science

Pakistan-built robots serve taste of future to Karachi diners

BY SAMEER MANDHRO

  • For now, the robots are proving a hit with families, who flock to the restaurant to watch the machines in action.
  • Wheeling its way between dining booths at a neon-lit restaurant in Pakistan's Karachi megacity, a robot waitress drops off a fresh pizza to an awaiting family.
  • For now, the robots are proving a hit with families, who flock to the restaurant to watch the machines in action.
Wheeling its way between dining booths at a neon-lit restaurant in Pakistan's Karachi megacity, a robot waitress drops off a fresh pizza to an awaiting family.
While global buzz is growing around agile humanoids made in the United States, China and elsewhere, robotic staff are a rare sight in the South Asian country.
"Welcome to Sicilian. My name is Robia," the waitress says in an electronic voice, looking down at diners through blinking red LED eyes.
Robia and her colleague Jojo are the DIY creations of Pakistani engineer Syed Osama Aziz, who is bringing a sci-fi reboot to Karachi's storied dining scene.
They have spherical white heads and dark visors, similar to a motorcycle helmet, and sport black shawls around their neck in the style of a dupatta worn by women in the region.
Aziz, who also established The Sicilian restaurant, developed the machines using mostly locally manufactured parts, designing their circuit boards and programming them in his own lab.
"The equipment, the things needed to make robots, are not so easily available here, so we had to arrange them," he told AFP.
The duo were developed through his robotics company, Cobot, with help from friends and some electronic components imported from abroad.
In China and other East Asian countries especially, robots are an increasingly common sight, shuttling food to customers in chain restaurants and taking on other jobs from cleaning to greeting hotel guests.
But they remain a novelty in Pakistan, where Aziz's home-grown creations are attracting curious diners.
When they arrive at a table, the customers must either take the food off their serving tray themselves or wait for a human member of staff to assist them.
 

Excited customers

 
Aziz, whose workshop manufacturing the robots is located in Multan in Pakistan's central Punjab province, hopes to make his creations more sophisticated in the future.
"Our plans are to add camera-based processing and image-processing features," he said.
"That means the robot will also recognise you by looking at you, and it will remember what you ordered last time, and accordingly it will be able to give you better suggestions."
For now, the robots are proving a hit with families, who flock to the restaurant to watch the machines in action.
"The best thing about it is that it introduces itself," excited customer Ahmer Zameer told AFP, explaining the robot also "tells you what things you can order".
Children reach out to touch the robots while their parents snap photographs and selfies.
Young customer Sonia Dilosker said she was impressed by the androids' appearance and preferred their approach to serving customers, "because it is a machine and without any emotions".
Restaurant staff say the novelty has brought a surge in customers -- some travelling from across the sprawling port city to meet the robots.
"Our business has grown. People are coming from far away for this. They have seen a robot for the first time here," staff member Hellen told AFP.
"People are excited to see this."
sma/csp/kaf

literature

Cyberattack shuts down internet fan fiction library

BY MANON BELIOT

  • Fan fiction is where many future authors learn their craft and where readers disover new writers.
  • One of the world's most popular fan fiction libraries has been dark for over a week after a cyberattack forced it offline, leaving millions of fans without access to their favorite stories.
  • Fan fiction is where many future authors learn their craft and where readers disover new writers.
One of the world's most popular fan fiction libraries has been dark for over a week after a cyberattack forced it offline, leaving millions of fans without access to their favorite stories.
Since October 29, anyone trying to reach Archive of Our Own -- known to devotees as AO3 -- has been met with a message saying the site is down for maintenance. 
"It's all over my social feed. People are in withdrawal," Johanna Sarriot, a writer and co-founder of pop-up romance bookstore Love Love Books in Los Angeles, told AFP. 
Fan fiction refers to stories written by amateur authors using characters and worlds from already-existing books, films or TV shows.
Contributors rewrite endings, invent romances or give minor characters their own adventures, then share them online for free.
Once a niche hobby, it has become a launchpad for the publishing industry, generating breakout smashes like E.L. James's "Fifty Shades of Grey," which began as "Twilight" fan fiction, and Ali Hazelwood's bestseller "The Love Hypothesis," which sprang from the "Star Wars" universe.
Launched in 2008, AO3 has become the genre's main home, with more than 11 million registered users and nearly 18 million works stored on the platform.
In 2019, it won a Hugo Award for Best Related Work, one of science fiction's most prestigious prizes.
"We wanted to make sure that there was a place for fans to engage in non-commercial creativity in ways that wouldn't barrage them with ads, censor them to placate advertisers, sell their data, or do anything but provide a space for fandom," said Harvard Law School professor Rebecca Tushnet, who helped found the Organization for Transformative Works (OTW) which runs the site.

'AO3 is not gone'

Romance fans have been among the hardest hit by the outage.
"Romance readers read really fast and they read a lot," explained Sarriot. 
For the bookseller, what is at stake goes beyond a website. Fan fiction is where many future authors learn their craft and where readers disover new writers.
"Writers who were able to cut their teeth on fanfiction, and just write as much as they wanted with all the joy and no constraints, can lead to really great novels," she said.
OTW says it does not believe the user database, which holds usernames, email addresses and encrypted passwords, was stolen, but cannot rule it out.
"AO3 is not gone! There is no data loss and we look forward to restoring the site as soon as possible," Tushnet added.
The non-profit says it hopes to have things up and running again within a month.
For web developer Yohan Mebrouk, the timeline is a sign of how serious the attack was.
The delay could reflect limited resources, he explained, or the need for a full software upgrade, which he called "an enormous job."
Well-run sites keep backup copies of their databases, he added. "If they did things right, everyone will get everything back."
In the meantime, Sarriot is happy to point stranded readers elsewhere.
"Come to Love Love Books and check out a romance novel!" she said, pointing to her shelves full of paperbacks.
mbt/hg/mlm

politics

Tech stocks stage partial rebound while dollar gains after US jobs data

  • After slumping 5.5 percent on Thursday, the tech-focused Nasdaq clawed back 1.1 percent on Friday.
  • Tech shares rose Friday, recovering a fraction of losses from the prior session's rout as the dollar rallied after better than expected US jobs data.
  • After slumping 5.5 percent on Thursday, the tech-focused Nasdaq clawed back 1.1 percent on Friday.
Tech shares rose Friday, recovering a fraction of losses from the prior session's rout as the dollar rallied after better than expected US jobs data.
Investors welcomed a partial recovery for the tech-heavy Nasdaq after its plunge on Thursday, which sent Asian and European markets tumbling.
But the Nasdaq technically entered a correction after dropping more than 10 percent over the past month, reflecting worries that sweeping victories by Democrats in this week's US elections would slow or derail the artificial intelligence data center buildout.
After slumping 5.5 percent on Thursday, the tech-focused Nasdaq clawed back 1.1 percent on Friday.
"Markets could receive an additional boost if the Democrats soften their tone on AI," said a note from Jose Torres of Interactive Brokers. 
In a CNN interview, Texas Governor-elect Gina Hiojasa, among the newly elected leaders that spotlighted the issue during the campaign, said communities need more say on if and where data centers get built.
"We have always been a pro-business state in Texas, but what changes now with me as governor is it has to be a partnership that benefits Texans," she said. "It is not that I hate data centers; I hate the way they are taking advantage of Texas, and that stops when I'm governor."
Investors were encouraged by a report showing that the US economy added 165,000 jobs last month, far more than analysts had expected.
While the US jobs data -- including a dip in the unemployment rate to 4.1 percent -- reassured investors on the strength of the economy, it also cleared the way for the US Federal Reserve to raise interest rates.
Investors widely expect the Fed to raise interest rates by a quarter percentage point at its next meeting in December to bring down stubborn inflation.
Those expectations were also reflected in the foreign exchange market, with the dollar rising against the euro and other major currencies.
"The prospect of a relatively more aggressive Federal Reserve has moved back into focus, making it increasingly difficult for the euro to recover ground consistently," said a note from Julian Pineda, a market analyst at Forex.com. 
Rising interest rates would further complicate the calculus for tech firms that have propelled markets worldwide over a four-year bull run.
"Big worries are colliding -- the build-out of data centers looks set to slow, which risks shredding the spending spree underpinning the boom, right at the time that investors have become super nervous about returns on this massive investment," said Susannah Streeter, chief investment strategist at Wealth Club.
The Tokyo and Seoul stock markets closed out their weeks with losses of six percent following the Nasdaq's 5.5 percent collapse on Thursday, its worst showing since April last year when Trump unveiled his wide-ranging global tariffs.
On Seoul's Kospi stocks index, chip giant SK hynix slumped 10 percent and its rival Samsung lost 6.4 percent.
The London, Paris and Frankfurt stock markets finished down, despite featuring few technology firms.

Key figures at around 2120 GMT

New York - Dow: DOWN 0.5 percent at 51,285.73 (close)
New York - S&P 500: UP 0.3 percent at 7,376.86 (close)
New York - Nasdaq: UP 1.1 percent at 24,579.42 (close)
London - FTSE 100: DOWN 1.0 percent at 10,673.94 (close)
Paris - CAC 40: DOWN 0.8 percent at 8,139.00 (close)
Frankfurt - DAX: DOWN 1.9 percent at 25,820.36 (close)
Tokyo - Nikkei 225: DOWN 6.0 percent at 57,434.77 (close)
Seoul - Kospi: DOWN 6.0 percent at 5,059.69 (close)
Hong Kong - Hang Seng Index: DOWN 1.3 percent at 24,481.86 (close)
Shanghai - Composite: DOWN 1.1 percent at 3,649.18 (close)
Brent North Sea crude: UP less than 0.1 percent at $98.32 per barrel
West Texas Intermediate: UP 0.3 percent at $90.83 per barrel
Euro/dollar: DOWN at $1.1138 from $1.1219 on Thursday
Pound/dollar: DOWN at $1.3350 from $1.3450
Dollar/yen: UP at 158.46 from 157.70 yen
Euro/pound: UP at 83.43 from 83.42 pence
burs-jmb/sla

AI

Dutch activists invade site of Microsoft's planned data centre

  • A coalition of campaign groups called "Stop Dirty Data Centers" has been pushing for a pause in the construction of such sites and called for action on Friday and Saturday in European countries.
  • Several hundred activists opposed to the construction of a Microsoft data centre in Amsterdam occupied the site on Friday, responding to a call for a two-day protest across Europe. 
  • A coalition of campaign groups called "Stop Dirty Data Centers" has been pushing for a pause in the construction of such sites and called for action on Friday and Saturday in European countries.
Several hundred activists opposed to the construction of a Microsoft data centre in Amsterdam occupied the site on Friday, responding to a call for a two-day protest across Europe. 
The collective Geef Tegengas, which campaigns for climate and justice, and the environmental group Fossielvrij entered the construction site in the Sloterdijk district, halting work for the day.
"It's so many things, so many struggles that come together in this data centre," Geef Tegengas spokesman, Nathan Timmer, told AFP.
"This thing will use as much energy as all of the households in Amsterdam combined... this thing will use a ridiculous amount of water."
Timmer, 30, also argued that the centre would be used  "to facilitate AI warfare" and what he called the "genocide" in Gaza.
The activists made it on to the site by scaling barrier fences despite a police presence.
Once inside they shouted political slogans and sprayed anti-Microsoft messages on walls under construction.
Dutch news agency ANP reported that police arrested between 150 and 200 protesters. Other activists quit the site voluntarily.
AI data centres require large amounts of electricity -- placing increased pressure on grids often still reliant on fossil fuels -- and, in many cases, water to cool down hot computing equipment.
A backlash is growing worldwide against their environmental impact, and there are also concerns over noise, pollution, their appearance and how many jobs the facilities actually create.
A coalition of campaign groups called "Stop Dirty Data Centers" has been pushing for a pause in the construction of such sites and called for action on Friday and Saturday in European countries.
Microsoft did not immediately respond to a request for comment.
sh/jj/ach

US

An AI reckoning? Investors brace for potential fallout

BY FLORIAN CAZERES

  •   - IPO anxiety -   The tech-heavy Nasdaq index had already wobbled before this week and now is officially in a correction, having fallen around 12 percent over the past month.
  • "Correction" is not the word investors want to hear, but the stampede out of AI and technology stocks this week has heightened doubts about the sustainability of the four-year bull run for markets.
  •   - IPO anxiety -   The tech-heavy Nasdaq index had already wobbled before this week and now is officially in a correction, having fallen around 12 percent over the past month.
"Correction" is not the word investors want to hear, but the stampede out of AI and technology stocks this week has heightened doubts about the sustainability of the four-year bull run for markets.
Further declines would have wide-ranging financial and economic consequences amid the massive AI spending, much of it fuelled by companies taking on huge amounts of debt.
AFP takes a look at how tech sector weakness could impact governments and households.
 

The AI motor

 
Outlays on data centres to train and power artificial intelligence models have become one of the main growth engines for the US economy, the world's largest.
"That contributed to 30 percent of growth in the United States in 2025 and 2026," said Bruno Cavalier, economist at the bank Oddo BHF, with spending on data centres surging by 150 percent.
The resulting demand for cutting-edge chips and other high-tech equipment has boosted share price gains across the technology sector.
But the centres also need power sources, intensive cooling systems, and old-fashioned construction, plumbing and electrical work, all providing money and jobs for local communities and beyond.
Opposition has been building over worries about the environmental and human costs of the huge server farms, including air pollution, higher energy costs, water use and noise.
The Democratic wave in this week's midterm elections in the United States raised investor fears of a regulatory crackdown -- prompting the heavy selling on Wall Street on Thursday.
Markets are pricing in the risk that "Democratic victories in the states with the most projects will make construction more expensive and longer to complete", said Kevin Thozet at the asset management firm Carmignac in Paris.
 

American households

 
Share ownership is widespread in the United States, where individual retirement funds invested in stocks and bonds are much more prevalent than government-funded pensions.
"Forty-three percent of household wealth is now exposed to the stock market, an increase of 10 percentage points from 2000," said Cavalier.
A sustained market downturn would directly impact their spending power, especially for retirees, creating a wider economic risk.
And "the richest 10 percent of Americans now account for half of consumer spending", said Alexandre Baradez, an analyst at the brokerage IG France.
"They're the ones who own the most stocks and have benefited the most from the tech stock gains," he said.
A market crisis could crimp their spending as well, meaning it would no longer compensate for the spending pullback among average Americans since the inflation spike provoked by the US war against Iran.
 

Capital access?

 
Rising interest rates complicate the picture, since they lift borrowing costs for the billions of dollars in debt AI firms have taken on to finance the data centre rollout.
"This is not an easy time, with shares falling and high rates on debt markets -- and little chance of any easing moves soon by central banks," Baradez said.
Tighter financial conditions and higher cash needs could make investors wary of putting up more money, or lead them to reduce exposure to the sector.
For Jacques-Aurelien Marcireau at Stilgar, an asset management firm in Paris, American politicians are aware of the risks.
"The truth is that Democrats will be pragmatic: there won't be any general halt to data centre construction," he told AFP.
But a slowdown "could even be healthy, by halting the most vulnerable projects", he said.
 

IPO anxiety

 
The tech-heavy Nasdaq index had already wobbled before this week and now is officially in a correction, having fallen around 12 percent over the past month.
That wiped off $2.5 trillion in market valuation for Nasdaq companies alone.
Investors now question the prospects for the imminent initial public share offering from Claude-maker Anthropic, which had been tipped to surpass SpaceX's earlier this year as the biggest ever.
It has reportedly been looking to raise a whopping $100 billion this month to buy the chips and other equipment for its computing needs, which could have valued the start-up at up to $2 trillion.
Now it may have to settle for less, potentially complicating its race to dominate the market.
Its rival OpenAI has already pushed back its IPO plans to sometime next year.
"Anthropic will be the big test," Baradez said.
fcz/tq/js/jhb

TikTok

EU tells TikTok to change 'harmful' For You feed

BY RAZIYE AKKOC

  • Friday's demand comes after the EU in February told TikTok to change "addictive design" features, and after it accused the company in July of exposing children to risks like cyberbullying and predatory behaviour.
  • The EU told TikTok on Friday to change the "harmful" design of its personalised For You feed or risk large fines under the bloc's digital content rules.
  • Friday's demand comes after the EU in February told TikTok to change "addictive design" features, and after it accused the company in July of exposing children to risks like cyberbullying and predatory behaviour.
The EU told TikTok on Friday to change the "harmful" design of its personalised For You feed or risk large fines under the bloc's digital content rules.
The European Union warned that TikTok's recommendation system exposes children and vulnerable individuals to potentially damaging content by pushing a stream of videos similar to those the user has previously engaged with. TikTok rejected the accusations.
Brussels accused TikTok of violating the EU's Digital Services Act (DSA) over the "harmful 'rabbit holes' effect", which happens when users are fed related videos based on an algorithm, in some cases leading to more dangerous content.
"The more a user interacts with a certain type of content, the more TikTok's 'For You' feed algorithm recommends similar content. This design can be harmful when such endless and repeated exposure involves, for example, content about sadness, isolation or body image," the European Commission said.
Even when users tell TikTok they don't like certain content, the platform's algorithm continues to suggest similar videos.
TikTok has made some tweaks to mitigate the risks but they were not effective, the EU said, demanding more radical design changes.
The commission gave examples of changes the platform could make, such as:
-- rely more on users' explicit preferences, because the EU found that the app's "not interested" button "may not be enough to change the content feed"
-- disable viewing time as an indicator to recommend content
-- increase the accessibility and ease of use of user controls, and regularly test its algorithm.
A TikTok spokesperson hit out at the EU's "series of statements that are fundamentally inaccurate and distort our platform's architecture and our commitment to user safety".
The spokesperson insisted TikTok has "a strong record of protecting our community while giving users direct and intuitive control over their content experience".
Friday's demand comes after the EU in February told TikTok to change "addictive design" features, and after it accused the company in July of exposing children to risks like cyberbullying and predatory behaviour.
The EU has ramped up its efforts to protect children, launching investigations into the world's most popular platforms under the DSA law -- part of the bloc's legal armoury targeting digital giants.
Brussels plans to add to this by forcing social media companies to make platforms safe before children use them, as part of plans to ban children under 13 from social media, and permit only limited, supervised access until the age of 15.
 

'Repeated exposure risks'

 
A senior EU official said the February findings on TikTok focused on features like notifications, whereas Friday's view was about the algorithm itself.
TikTok is popular because of that algorithm, say observers, and the official admitted the EU was targeting its engagement-based business model.
TikTok can now access the EU's files and defend itself against the accusations.
But if the EU regulator's views are confirmed, the commission can impose a fine of up to six percent of the company's total worldwide annual turnover.
In its preliminary findings, the EU said TikTok did not "adequately assess" how its recommendation algorithm put users' wellbeing at risk.
The EU worried that repeated exposure to certain content could lead to "increased feelings of sadness or anxiety, body dissatisfaction, lower self-esteem, or worsening of existing eating disorders".
Brussels also told Meta in July to change the "addictive design" of Facebook and Instagram in an investigation started under the DSA.
TikTok and Meta are subject to similar scrutiny around the world, especially in the United States where there have been claims lodged with the courts over the dangers the platforms pose to young people.
TikTok in September avoided a trial after agreeing to pay Alabama at least $100 million and enact time limits and other restrictions for teenage users.
raz/sbk

markets

Beijing seeks to bolster AI shares

BY JULIEN GIRAULT, WITH LUNA LIN IN BEIJING

  • On Tuesday, net capital inflows into nine exchange-traded funds (ETFs), commonly used by Beijing to prop up the market, reached 120 billion yuan ($18 billion) -- a near-record, according to Bloomberg data, targeting tech stocks in particular.
  • After recent sharp drops for some Chinese tech stocks, Beijing is trying to prop them up -- seeing the sector, and in particular AI, as a central strategic priority.
  • On Tuesday, net capital inflows into nine exchange-traded funds (ETFs), commonly used by Beijing to prop up the market, reached 120 billion yuan ($18 billion) -- a near-record, according to Bloomberg data, targeting tech stocks in particular.
After recent sharp drops for some Chinese tech stocks, Beijing is trying to prop them up -- seeing the sector, and in particular AI, as a central strategic priority.
In August, buzzy Chinese startup Unitree, whose walking, dancing humanoids are often used by AI researchers trying to make robots act autonomously, went public.
Its shares skyrocketed on its Shanghai debut in August -- but almost three months later they are down 40 percent from their peak that day.
Chinese chipmakers Moore Threads and MetaX -- touted as challengers to US behemoth Nvidia -- have both plunged by around 50 percent over the past six months.
Similar steep falls can be seen across Shanghai's tech-focused Star50 index, China's answer to Wall Street's Nasdaq, which is down only around six percent over the same period.
Volatility on Wall Street hasn't helped -- the Nasdaq plunged 5.5 percent on Thursday -- with fears that the vast sums spent on AI infrastructure will not translate into big profits.
However, the Chinese government is taking steps.
Officials this week reactivated liquidity-injection tools aimed at encouraging share purchases while boosting the trading volume of state-backed investment funds, according to local media reports.
On Tuesday, net capital inflows into nine exchange-traded funds (ETFs), commonly used by Beijing to prop up the market, reached 120 billion yuan ($18 billion) -- a near-record, according to Bloomberg data, targeting tech stocks in particular.
The Financial Times also reported this week that Chinese regulators had told banks to halt company listings in an attempt to boost market liquidity.
"Chinese technology shares have been under significant pressure, with investors concerned about weak domestic demand, intense competition and disappointing earnings," said Charu Chanana, chief investment strategist at Saxo Markets.
However, now "fresh policy support, lower valuations and hopes of improving sentiment are encouraging investors to revisit a sector that has struggled for much of the year", she added.

Packed to the rafters

Overall, tech investors in China remain optimistic, said Ajin Jin, partner at Shanghai-based investment firm Kunquan.
"In both Beijing and Shanghai, weekly roadshows are still taking place, and every single one is packed to the rafters, with people even sitting in the aisles," he said.
Li Daxiao, one of China's most prominent investment gurus, also said "the burgeoning development of the industry is unstoppable" and should be clearly distinguished from "share price fluctuations".
"Technology shares have been affected by external risks and have fallen in tandem with their counterparts in the US, Japan and South Korea," he told AFP.
Stephen Inness of SPI Asset Management said that "a state-backed effort would add to steps taken by China last month to stabilise its equity market".
However, he warned the measures "have so far proved insufficient against the backdrop of a weak domestic economy and doubts over AI earnings".
ll-jug/kaf-stu/dan/fox

AI

Silicon Valley faces Democratic reckoning after betting on Trump

BY BENJAMIN LEGENDRE

  • "There's only one thing that sinks the US economy, it's the AI trade," and "the only way to sink the AI trade is to sink the data centers," Chamath Palihapitiya warned during an election-night livestream of the All-In podcast, which he co-hosts with other prominent pro-Trump Silicon Valley investors.
  • Silicon Valley bosses who spent two years cozying up to US President Donald Trump now face a Democratic Congress already threatening subpoenas, after midterm elections in which antipathy toward artificial intelligence swayed voters.
  • "There's only one thing that sinks the US economy, it's the AI trade," and "the only way to sink the AI trade is to sink the data centers," Chamath Palihapitiya warned during an election-night livestream of the All-In podcast, which he co-hosts with other prominent pro-Trump Silicon Valley investors.
Silicon Valley bosses who spent two years cozying up to US President Donald Trump now face a Democratic Congress already threatening subpoenas, after midterm elections in which antipathy toward artificial intelligence swayed voters.
"AI company CEOs should be called to testify under oath," Greg Casar, chair of the Congressional Progressive Caucus, said on Tuesday.
Senator Ed Markey, for his part, warned tech executives they would need to "plant new forests for all of the subpoenas we'll be printing."
It is a cold shower for tech bosses who rallied behind Trump after his 2024 win, with Elon Musk, Meta's Mark Zuckerberg, Amazon founder Jeff Bezos and others front and center at his January 2025 inauguration.
Musk went furthest, spending more than $250 million to help elect Trump and then taking charge of his DOGE cost-cutting drive. He poured nearly half a billion dollars more into backing Republicans in the midterms.
Once they take control of both chambers on January 3, Democrats will control the committees that can compel an executive to testify or hand over internal documents under threat of prosecution.
In 2020, the heads of Amazon, Apple, Facebook and Google were called to answer for their market dominance.
Musk, OpenAI's Sam Altman or Nvidia's Jensen Huang -- one of Trump's closest advisers on AI -- could now face the same ordeal.
Democrat Abdul El-Sayed, elected senator for Michigan despite more than $5 million spent against him by America PAC, the political committee funded by Musk, wants to go further.
He is calling for a "full-on pause" in the development of the most powerful AI systems and a US-China treaty to "pause and ban superintelligence," meaning AI that would surpass human intelligence.
To pass legislation, however, lawmakers will need to rally a party divided on the issue and overcome the veto of Trump, who remains president until January 2029.
No major tech boss had commented on the election as of Thursday except Musk, who joked: "Data centers in space looking pretty appealing right now."
Parts of the industry had begun distancing themselves from Republicans before the vote.
The main pro-AI political fund, financed by OpenAI president Greg Brockman and venture capital firm Andreessen Horowitz -- one of whose cofounders backed Trump in 2024 -- received no new donations from April through the end of September, as public distrust of AI grew.
Meta has hired former Democratic staffers in recent months to represent it in Congress, according to Axios.

'Shapeshifters'

Criticism is also coming from the Republican side.
"Voters are primed to reward whichever political party delivers on holding Big Tech accountable," Daniel Cochrane, a senior fellow at the conservative Institute for Family Studies, told AFP.
"In this election, concerns about AI's social and economic impacts shifted votes," he said, pointing to data centers as well as AI's effects on jobs and children.
According to an NBC News exit poll, AI and data centers were "a crucial or a major factor" for 56 percent of voters in their choice for Congress.
"Let's not confuse an anti-Trump wave with an anti-tech mandate," said Adam Kovacevich, CEO of Chamber of Progress, an influential centrist coalition funded by major tech companies.
"There's only one thing that sinks the US economy, it's the AI trade," and "the only way to sink the AI trade is to sink the data centers," Chamath Palihapitiya warned during an election-night livestream of the All-In podcast, which he co-hosts with other prominent pro-Trump Silicon Valley investors.
Even market analyst Dan Ives, known for his bullish outlook on tech growth, now estimates that 10 percent of the 800 to 1,000 new data centers planned in the United States "will not happen."
"The tech bros are shapeshifters," said Cochrane, who expects Zuckerberg "to have another mea culpa" before an "aggressive Dem congress." 
"Ultimately, the tech bros are for the tech bros," he said.
bl/arp/md

AI

Big tech, green groups in talks over data center guardrails

  • The "Responsible Data Center Collaborative" counts Google, Meta, Microsoft, Oracle, Amazon, Nvidia and Anthropic among its members, alongside environmental stalwarts the Natural Resources Defense Council (NRDC), World Wildlife Fund (WWF) and Sierra Club.
  • A coalition of tech giants and environmental groups is negotiating a national framework for data centers, according to a document reviewed by AFP on Thursday.
  • The "Responsible Data Center Collaborative" counts Google, Meta, Microsoft, Oracle, Amazon, Nvidia and Anthropic among its members, alongside environmental stalwarts the Natural Resources Defense Council (NRDC), World Wildlife Fund (WWF) and Sierra Club.
A coalition of tech giants and environmental groups is negotiating a national framework for data centers, according to a document reviewed by AFP on Thursday.
The "Responsible Data Center Collaborative" counts Google, Meta, Microsoft, Oracle, Amazon, Nvidia and Anthropic among its members, alongside environmental stalwarts the Natural Resources Defense Council (NRDC), World Wildlife Fund (WWF) and Sierra Club.
The members began holding meetings in September and will continue through February 2027, backed by the Rockefeller Foundation, a philanthropic organization, and Invest in Our Future, which is affiliated with it, the document said.
The matter was first reported by The Hill. 
Their purpose is to hash out guidelines for areas including energy and infrastructure, water and natural resources, community engagement, economic and community value, transparency and accountability, and implementation.
Given the mix of participants, the document is blunt about how they should behave: "Check your assumptions by asking clarifying questions" and "attack problems, not people."
It also says the discussions will be governed by "Chatham House rules" which means "conversations stay in the room and learnings leave (without attribution)."
Data centers are key to the artificial intelligence buildout but have angered communities across the United States. 
New power plants have driven up electric rates and the facilities create noise pollution. Cooling systems can draw large volumes of municipal water, while mists from cooling towers can carry harmful bacteria.
To avoid raising local power bills, newer data centers generate their own electricity, often with noisy, polluting gas-fired plants.
Voters in dozens of American cities approved restrictions or outright bans on data centers during Tuesday's midterm elections, in one of the clearest signals that a groundswell of anger over the AI boom has cut across political divides.
ia/acb

Zelda

Gamers get nostalgic as Zelda remake hits shelves

  • In a five-star review for the Guardian newspaper, Keza MacDonald wrote: "Nintendo's modern Zelda team have turned one of the best games of the 1990s into one of the best games of the 2020s, and made time-travellers out of us all."
  • Gaming fans around the world excitedly welcomed the release on Thursday of Nintendo's latest addition to its long-running Zelda series, "Ocarina of Time".
  • In a five-star review for the Guardian newspaper, Keza MacDonald wrote: "Nintendo's modern Zelda team have turned one of the best games of the 1990s into one of the best games of the 2020s, and made time-travellers out of us all."
Gaming fans around the world excitedly welcomed the release on Thursday of Nintendo's latest addition to its long-running Zelda series, "Ocarina of Time".
The game, a remake of a classic 1990s edition, is being released as part of the 40th anniversary of the action-adventure saga.
It has already won widespread praise, with a score of 96 out of 100 on aggregator site Metacritic by the eve of its release, putting it among the best-reviewed titles of the year so far.
"I used to play (Zelda) a lot on the Nintendo 64 when I was little, so I'm really excited about this remake, especially the improved graphics," 27-year-old fan Takumi Kawata told AFP in Tokyo, referring to the console first released in 1996.
His brother Kousuke Kawata, 31, added that Zelda visuals are "absolutely stunning, especially the beautiful forests".
The series centres on the adventures of green-clad warrior Link, who battles the evil demon Ganondorf across the ages trying to save Princess Zelda.
US tourist Marissa Olvera reminisced that she had spent "a whole summer" playing the original "Ocarina of Time" when she was a child.
"I just loved it. It was a bonding experience with my brothers," she told AFP, adding that the trailer for the remake had "brought back a lot of those feelings".

Rave reviews

Created in 1986 by Nintendo legend and Mario inventor Shigeru Miyamoto, the Zelda series has enjoyed growing success.
Its last two main installments, "Breath of the Wild" (2017) and "Tears of the Kingdom" (2023), sold more than 55 million copies between them.
Eiji Aonuma, who produced the new "Ocarina of Time", told AFP last month he was "all nervous" about the release.
"I don't know how it's going to turn out, I can't stop thinking about it," he said.
The reviews so far have gone some way to allaying his fear, with US outlet IGN calling it an "exceptional remake of an incredible game and a rich adventure that stands on its own".
In a five-star review for the Guardian newspaper, Keza MacDonald wrote: "Nintendo's modern Zelda team have turned one of the best games of the 1990s into one of the best games of the 2020s, and made time-travellers out of us all."
Market analysts will be watching to see how much the title can boost sales of the Switch 2, which became the world's fastest-selling games console on its 2025 release.
Nintendo reported a drop in quarterly net profit on Thursday as those sales, but maintained its annual forecasts.
kf/jxb/eab/tc/kaf-cg/mjw

tech

'We're not for sale': Americans want to pump brakes on data centers

BY THOMAS URBAIN

  • - Last straw - In Ottawa, a small town near Kansas City, residents also overwhelmingly approved a ban on large data centers. 
  • Voters in dozens of American cities approved restrictions or outright bans on data centers during Tuesday's midterm elections, in one of the clearest signals that a groundswell of anger over the AI boom has cut across political divides.
  • - Last straw - In Ottawa, a small town near Kansas City, residents also overwhelmingly approved a ban on large data centers. 
Voters in dozens of American cities approved restrictions or outright bans on data centers during Tuesday's midterm elections, in one of the clearest signals that a groundswell of anger over the AI boom has cut across political divides.
"Democrat, Republican, that didn't matter," said Alisha Lange, a farmer from Piqua, Ohio. "People want their town back."
She was one of the driving forces behind the grassroots group Save Piqua, which advocated for a ban on data centers with a capacity of more than 25 megawatts, effectively ruling out any large-scale project.
The ban passed in a ballot measure on Tuesday and now jeopardizes the construction of a data center worth more than $1 billion planned by Meta, even though the project had already been approved.
The 42-year-old farmer wept with joy when she learned that two-thirds of voters approved the measure.
"We're just a little farm town, and we would much rather just have a cornfield or have some sort of industry that provides an income to a larger amount of people," she said, referring to the relatively small number of jobs that data centers tend to create after they're built. 
While ballots are still being counted around the country, more than three dozen local measures have already passed that resulted in tighter regulations for data centers, according to consulting firm Redbud Consulting.
The outcomes add to restrictions already imposed by nearly 600 jurisdictions across the country, according to Electric Choice, including entire states in some cases, like New York and Texas.
Meta isn't the only big tech company spending billions on data center construction -- Google, Microsoft and Amazon are also investing heavily in an effort to meet growing demand for their AI products and services.  

Last straw

In Ottawa, a small town near Kansas City, residents also overwhelmingly approved a ban on large data centers. 
"We're not for sale," Michael Lewis said. 
Lewis is a member of a grassroots organization called Ottawa United, and ran for Congress this cycle as a Democrat, though he lost on Tuesday to the incumbent Republican.
"You can dangle some money in our face, it doesn't mean we're going to give away our resources and deal with the environmental issues that stem from" data centers, he said.
In Janesville, Wisconsin, a majority of city council members were initially open to a proposal from developer Viridian, which was ready to put $8 billion on the table to build a data center on a former General Motors site that closed in 2008.
But communities are starting to ask for more than job creation potential and total investment when scrutinizing new construction projects. 
"What we're seeing now is that there is a much broader range of indicators that indicates a good project or a bad one," Cathy Erdman, a member of Janesville for Responsible Development, said.
A project's impacts on local tax streams, broader infrastructure, quality of life and other land uses "can be equal to or greater in importance than the number of jobs or just the dollar amount in capital investment," Erdman said.
Under pressure, the city council ultimately rejected the Viridian project in June, then adopted a one-year moratorium, before a majority of voters on Tuesday backed new restrictions.
Any large-scale project on the former General Motors site must now be approved by referendum.
While some in Janesville went to the polls to express their outright opposition to data centers, others simply wanted "more transparency," Erdman said.
"I believe that is the general consensus community-wide: that that project, in particular, was very poorly handled by the powers that be," the 45-year-old activist said.
In Piqua, the Meta project was the last straw, Lange said, leading to the removal of two city council members and the mayor.
In Ottawa, Lewis will remember the data center protest joining forces across the political aisle with a Republican official to win a court order for a referendum.
"It was wonderful," he said. "That was the true meaning of community right there."
tu/bl/smb/jgc/mlm/lga

schools

Google turns off AI for schoolchildren by default

  • The feature was switched on by default for students at schools that had already enabled Gemini.
  • Google has said it will switch off its AI tools for schoolchildren by default, after drawing criticism for pushing artificial intelligence features to students while schools are still debating how the technology should be used in classrooms.
  • The feature was switched on by default for students at schools that had already enabled Gemini.
Google has said it will switch off its AI tools for schoolchildren by default, after drawing criticism for pushing artificial intelligence features to students while schools are still debating how the technology should be used in classrooms.
The search engine juggernaut dominates technology in American schools. Its education apps, including Gmail, Docs and the Classroom platform, are used in classrooms across the country, and its low-cost Chromebook laptops are the devices US schools most often issue to students.
Under the changes Google announced in a blog post on Tuesday, Gemini, Gemini Notebook, Gemini in Classroom and Search will be turned off for K-12 student accounts until school administrators switch them on.
Google told AFP the new controls would be rolled out worldwide, beginning in the United States, but that it had not yet set a date.
The changes also mean schools will not be able to give students Google Search without its AI features.
Chris Phillips, Google's vice president for education, told The New York Times that "Search and A.I. are deeply integrated," as AI features are now built into the company's search engine.
If a school bans AI, he said, students "will not have access to search, and they'll see a message from their school that indicates it's not available."
New York Mayor Zohran Mamdani accused Google on X of "trying to use its monopoly power to force AI into our public schools."
"No corporation should have the power to dictate how we teach our children," he wrote.
More than 150 million students worldwide use Google Classroom, according to a letter Senators Richard Blumenthal and Josh Hawley sent to the company last month asking why AI features had been made the default on school devices.
In August, Google opened its Gemini chatbot in Classroom to students of all ages.
Students were shown starter prompts tied to their specific class and assignment, offering guided help or study guides, practice quizzes and flashcards built from their course materials.
The feature was switched on by default for students at schools that had already enabled Gemini.
New York and Los Angeles both announced plans in September to restrict access to AI tools on school devices.
"We are reviewing Google's announcement and its potential implications for schools, educators and students," a spokesperson for the Los Angeles Unified School District told AFP.
The spokesperson said the district believes AI can improve teaching and learning if it is introduced carefully and with clear safeguards, and pointed to district principles requiring that technology support rather than replace teachers and protect student privacy.
New York City Public Schools (NYCPS) are seeking more details from Google about the policy change, but the Gemini chatbot and AI mode are already disabled on Chromebooks used by students in NYC, a spokesperson for the district told AFP.
"NYCPS has also made requests to Google to disable the AI Overview displayed in the Google search result for students," the spokesperson said.
In a statement on X, Google said Tuesday's changes give "schools more choice, not less."
"We know there's concern about students using AI to shortcut schoolwork, and schools have told us they want control," Google said.
arp/smb/sla

internet

Harry and Meghan to discuss online harms at EU parliament event

  • The event will be Meghan's first public appearance since revelations that police in the Philippines have found child sexual abuse material on devices at the former home of her estranged father, Thomas Markle.
  • Prince Harry and his wife Meghan will speak in Brussels about the dangers children face on social media, in their first public appearance since allegations emerged that her father had possessed child abuse images.
  • The event will be Meghan's first public appearance since revelations that police in the Philippines have found child sexual abuse material on devices at the former home of her estranged father, Thomas Markle.
Prince Harry and his wife Meghan will speak in Brussels about the dangers children face on social media, in their first public appearance since allegations emerged that her father had possessed child abuse images.
The couple, also known as the Duke and Duchess of Sussex, will attend the Tuesday opening of a video installation commemorating children who have lost their lives due to harm experienced via social media, a spokesperson told AFP.
The royals will open The Lost Screen Memorial at the European Parliament displaying images of children affected by online abuse and exploitation and give a joint speech before taking part in a panel event with international activists.
The event will be Meghan's first public appearance since revelations that police in the Philippines have found child sexual abuse material on devices at the former home of her estranged father, Thomas Markle.
It will also be the first time that the couple appear together at a public event abroad since returning to live in Britain this summer.
The event, hosted by Spanish MEP Antonio Lopez-Isturiz, highlights the "urgent need to make the digital world safer for every child", a statement said, calling it a "call to action" for policymakers, technology companies and society to do more to protect children.
The couple, who have two young children, Archie and Lilibet, together unveiled the installation in New York in April 2025 and it has also been shown in Los Angeles.
Meghan gave a speech at its opening in Geneva in May.
The installation shows 50 enlarged images of children who died as a result of harm they suffered on social media, taken from their parents' lock screens.
At the event in Brussels, the European Parliament chief Roberta Metsola will give an address, as will Sandra Spitaels, whose 15-year-old son Glenn committed suicide in 2017 after a nude image of him was posted online.
am/jkb/phz

technology

Google unveils $15 bn data centre investment in Malaysia

  • It said the $15 billion investment will be made "over the next two years (2027 to 2028)", expanding on a $2 billion investment announced in 2024.
  • Google announced plans on Wednesday to invest $15 billion in Malaysia over the next two years to build data centres and other infrastructure as demand booms for artificial intelligence services.
  • It said the $15 billion investment will be made "over the next two years (2027 to 2028)", expanding on a $2 billion investment announced in 2024.
Google announced plans on Wednesday to invest $15 billion in Malaysia over the next two years to build data centres and other infrastructure as demand booms for artificial intelligence services.
The country is Southeast Asia's fastest growing hub for new data centres -- warehouse-like buildings crammed with servers that can train and run AI tools such as chatbots and image generators.
"Malaysia is one of Southeast Asia's most dynamic digital economies, driven by 99 percent internet penetration and rapid AI adoption across government agencies, businesses, and consumers," the US tech giant said in a statement.
It said the $15 billion investment will be made "over the next two years (2027 to 2028)", expanding on a $2 billion investment announced in 2024.
AI data centres require large amounts of electricity -- placing increased pressure on grids often still reliant on fossil fuels -- and, in many cases, water to cool down hot computing equipment.
A backlash is growing worldwide against their environmental impact, and there are also concerns over noise, pollution, their appearance and how many jobs the facilities actually create.
Google pledged Wednesday that its data centre campuses in Malaysian industrial zones would be "energy-efficient".
The company said it would work to "bolster local water security, and add more than one gigawatt of clean energy capacity" to the national grid.
"Once operational, the facilities are projected to sustain more than 9,000 local jobs per year, with wages averaging 45 percent above the national median," it added.
Malaysia accounts for more than half of under-construction data centre capacity among Southeast Asian countries, including Indonesia, Thailand, the Philippines and Vietnam, according to market intelligence firm DC Byte.
The International Monetary Fund forecasts that Malaysia's economy will grow by 5.2 percent in 2026, partly encouraged by data centre activity.
But by 2035, data centres are forecast to swallow up one-third of Malaysia's total energy consumption, the deputy prime minister told parliament in July.
The government said at that time it would only approve data centre projects after ensuring sufficient energy and water supply capacity.
In October, a Finnish authority ordered work on two of Google's data centre sites in the country halted until mandatory environmental impact assessments have been completed.
bur-kaf/mjw

smartphones

Apple AI tools come to China after two-year delay

BY LUNA LIN

  • The long-anticipated Chinese rollout of Apple Intelligence "fills a long-standing gap in Apple's AI offering", said Ivan Lam, a senior analyst at Counterpoint Research.
  • Apple began on Wednesday a long-delayed trial rollout of in-house artificial intelligence tools for Chinese users, playing catch-up to homegrown smartphone makers in a major market for the US giant.
  • The long-anticipated Chinese rollout of Apple Intelligence "fills a long-standing gap in Apple's AI offering", said Ivan Lam, a senior analyst at Counterpoint Research.
Apple began on Wednesday a long-delayed trial rollout of in-house artificial intelligence tools for Chinese users, playing catch-up to homegrown smartphone makers in a major market for the US giant.
The on-device features dubbed "Apple Intelligence" were first released in the United States and other countries in autumn 2024, marking the firm's first major push into generative AI.
But strict Chinese regulatory requirements for the fast-evolving technology have held back its launch there -- as well as the fact that models from Google and OpenAI, which Apple has used elsewhere to boost its own AI capabilities, are not available in the country.
Apple, which has largely stayed out of the global AI investment frenzy, announced it will work with a Chinese partner -- Alibaba's Qwen -- to integrate more powerful AI features into devices alongside its own basic AI offerings.
Wednesday's trial rollout is restricted to developers, but public access to Apple Intelligence services in China will come in December, the company said in a statement at the weekend.
The long-anticipated Chinese rollout of Apple Intelligence "fills a long-standing gap in Apple's AI offering", said Ivan Lam, a senior analyst at Counterpoint Research.
However, it does not necessarily give the Californian company "a significant advantage" in a crowded market, Lam said.
The initial release includes 18 AI features, from live translation of phone calls to tools for summarising and proofreading messages, but not the latest upgrade of iPhone assistant Siri.

'Concession'

Chinese regulations require AI service providers to register with the country's internet watchdog, and there are requirements to use local data centres and process data domestically.
Apple only secured regulatory approval in July, after reportedly holding talks with Chinese tech firms including Alibaba, Baidu, ByteDance and Tencent about partnerships on Apple Intelligence.
Major US AI firms, including Google, OpenAI and Anthropic, have either banned Chinese users or restricted access to their services, and are unlikely to secure regulatory clearance.
Working with local partners on Apple Intelligence "signifies a structural concession by Apple regarding its pricing and defining power over the AI experience in the Chinese market", Zhang Yi, founder of research firm iiMedia, told AFP.
Apple's delay in introducing AI tools in China was not due to lags in research and development, but rather "compliance costs", he said.
Despite reporting a 21 per cent increase in revenue in the Greater China region, which includes Hong Kong and Taiwan, for the past quarter, Apple's new CEO John Ternus still faces significant challenges in an evolving market.
According to research firm Omdia, more than 110 million iPhones in China can support Apple Intelligence, providing a foundation for quick adoption -- but competition is fierce.
"Chinese mobile phone manufacturers and internet companies are actively developing AI features and agent ecosystems," Zhong Xiaolei, a senior analyst at Omdia, told AFP.
ll/kaf/abs

crypto

Binance reveals hackers used fake job interviews, malicious code in $6-bn heist

  • In September, the attackers approached another Binance engineer through a legitimate job-search platform, offering him an interview for a position in a fictitious company.
  • Binance released on Tuesday a detailed investigation into the $6 billion hack of its cryptocurrency platform, showing how attackers used fake job interviews and malicious code to steal the funds.
  • In September, the attackers approached another Binance engineer through a legitimate job-search platform, offering him an interview for a position in a fictitious company.
Binance released on Tuesday a detailed investigation into the $6 billion hack of its cryptocurrency platform, showing how attackers used fake job interviews and malicious code to steal the funds.
The crypto exchange said it had replenished 98 percent of the affected assets by drawing on its own funds, with the remaining amount involving little-traded currencies that are harder to acquire. 
The attack, carried out over about four hours on October 29, was attributed by the US Federal Bureau of Investigation (FBI) to North Korea-affiliated hackers.
Pyongyang has denied the allegations, with the powerful sister of leader Kim Jong Un calling the claim "absurd rhetoric."
Around $2.85 billion of the stolen $6.34 billion has been frozen, Binance said, meaning that the attackers could not use it or convert it to another currency.
Binance is the world's largest crypto exchange by daily trading volumes. News of the hacking attack, which is thought to be the biggest ever crypto heist, caused the prices of most major cryptocurrencies to fall sharply.
According to Binance, attackers took their first steps towards the breach as early as June.
First, they targeted a platform engineer by getting them to download a virus-laden version of the widely-known programming software Homebrew. 
Binance's cybersecurity teams identified the malware and removed it, but the hackers had already had time to harvest information about the engineer and his colleagues. 
In September, the attackers approached another Binance engineer through a legitimate job-search platform, offering him an interview for a position in a fictitious company.
After building trust with the subject through the sharing of social media photos and videos, they asked the engineer to download a test exercise as part of the interview process. 
The exercise contained malware that deployed "sophisticated obfuscation and evasion techniques," Binance said, including replacing a legitimate development tool component with a malicious version.
That malware was critical to the success of the attack, as it allowed the hackers to establish a connection with Binance's internal systems.
On October 20, the attackers used the access they had gained and an internal AI tool to gain control over software used by two other engineers.  
Nine days later, they triggered malicious code that initiated the attack.
tu/elm/aha/bs

food

At least 11 European countries snared in freezer cyberattacks

  • Four supermarket outlets were affected in Germany and six in Romania, officials and company representatives said on Tuesday.
  • Germany, Romania and the UK on Tuesday joined the growing list of European nations hit by cyberattacks on supermarket refrigeration systems in recent days.
  • Four supermarket outlets were affected in Germany and six in Romania, officials and company representatives said on Tuesday.
Germany, Romania and the UK on Tuesday joined the growing list of European nations hit by cyberattacks on supermarket refrigeration systems in recent days.
Italy's government held emergency talks with intelligence chiefs on Tuesday in the wake of the issue, which prompted the EU to warn of a deliberate attempt to disrupt the bloc's food supply chain.
The motive for the cyberattacks is not known and nobody has claimed responsibility yet.
EU spokesman Thomas Regnier said EU agencies were also talking with national authorities on Tuesday to come up with a collective response.
"These are not isolated incidents, and of course, we have to be collectively resilient and prepared for such cyber incidents," he said. 
But he refused to point the finger at any potential culprits. 
Italy and the Netherlands were among the 10 or so EU countries hit, just weeks after hundreds of French supermarkets were forced to throw away spoiled stock in a similar incident.
Britain, which is not in the bloc, was also affected.
National cybersecurity chief Richard Horne told the BBC the issue affected "the whole food supply chain".
He said staff had "worked through the night" to alert those at risk, adding that there had been a couple of instances where "vulnerable technology had been exploited".
The British Association of Independent Meat Suppliers (AIMS) also told the BBC two of its members saw their fridges begin to defrost, while another saw the temperature of their cold units shoot up.
EU food and agriculture commissioner Christophe Hansen said on Monday the attacks were a "deliberate attempt to disrupt our food supply chain".
Italy's government said as yet the impact was "highly limited" and affected outlets with outdated security systems.
"Only a small number of commercial units have been affected," the statement said.
Ministers nevertheless met intelligence chiefs and the National Cyber Security Agency, which also issued technical advice for retailers to protect their systems.

'Semi-amateurish'

There were reports of around 20 outlets affected nationwide in Italy.
Giuseppe Spadaro, director of a group that manages the Coop supermarkets in Sicily, said four of the company's 300 stores were hit.
"We estimate the damage at several thousands of euros. We had to throw away everything. The shelves are empty," he was quoted as saying by the Corriere della Sera daily.
Four supermarket outlets were affected in Germany and six in Romania, officials and company representatives said on Tuesday.
"No other infrastructure was compromised," Dan Cimpean, director of Romania's National Cyber Security Directorate (DNSC), told AFP.
"It seems semi-amateurish. We do not suspect, under any circumstances, that it is a state actor."
Carlos Cancinos, spokesperson for the Coop chain in Sweden, said the group's seven affected shops had resumed "normal functioning".
Refrigeration equipment and freezers affected overnight from Monday to Tuesday were connected to "old 4G networks", he said. 
"We have launched a major programme to modernise a large part of our store network, equipping connected refrigerators and ensuring that they operate via 5G networks. This helps to solve a large share of problems of this kind," he said.
Belgium was among the hardest-hit countries, with dozens of stores across several grocery chains reporting refrigeration failures. Federal prosecutors there have opened an investigation.
burs/sbk/jxb

semiconductors

Nvidia orders data center clients to reduce China exposure: FT

  • Nvidia is the world's most valuable company thanks to the global drive to build data centres packed with specialist computer chips and other hardware that runs AI systems. bur-lkd/lga
  • US chipmaking giant Nvidia is ordering its data center partners to scale back work with Chinese customers, the Financial Times reported on Tuesday. 
  • Nvidia is the world's most valuable company thanks to the global drive to build data centres packed with specialist computer chips and other hardware that runs AI systems. bur-lkd/lga
US chipmaking giant Nvidia is ordering its data center partners to scale back work with Chinese customers, the Financial Times reported on Tuesday. 
Nvidia's cloud computing clients must now sell at least 51 percent of their services to non-Chinese companies, the FT reported, citing five people with knowledge of the matter. 
The new rules only affect future contracts and not existing agreements, the report said.
Nvidia neither confirmed nor denied the report to AFP, but said in a statement that "datacenters and clouds built on American chips are the foundation that American model and application makers need to compete worldwide." 
"Longstanding U.S. policy encourages every cloud to build on American infrastructure and to win business everywhere they can," the statement said.
"America cannot win the AI race unless America's AI infrastructure wins worldwide -- if the world's clouds rely on foreign-made chips, the model and application layers will be foreign-built as well."
Asked about the report, Chinese foreign ministry spokeswoman Mao Ning said she was not aware of it.
"As a matter of principle, we consistently hope that all parties can provide a fair, transparent and non-discriminatory business environment for companies from all countries, including Chinese companies," she said. 
China and the United States are locked in a race for dominance in the fast-moving artificial intelligence sector.
In recent months, the two countries have traded accusations of stealing AI technologies and jeopardising security.
Nvidia is the world's most valuable company thanks to the global drive to build data centres packed with specialist computer chips and other hardware that runs AI systems.
bur-lkd/lga

cybercrime

South Korea investigates online leak of president's passport details

  • Local media reported on Monday that a computer file containing personal details, including the passport numbers and dates of birth of the first couple and senior officials accompanying Lee on an overseas trip, had been uploaded to a website in June.
  • South Korea is investigating an online leak of passport details belonging to President Lee Jae Myung and more than 160 other officials, a government minister said on Tuesday.
  • Local media reported on Monday that a computer file containing personal details, including the passport numbers and dates of birth of the first couple and senior officials accompanying Lee on an overseas trip, had been uploaded to a website in June.
South Korea is investigating an online leak of passport details belonging to President Lee Jae Myung and more than 160 other officials, a government minister said on Tuesday.
Local media reported on Monday that a computer file containing personal details, including the passport numbers and dates of birth of the first couple and senior officials accompanying Lee on an overseas trip, had been uploaded to a website in June.
The file was reportedly uploaded to VirusTotal, a malware-scanning platform, ahead of Lee's visit to Europe, and remained accessible for nearly five months.
It remains unclear who uploaded it and why.
The foreign ministry has since renewed the passports of all those affected, including Lee, and had the exposed information removed from the website, ministry spokesman Park Doo-soon said at a news briefing on Tuesday.
"The foreign ministry is now working with the National Police Agency and other agencies to determine how the information was leaked," he said.
Park said the ministry was treating the incident as a data leak rather than a hacking attack, and was examining computers and email accounts used to prepare the presidential trip.
The leak was a "serious matter that must not be taken lightly", Presidential Chief of Staff Lim Kwang-hyun told an emergency meeting on the breach, according to Lee's office.
The incident has raised "need for a thorough investigation into how the leak occurred, the extent of the information exposed and whether any actual harm had resulted", he said.  
kjk/mjw

food

Fridge cyberattack hits supermarkets across Europe

BY OLIVIER BAUBE WITH PIA OHLIN IN STOCKHOLM

  • "They are a deliberate attempt to disrupt our food supply chain and undermine consumers' confidence in it -- a warning that we must take seriously," he said.
  • A suspected cyberattack hit grocery stores across at least nine European countries on Monday, prompting the EU to warn of a deliberate attempt to disrupt the bloc's food supply chain.
  • "They are a deliberate attempt to disrupt our food supply chain and undermine consumers' confidence in it -- a warning that we must take seriously," he said.
A suspected cyberattack hit grocery stores across at least nine European countries on Monday, prompting the EU to warn of a deliberate attempt to disrupt the bloc's food supply chain.
Supermarkets in Belgium, Estonia, Finland, Ireland, Italy, Luxembourg, the Netherlands, Portugal and Sweden reported refrigeration problems that forced some retailers to discard food after temperatures rose overnight.
The incident came three weeks after a cyberattack disrupted French grocery stores and forced hundreds of outlets to throw away spoiled stock.
No one has claimed responsibility for either incident.
EU food and agriculture commissioner Christophe Hansen said the bloc was taking the incidents "very seriously".
"They are a deliberate attempt to disrupt our food supply chain and undermine consumers' confidence in it -- a warning that we must take seriously," he said.
In Italy, at least four Coop supermarkets in Sicily and one business in Tuscany were affected, according to media reports and the national cybersecurity agency.
The Italian government was to hold a cybersecurity meeting with intelligence services on Tuesday.
"The attack happened at 4:00 am (0300 GMT)," Giuseppe Spadaro, head of the Radenza Group which manages Coop-branded supermarkets, told Italian news agency ANSA.
The attack targeted "our remote-control system for frozen food sections" and had since been resolved, he added.
Belgium was among the hardest-hit countries, with dozens of stores across several grocery chains reporting refrigeration failures. Federal prosecutors opened an investigation.
"It's a cyberattack, there is no other reason," said Mich Vergauwen, a spokesman for Carrefour Belgium which reported disruption in about 15 of its 700 stores.
Ikea said its Mons outlet also experienced a cyberattack that caused freezer temperatures to rise.
Belgian refrigeration company Sabcobel said it had begun securing its installations after the October attack in France, but some systems had yet to be upgraded and were among those affected on Monday.
Delhaize, Colruyt and Intermarche also reported refrigeration problems at stores in Belgium, while four Delhaize outlets in Luxembourg were affected.
In Ireland, retailer Musgrave said 13 Supervalu stores were impacted by what it described as a cyber incident.
A small number of corner shops in the Netherlands also suffered disruptions.
In Portugal, the Mosqueteiros group confirmed to AFP incidents at seven Intermarche stores linked to a cyberattack, which "appears to have targeted an external service provider responsible for managing the refrigeration systems".
In Sweden, Coop confirmed six of its around 800 stores had been hit.
"We can confirm that it is a cyber incident. The problem has now been resolved and the fridges and freezers are working as normal," spokesman Carlos Cancino told AFP.
He said Coop did not know who was behind the attack, but noted that the affected stores had been connected to a 4G mobile network, not the updated 5G system, which made them "more exposed" to hackers.
 

'Vulnerabilities'

 
Swedish chain Ica also said "a limited number" of its stores experienced refrigeration problems, without confirming a cyberattack.
According to Ica spokesperson Helena Johansson, the affected stores were all connected to an external refrigeration control system provided by Danish manufacturer Danfoss -- the same supplier targeted in the October 12 cyberattack on French retailers.
She referred all queries to Danfoss, which had yet to respond to an AFP request for a comment.
In Finland, irregular temperatures were also detected in a few smaller grocery stores belonging to the Kesko group.
"We are currently investigating the cause, and interference via the network has not yet been ruled out," Petri Kayhko, head of Kesko's corporate security, told AFP.
Roni Kokkola, an information security specialist at the National Cyber Security Centre at Finland's Transport and Communications Agency Traficom, would not rule out cyber interference.
"We've seen this in Europe in the past few months, so it spreading beyond France to other countries has been expected," he told AFP.
"Nowadays, it's very typical that all kinds of vulnerabilities are being exploited ever more actively and rapidly." 
In Estonia, at least one supermarket was attacked by hackers who raised the temperature in the refrigeration units to 50C, public broadcaster ERR reported.
The units were older models delivered by Danfoss, similar to the ones targeted in the France attack, Jann Mutso, director of Kitman Coldtech, a manufacturer and distributor of advanced refrigeration systems, told ERR.
po-bur-ec/sbk/rmb/cms/abs