 bn for favouring its price comparisons

AFP
July 1, 2026

Image: — © AFP Idrees MOHAMMED

A Swedish market court on Wednesday ordered Google to pay some 14.3 billion kronor ($1.46 billion) in damages, plus interest, to price comparison site Pricerunner for promoting its own shopping comparisons in search results.

The Patent and Market Court in Stockholm said “Pricerunner is deemed to have suffered damage as a result of Google having, for many years, unlawfully favoured its own price comparison service.”

The case concerns Sweden, Denmark and the UK and the court said it had ordered Google to pay “just over” 1 billion Swedish kronor, 675 million Danish kroner ($103 million) and £950 million ($1.26 billion) as well as “accrued interest” amounting to approximately 400 million Swedish kronor, 250 million Danish kronor and £300 million.

The court noted that Pricerunner, which filed a lawsuit in 2022 and has since been acquired by fintech giant Klarna, had requested “significantly higher damages, but was not entirely successful in its claim.”

“In many ways, this is a complex and wide-ranging case, and although Pricerunner has not been entirely successful in its claim, the damages awarded are undoubtedly the largest ever ordered in a Swedish competition case,” judge Linda Kullberg said in a statement.

Pricerunner had requested damages totalling 64 billion kronor and an additional 14 billion kronor in interest.

Dan Greaves, head of communications and policy at Klarna, said the ruling “supports a healthier, more competitive market for the way people compare products and services — and that is good for everyone who shops.”



– Antitrust laws –



A Google spokesman meanwhile told AFP: “We don’t agree with the court’s decision, we are reviewing and will consider our legal options. The changes we made to shopping ads back in 2017 are working successfully.”

Pricerunner filed its suit following a European Union General Court ruling that Google “breached EU antitrust laws by manipulating search results in favour of their own comparison shopping services”.

The trial began in October last year, and a ruling was first scheduled for April 15 but was delayed several times due to the complexity of the case.

Originally, Pricerunner said it was suing Google for around $2 billion but said at the time it expected the “final damages amount of the lawsuit to be significantly higher” given that “the violation is still ongoing”.

The European Union General Court in 2021 upheld a 2017 European Commission ruling “that Google had violated competition law by favouring its own shopping service”.

The ruling was then upheld again in 2024 by the EU Court of Justice.



– Continued abuse –



In the Swedish case, the court noted that Google had argued that the violation ended in 2017, while Pricerunner had argued that it continued after 2017 and up until at least 2023.

Pontus Scherp, the lawyer representing Pricerunner, told AFP ahead of Wednesday’s ruling that his team had argued that the changes Google implemented in 2017 were “mostly cosmetic”.

The court found that “Google’s abuse continued for longer than Google had claimed, and that the abuse caused Pricerunner harm.”

However, the court also said that part of the claim had been brought too late and it was also not awarding Pricerunner any compensation for “ongoing harm” after the abuse stopped.

“The judgment shows that competition law provides real protection for companies that suffer harm as a result of competition law infringements committed by dominant firms such as Google,” Scherp said in a comment after the ruling.

Scherp said he thought it should be “highlighted that the court found that Google never ceased its infringement” following the EU commission’s decision.

The damages awarded for the period Google was deemed at fault — 15 years in the UK and 10 years in Sweden and Denmark — were also lower than what Pricerunner had requested, he noted.

Taiwan raids tech firms in China AI chip smuggling probe

AFP
June 30, 2026

The United States restricts the export of its most cutting-edge chips to China – Copyright AFP I-Hwa Cheng

Taiwanese investigators have raided the Taiwan offices of US company Super Micro Computer and two other tech firms, a prosecutor said Tuesday, as part of an expanded probe into the alleged smuggling of Nvidia AI chips to China.

Prosecutors said in May they were investigating the shipment of “high-end” AI servers containing advanced Nvidia chips to China, Macau and Hong Kong, in violation of US export controls.

Nine people are now under investigation, up from three previously, Huang Sheng, head prosecutor in the Keelung Prosecutors Office, told AFP.

They are accused of forging documents so they could ship roughly 50 servers made by Super Micro Computer to China.

Some of the servers were cleared by Taiwan customs and sent to China via Japan, an official previously told AFP on the condition of anonymity.

Twelve sites were raided on Monday as part of the probe, the prosecutors office said in a statement.

They included the homes of six people and offices of the companies they worked for — Nasdaq-listed Super Micro Computer and Taiwan-listed firms Albatron Technology and Chief Telecom.

The United States restricts the export of its most cutting-edge AI chips to China, partly over concerns the technology could be used by Beijing’s military.

But it is not a criminal offence in Taiwan — a situation lawmakers and experts say needs to change — with Taiwanese prosecutors relying on other laws to go after offenders.

Lawmaker Chung Chia-pin, who belongs to President Lai Ching-te’s Democratic Progressive Party (DPP), plans to propose an amendment to the Foreign Trade Act to include a “mainland China semiconductor chip clause” that would make exporting chips there illegal.

Chung told AFP Tuesday that a loophole in the law was created under former president Ma Ying-jeou, who belongs to the Kuomintang party, and successive DPP-led governments have failed to close it.



– Shares fall sharply –



Top-end chips made by US titan Nvidia — the world’s most valuable company — are used to train and run AI systems.

In response to Washington’s export restrictions, China has been accelerating efforts to develop its own AI chips and break away from reliance on US hardware.

This month, Taiwanese Deputy Economic Affairs Minister Ho Chin-tsang said Taiwan and the United States “will work to implement our shared export control goals”, but the government has not provided details.

Chris McGuire, an expert on China and AI at the US-based Council on Foreign Relations, said chip smuggling was a “really significant problem” in Taiwan and Southeast Asia.

“It’s really, really important that allies align with the United States on all of these policies and also legal authorities,” McGuire, who worked at the National Security Council under former US president Joe Biden, told a forum in Taipei this month.

“It’s not a criminal violation in Taiwan to export AI chips to China, obviously it is under US law, but it’s not under Taiwanese law. That needs to change, right?”

Super Micro Computer, Albatron Technology and Chief Telecom have said separately they are cooperating with investigators. Their shares have seen sharp falls this week.

Prosecutors say it is too early to know if the case is linked to a Nvidia chip smuggling case involving Super Micro Computer employees in the United States.

A US indictment unsealed in March showed employees of the company allegedly raked in billions of dollars diverting Nvidia AI chips to China in breach of export controls.

Op-Ed: Military AI is reinventing war at the expense of all previous theories and many sacred cows

Paul Wallis
June 27, 2026

DIGITAL JOURNAL

A Ukrainian serviceman of an air reconnaissance squad of the 45th Brigade carries a Leleka reconnaissance UAV after its landing at a position in Donetsk region – Copyright AFP Genya SAVILOV

If you give someone a weapon nobody knows quite how to use, nobody can be sure what’s going to happen. A raging torrent of military AI news is reshaping everything from logistics to intelligence to basic tactics daily. It’s turning strategy into a multidimensional universe in its spare time.

Certainties have become gaping holes in strategy. AI can be trained in operational modes that make old-style strategies redundant. These changes create new vulnerabilities requiring innovative countermeasures. Technologies may have to be designed from scratch.

The scale and scope of strategic targeting are also now much wider, making defence that much harder. A single coordinated AI strike could and will target infrastructure, communications, economic networks and put an entire country out of business in a few hours.

Mass production of AI weapons and agents adds further dimensions. These systems are coming online fast, and at the operational level, they work. Ukraine is using various AI-enabled systems very effectively. Even at this early stage, a search of “Ukraine AI warfare” generates vast amounts of related information.

“Asymmetric warfare” is one thing, but military AI is a whole new type of asymmetry. In asymmetric warfare, the capabilities of forces differ significantly. With military AI, those capabilities can change overnight.
Military AI in sandbox mode

There are few things less popular with the world’s military than saying warfare is simply becoming a video game. From the AI perspective, however, it can’t really be anything else.

Gamers in the meantime are finding out the daunting fact of how effective upgraded AI can be in gaming scenarios. In one notable case, the AI not only deployed competently in strict best practice defensive positions but was also able to counterattack. That outcome is textbook good tactics. The after action report in that game was downright gruesome. Even at the sim level, this long-established game suddenly became a very tricky affair, and the highly experienced human player lost.

The sandbox analogy for military AI is perhaps too appropriate. That makes it even more dangerous. You can test scenarios. You can model operations. You can add elements to your combat force and your tactics. You can force asymmetry on your opponent.

This generation of military AI is a mix of both coexisting conventional systems and autonomous units. It’s barely the beginning of the beginning. AI is driving a top-to-bottom redesign of conventional systems while new platforms create new capabilities. The number of new degrees of difficulty is exploding.
Military AI at the geopolitical level

If grunt-level AI is becoming so much tougher, at the geopolitical level, it’s more like 5-dimensional chess. The China vs US dichotomy is a case in point. Even the rules of AI governance are under a type of top-level scrutiny that those rules can barely meet.

With this situation come a range of operational realities. China’s current visible military AI profile seems to be mirroring the US, but only to a point. At a recent unveiling of robotic military tech, the familiar robot dogs, drones, and UAVs were highly visible. Also visible were “robot soldiers”, the much predicted and barely adequately described antithesis of human warfare.

At the moment they look more like updated Terra Cotta Warriors, but they could well be a factor in real operations. That’s also a form of sabre-rattling, and it’s making a deep impression geopolitically. You can almost hear the budget calculations scrambling to adjust to unknown threat levels.
The first casualty of AI war is the old military industrial complex theory

This level of military AI is really just camera fodder at the moment, but it’s readying the geopolitical market for massive changes. It’s also reshaping the industrial base, down to the component level. The military industrial complexes of old are becoming fully automated industrial complexes.

The problem for the military industrial complexes is that the old industrial certainties are gone, and they won’t be coming back. AI may be a gamechanger, but they’ll have to guess which games they’re playing.

That means:

Forget the old super-expensive tech scenarios. The Ukraine war has shown how fatal expensive targets are. This is no longer “measure vs countermeasure”. It’s about survivability.

A seventh-generation air capacity will be essential, and soon. The sixth-generation fighters are already under stress from incoming AI capabilities. UAVs may be the only option, and combat effectiveness is the only criterion.

Land warfare will never be the same. Imagine AI deploying minefields, loitering munitions making areas uninhabitable, AI-guided bullets, and more. Old tech is very unlikely to be able to manage situations it was never designed for. Ammo types are likely to be the first on the scrapheap as the military tech evolves.

Naval warfare needs to evolve, fast. Big targets and ridiculously long lead times to production are deathtraps for modern navies. There are simply too many problems with the old methods.

Military intelligence now includes everything and everywhere. Remember that drones used to be toys. Now, they’re essential military hardware. Everything is their target. Cyberespionage and AI agents are redefining and redirecting military intelligence on an exponential scale across whole sectors. Any form of tech that can be adapted to military use is now an intelligence issue.

Modern military logistics can’t use ponderous supply chains. Ukraine has shown that onsite capacity is critical for maintaining and supporting combat capabilities. Russia has shown that these supply formats are obsolete and completely unworkable. The sedate pace of military supply can’t work. Modern combat needs support to deliver ASAP.
The perspective of those on the receiving end has also changed

Nothing is AI-proof. Control of military AI is a true guessing game. Autonomous AI may make decisions that aren’t under the effective control of military organizations. There may or may not be Off switches for these types of AI.

AI may also not have any connection with the rules of war. Does AI take prisoners? Probably not. Definitely not, if it doesn’t know how. Will AI respect civilian targets? Maybe, but how, and can anyone be sure it will?

What about collateral damage? It was a big issue, and now it’s inevitable. Higher lethality will apply indiscriminately, despite any efforts otherwise. How do you tell an AI not to bomb a specific target when it’s on seek-and-destroy missions?

Typically in modern wars, civilian casualties are much higher than military. The realities of AI warfare are much more dangerous. Let’s stop being smug about dollars and focus on survival.

____________________________________________________

Disclaimer
The opinions expressed in this Op-Ed are those of the author. They do not purport to reflect the opinions or views of the Digital Journal or its members.

Op-Ed: Canada tries to promote AI safety and equity at the UN

Paul Wallis
June 29, 2026

DIGITAL JOURNAL

Image:— © AFP

In the face of constant AI revelations and its ever-spreading economic effects, Canada is pushing for global action and consensus on AI rules, safety, and equity.

At the same time, a clear AI gap between wealthy and developing countries is widening. Canadian Ambassador to the United Nations David Lametti made the point succinctly to the Canadian press recently:

“The UN remains critically important, (it) remains perhaps the only institution in the world that can convene that kind of discussion on a more or less equal footing between Meta, Amazon Web, Microsoft, Apple and Google — and all of these other countries.”

The World Economic Forum (WEF) has come to more or less the same view from a different direction based on AI’s impact on global business and wider applications on trade. Loans and finance are particularly sensitive, and they’re already creating issues for developing countries. Exclusion from credit markets based on AI loan decisions alone creates capital problems that can stymie business growth and development.

The WEF estimates that “The global trade finance gap stands at $2.5 trillion, concentrated almost entirely in the emerging market corridors where transaction data is thinnest.”

This is where the “equity” issue espoused by Canada becomes critical. Such a huge shortfall in capital effectively shuts down trade almost entirely due to AI procedural behaviour in managing finance.
A tricky reality for the developing world

The UN’s own data supports this view with some added caveats. The many obvious opportunities for developing countries include significant burdens on their capital and raise many questions regarding even their capacity to develop.

According to the UNDP Regional Bureau for Asia and the Pacific (RBAP):

The infrastructure investment needed to advance AI is already a magnitude more than the current SDG financing gap. This is why widespread AI adoption must be treated as a central development objective.

The adoption and application realities are much more demanding than for the least developed countries. The UN is advocating AI as a critical development asset, but 1.2 billion people are being left behind.

Canada’s broad-brush UN approach to these problems may be the only realistic way of making the issues visible or even actionable. The vast mix of problems simply has no coherent profile. The world’s priorities are elsewhere. The slow drip of headlines for AI adoption in developing countries are a predictably chaotic and ineffectual sprinkle of anecdotes compared to the high-profile Big Tech news that drowns out all other considerations.
A polarized trade environment and China’s AI push into developing markets

The much bigger picture is a destructive form of technological polarization at the worst possible time. Geopolitics has created a series of minefields. Globalization is the world’s trade reality, and its reach is total throughout all economies.

Apart from Canada, the West isn’t providing much acknowledgment, much less leadership, for the rest of the world. China, however, was already promoting a “Digital Silk Road” earlier this year. According to the East Asia Forum, “This ambitious undertaking aims to link nearly 150 countries across Africa, Asia, Europe and beyond through a modern network of rails, roads and ports.”

The Chinese initiative is backed up by DeepSeek and its built-in business support networks. China is deploying AI and automation at an extraordinary speed.

The contrast between China and the US AI couldn’t be sharper, and it’s a grim dichotomy:

US trade policies have effectively created an obstacle course. These policies and the often-negative reactions to them now define global trade in the short term.

The Chinese open-source AI approach is the exact opposite of the dogmatic US proprietary stance. Open source is by far the easiest option for introducing AI.

Big capital investment in AI is a physical impossibility for many nations. The US may have priced itself out of some markets entirely.

Adopting a whole new class of technology and its support systems without even basic guidelines is equally impossible and unrealistic.

At ground level, Chinese AI and automation are being deployed across all industrial sectors, whereas US AI is far more high-end and exclusionary.

A US or China version of standardization of AI practices is therefore likely to be a major sticking point for the world as a whole. The two environments aren’t compatible.
Canada to the rescue? Maybe, and at least it’s a start.

By taking on the seemingly rather thankless task of addressing the realities of global AI adoption, Canada is lifting a very heavy load.

At the risk of euphemism, rules, fairness, and equity haven’t exactly been hot topics for Big Tech, Big Money, or anyone else. The US, in particular, has been actively trying to block AI regulation at the Federal level.

For global AI to work at all, standards are essential. Consider the possibilities of AI generating trade disputes and exacerbating the current situations. Equity and common practice are the basis of trade.

Failure to address these issues simply guarantees future problems, disputes, and dysfunction on a global scale. The world should be paying attention.

________________________________________________

Disclaimer
The opinions expressed in this Op-Ed are those of the author. They do not purport to reflect the opinions or views of the Digital Journal or its members.

Canada’s AI buildout left out the water bill

Digital Journal Staff
July 2, 2026 

Photo by Getty Images on Unsplash

Canada wants a lot more AI compute on home soil, but it hasn’t decided how to account for the water that keeps it running.

The federal “AI for All” strategy, released June 4, projects the country will need 5.5 gigawatts of commercial compute over the next four years, and commits to 850 megawatts of domestic capacity by 2030. But there isn’t much detail on how water and land use will be measured or managed.

On power, Canada has a real edge. According to The Conversation, more than 83% of the grid runs on low-emission sources, which can cut a data centre’s operating emissions by up to 90%.

Water doesn’t get the same answer.

Wafr Technologies announced on July 2 that it raised $100 million toward a $300-million goal to build an AI research lab in Canada. This lab would be anchored on a proprietary cooling technology the company says cuts data centre water use by up to 95% and cooling power draw by up to 80%.

According to their press release, a typical data centre uses up to 10 million litres of water per megawatt annually, with cooling consuming 30 to 45% of total electricity load. As workloads grow, more the real cost and the compliance questions come from the data centre underneath, instead of the model itself.

So far, Wafr’s technology has been demonstrated in India and Dubai, but not yet at Canadian scale. The lab is planned, not built, and the raise is a third of the way there.

“Our vision is to build a globally recognized AI research lab in Canada and be a leader in how we can reduce the impact to water and energy,” says Bikram Singh, Wafr Technologies Co-founder and CEO.

When CIOs or CISOs sign a cloud or data centre contract, asking what the provider does on cooling and water is a fair question. It might even be one your board or CFO will get to before you do.

Final Shots

Canada’s AI strategy leans on the clean grid to handle emissions but leaves water use largely unaddressed, so the resource question falls to whoever runs the compute.

Cooling can eat 30 to 45% of a data centre’s electricity load and millions of litres of water per megawatt, which puts efficiency in the operating budget, not the sustainability report.

Cooling and water use are now fair questions on any compute contract, and the answer is the kind of thing a board or CFO might reach before the technology leader does.

A quarter of Canadian leaders don’t see AI coming for them

Digital Journal Staff
June 30, 2026

Photo by Andrea Piacquadio on Pexels

A quarter of Canadian business leaders believe AI will have minimal impact on their organization over the next four years, according to BDO Canada’s AI Vision Report, released last week. That finding might be the most telling number in the report.

Anyone managing an enterprise technology stack has watched AI arrive through vendor updates they didn’t initiate.

Gartner projects that by 2028, a third of enterprise software applications will include agentic AI capabilities, up from less than 1% in 2024.

These features are being built into the platforms Canadian businesses already run, whether they asked for them or not. The leaders who think AI won’t touch them may already be running it.

The report, based on a BDO Canada survey of 520 Canadian business leaders who are members of the Angus Reid Forum, also found that 46% are experimenting with AI without achieving measurable ROI and only 18% have embedded AI into workflows and operations.

Digital Journal has reported the same pattern over the past two weeks. AI projects are clearing launch and missing ROI, and in some cases organizations are pulling live agents back into the sandbox after deployment.

“The next gap will not be between organizations using AI and not using AI. It will be between those redesigning work around AI and those funding disconnected pilots,” says Bill Syrros, national AI leader at BDO Canada.

The harder question is how many organizations know which side of that line they’re on. The report found only 18% have embedded AI into workflows and operations. The rest are either experimenting, holding back, or facing AI through the software they already use.

The BDO data found 46% of respondents can’t prove value from the AI they chose to adopt. Another 27% expect minimal impact, even as BDO says AI is becoming harder to separate from the enterprise software companies already use.

You can’t measure what you chose to buy if you can’t see what you’re already running.
Final shotsBDO Canada surveyed 520 business leaders and found 27% expect AI will have minimal impact on their organization over the next four years. Gartner projects a third of enterprise software will include agentic AI capabilities by 2028.
The survey found 46% of respondents experimenting with AI without measurable ROI, and only 18% have embedded AI into workflows and operations.
Every AI board update should include what is running, who owns it, and what changed because of it.

Physicists and Claude “collaborate” to prove a ten year old jamming conjecture

Nobel laureate Giorgio Parisi and physicist Francesco Zamponi, with the help of artificial intelligence, solve a mathematical problem that had remained open for more than a decade and document the entire process

Sissa Medialab

Facebook

WeChatBlueskyMessageWhatsAppEmail

A mathematical problem that had remained unsolved for more than ten years in the physics of complex systems has finally been resolved through an unusual collaboration: one involving two theoretical physicists and an artificial intelligence system. In a study published in the Journal of Statistical Mechanics: Theory and Experiment (JSTAT), Giorgio Parisi, Nobel Prize winner in Physics, and Francesco Zamponi, physicist at LaSapienza University of Rome, show how the AI model Claude contributed to finding the proof of a mathematical relation that had resisted researchers’ efforts for years. 

Beyond its scientific significance, the result offers a concrete glimpse into how artificial intelligence is transforming the work of researchers.

In physics, jamming describes the formation of a kind of “traffic jam” of particles: a system that is initially fluid suddenly becomes rigid while remaining disordered. Originally introduced to describe materials such as foams and granular matter, the concept has proved surprisingly general and is now also used in fields such as neuroscience and artificial intelligence.
In 2014, Giorgio Parisi, Emeritus Professor at LaSapienza University of Rome and recipient of the 2021 Nobel Prize in Physics, Francesco Zamponi, Professor of Physics at LaSapienza University of Rome, and collaborators developed a theoretical description of jamming and noticed a surprising relationship: two mathematical parameters of the model, denoted by a and b, always added up to one, as numerical calculations showed with extraordinary accuracy.

A surprising relationship

This relationship, explains Zamponi, co-author of the new study together with Parisi, yields the same physical laws obtained through a different theoretical approach to jamming developed almost simultaneously by French physicist Matthieu Wyart (EPFL, Lausanne). In other words, it suggests that two very different ways of describing the phenomenon actually lead to the same conclusions.

The result emerged clearly from numerical calculations from the very beginning, but no one could explain why it was true. For years, researchers searched for a mathematical proof of the relation, convinced that some deeper structure of the theory lay behind its apparent simplicity.

A persistent obsession

After several unsuccessful years, the problem gradually faded into the background. Not for Giorgio Parisi, however. “It really bothered him that we had never managed to prove it,” Zamponi recalls.

When the first generative AI models began to appear, Parisi identified this old problem as an ideal test case. Claude was chosen because it “seemed to have somewhat more advanced mathematical reasoning abilities,” says Zamponi.

The problem, after all, was well defined: a clear conjecture, relatively simple mathematics, and an answer that was known numerically but had never been formally proven.
The prompt given initially was not to find the proof. Parisi asked the model to reproduce the numerical calculations developed by the group more than a decade earlier, in order to understand how far it could go in tackling a real mathematical problem.

Once Claude was able to reproduce the result, the researchers’ next question came almost naturally: if a+b equals one, can you also prove why?

“Quite quickly, Claude came up with an initial idea that was essentially correct,” says Zamponi.

The proof still contained errors and required several rounds of verification and revision by the authors, but the underlying intuition turned out to be the right one.
Yet the surprise was not only the AI’s result. For years, the researchers had been searching for a deep explanation of the relation, imagining that it concealed a new mathematical structure or an unknown symmetry. “We were hoping this would reveal some new understanding of the equations,” Zamponi explains.

Instead, the solution turned out to be much simpler: “The answer was right there, and we simply hadn’t seen it.”

The proof therefore confirms that two very different theoretical approaches to jamming, developed independently by Parisi and collaborators on the one hand and Wyart and collaborators on the other, do in fact lead to the same physical laws.
 

---

Method of Research

Data/statistical analysis

Article Title

A proof of an identity for the critical exponents of jamming

Article Publication Date

1-Jul-2026