United News of Bangladesh

Protesters broke into a construction site for a new Microsoft data centre in Amsterdam on Friday, with some climbing a crane and setting off flares, while others clashed with police amid growing opposition to large-scale data centre projects across Europe. Nathan Timmer of the protest group Geef Tegengas, meaning Push Back, told local broadcaster AT5 that demonstrators had occupied the site to prevent construction from continuing. Activists across Europe planned demonstrations and other actions on Friday and Saturday, demanding a halt to large-scale data centre developments over concerns about their environmental and social impacts, particularly their heavy consumption of electricity and water. "We don't want it to be built. Nobody needs it," Timmer said, arguing that the facilities would support artificial intelligence, surveillance and warfare. Pro-Palestinian activists also joined the demonstration in Amsterdam. Police began removing protesters from the site by late afternoon, although some remained. Dutch police said they detained between 150 and 200 people on suspicion of trespassing. Microsoft acknowledged the demonstration and said it respected people's right to peaceful protest but opposed unlawful activities that endangered public safety. The Amsterdam protest was part of the "Stop Dirty Data Centres" campaign, involving dozens of grassroots and civil society organisations in 15 European countries, including Ireland, Poland, Denmark, Finland, Spain, Italy, the Netherlands, France and the United Kingdom. Owen Espley, coordinator of StopDirtyDataCentres.eu, said there was a growing disconnect between political leaders promoting artificial intelligence and public concerns about its expansion. In Germany, around a dozen protesters gathered outside a data centre near Berlin, calling for an end to construction without proper consideration of environmental consequences. In Poland, approximately 100 people demonstrated outside parliament in Warsaw, representing communities where new data centres are planned. Italy has also witnessed growing opposition, particularly in the northern Lombardy region, a major hub for digital infrastructure. Activists say the rapid expansion of facilities supporting cloud computing and artificial intelligence is outpacing public consultation and environmental oversight. On October 10, more than 1,000 people and representatives of around 40 local committees gathered at Milan's Piazza della Scala for what organisers described as Italy's first national demonstration against data centres. They demanded a moratorium on new projects and a national strategy to regulate the industry's expansion, following a series of increasingly large protests in towns surrounding Milan. According to PwC's Global Data Centre Outlook 2026–50, Europe is projected to attract $5.6 trillion in cumulative data centre investment between 2026 and 2050, representing about 18% of global spending. The United States is expected to attract $15.1 trillion during the same period.

Three people have been indicted on charges of wire fraud and aggravated identity theft for allegedly using artificial intelligence to impersonate corporate executives and lawyers in a scheme that defrauded companies of approximately $129 million, prosecutors said Thursday. According to an indictment filed in a US District Court in New York, Jacob Soall, 41, Sharon Hazan, 54, and Guy Sheetrit, 42, targeted thousands of companies worldwide over a period of more than 18 months using AI-generated impersonations. Their alleged victims included hospital networks in Missouri and California, a Rhode Island-based bank, a financial technology company in the United Kingdom, a Massachusetts mutual fund company and a Florida-based pet food business. The three defendants, who have links to Spain, the United Kingdom, Israel and the United Arab Emirates, also face money laundering charges, according to court documents. Soall and Sheetrit's lawyers did not immediately respond to requests for comment. It was not immediately clear whether Hazan had legal representation. The US Attorney's Office for the Southern District of New York also did not immediately respond to a request for comment. The suspects were arrested between June and September. According to prosecutors, Hazan was convicted in a New York court in 2008 of conspiracy to commit bank fraud. The defendants, who reportedly used nicknames including "Rocky Balboa", "Superman Cash" and "Fat Boy", primarily targeted corporate accounting departments. They allegedly sent emails pretending to come from company chief executives and requested assistance with confidential financial transactions. They then arranged telephone or video calls during which they used AI technology to imitate the executives' voices and, in some cases, their appearances. Employees were subsequently instructed to contact lawyers at major law firms, whom the suspects were also prepared to impersonate, according to the indictment. Soall primarily posed as British lawyers and reportedly used a handwritten script titled "Lawyer Call" to make himself sound like a genuine legal professional while communicating with victims. The impersonated lawyers then instructed employees to transfer large sums of money into bank accounts in different countries. The money was subsequently converted into cryptocurrency to conceal its origins before being distributed among the defendants and other alleged participants, court documents said. FBI Assistant Director in Charge James C. Barnacle Jr said the case demonstrated how advanced technology could be exploited to carry out large-scale financial fraud. He said the FBI remained committed to adapting its investigative capabilities to identify and disrupt criminals using emerging technologies to target individuals and businesses. If convicted, the defendants could face maximum sentences of 20 years for each of the charges, except aggravated identity theft, which carries a mandatory two-year prison term, according to the US Attorney's Office.

Google, Meta and TikTok have filed separate lawsuits against California, seeking to block a new law aimed at protecting children from harmful social media features, arguing that it violates their constitutional right to freedom of expression. The legislation, signed by Governor Gavin Newsom in September, is scheduled to take effect on Jan 1, 2027. It prohibits technology companies from providing users under 16 with potentially addictive features, including personalised feeds, endless scrolling and automatically playing videos. The law also requires companies to take "reasonable measures" to verify users' ages and prevent children under 16 from accessing these features. The three companies filed separate lawsuits in federal courts in San Francisco, Los Angeles and San Jose. Besides challenging the law under the First Amendment, they argue that it violates Section 230 of the 1996 Communications Decency Act, which generally protects technology platforms from legal liability for content posted by third parties. The legal challenges come amid growing scrutiny of social media companies over the impact of their platforms on children's well-being. Schools, government authorities and individual plaintiffs have filed thousands of lawsuits alleging harm to young users. In August, Meta agreed to pay up to $18 billion and introduce stronger child-safety measures on Facebook and Instagram under a landmark legal settlement. The agreement ended a trial concerning teenage social media addiction and resolved claims brought by nearly every US state over the platforms' alleged role in harming children's mental health. The legal action focused on platform features designed to keep young users engaged for extended periods. In September, TikTok also reached a settlement with Alabama over allegations involving addictive features. California Attorney General Rob Bonta's office, which is representing the state in all three cases, said Thursday that it was reviewing the lawsuits and remained committed to defending legislation designed to protect children online. TikTok described the California law as an unconstitutional restriction on free expression, arguing that it undermines the established practice of recommending and organising content according to users' preferences, a system also used by bookstores, television networks and streaming services. The company said it had introduced several safety measures for younger users and argued that the legislation disregards progress in online child protection, restricts free speech and undermines parents' authority to decide what content is appropriate for their children. Meta argued that the legislation violates the First Amendment by allowing the government to determine how information and ideas are organised, displayed and shared with California residents. The company said it would be forced either to remove features that help maintain age-appropriate experiences for teenagers or prevent thousands of young users in California from connecting with family members, friends and their communities. Google said parents were best placed to make decisions for their children and highlighted the parental controls and safety protections already available through its services. "Overly broad laws like this actually break the systems we've built to help kids find high-quality, age-appropriate videos," Google spokesman José Castañeda said.

A cyberattack targeting refrigeration systems at three French blood bank facilities attempted to raise storage temperatures to extremely high levels, but security systems detected the incident before any blood products or samples were damaged, French authorities said. The incident follows reports this week that supermarkets in several European countries experienced disruptions after hackers remotely accessed their refrigeration systems, causing refrigerators and freezers to warm up and resulting in millions of euros in spoiled food. The French Blood Establishment (EFS) said in a statement released late Tuesday that temperature and defrosting settings at three plasma storage facilities were raised to extremely high levels on Monday. Monitoring systems quickly detected the abnormal activity, allowing staff to intervene before any damage occurred, the organisation said. French weekly newspaper Le Canard Enchaîné, which first reported the attack, said hackers had attempted to change storage temperatures from the normal minus 30 degrees Celsius (minus 22 degrees Fahrenheit) to as high as 110 degrees Celsius (230 degrees Fahrenheit). The attack targeted a supplier of refrigeration equipment used at three EFS facilities. The affected storage units contained plasma bags intended for transfusions and plasma used in medicine production. The products were either transferred to other facilities or secured, according to the EFS. "No health consequences have been identified, and no blood products were lost," the organisation said. The EFS added that blood collection and distribution activities continued without interruption, ensuring uninterrupted supplies for patients at hospitals and other healthcare facilities. Coolworld, the company supplying the refrigeration equipment, said it was aware of the cybersecurity incident. In a statement, the company said the vulnerability reported in the media involved a component supplied by a third party and was beyond its direct control. It added that only a limited number of units were affected and that appropriate measures were being taken. France's national cybersecurity agency and the Health Ministry are investigating the incident. The latest attacks on refrigeration systems resemble a cyberattack in mid-October that affected refrigerators and freezers at hundreds of supermarkets and independent stores across France, Belgium and the Netherlands. Dutch supermarket chain Ahold Delhaize, one of the companies affected by the earlier incident, said Wednesday that the cyberattack could reduce its fourth-quarter sales by up to 250 million euros ($281 million) and lower its operating income by as much as 100 million euros. "While most of the disruption was contained within two weeks, work continues to fully restore our normal operations," CEO Franz Muller said in a statement accompanying the company's quarterly earnings report. He also warned that cyber threats had become an increasingly serious reality for businesses. The company had earlier attributed the disruption to "malicious actions by others" and proactively shut down some of its systems, including websites and applications for several brands operating in Europe and the United States.

Researchers have shown that an artificial intelligence (AI) model can learn and improve its predictive ability using only data it generates itself, without exposure to real-world information, Israel's Tel Aviv University (TAU) said in a statement Wednesday. The study, published on the scientific website arXiv, presented a new approach to AI training that could help machines develop learning skills without relying on large collections of human-created material, according to the statement. Researchers from TAU and Stanford University trained two computer models together, with one creating symbol sequences and the other learning to predict the next symbol. The first model gradually improved at creating learning exercises that helped the second model become better at making predictions. The team tested the trained model on several types of information, including language, images, sounds, music, computer code, and DNA sequences. The results showed that its ability to make predictions improved as more computing power was used during training. The researchers stressed that the method can't teach AI facts about the real world without exposure to real-world information. However, it could help future AI systems learn more efficiently when they encounter real-world data.  

Google will restrict students' access to its search engine in schools that choose not to adopt the company's artificial intelligence (AI) tools, the technology giant announced Tuesday. The decision means schools using Google's classroom software will have to enable AI features to retain access to the company's search engine, which is increasingly integrated with AI technology. The Mountain View, California-based company has faced growing criticism from educators, child development specialists and parents over the use of AI in schools. Concerns intensified in August when Google activated its Gemini chatbot for many K-12 students using Google Classroom. In its latest announcement, Google said enabling its AI services would no longer be the default setting for school administrators. Instead, K-12 schools will be able to independently activate or deactivate individual services, including Gemini, Gemini Notebook and Google Search. All these tools will remain switched off by default until administrators enable them. However, schools that reject AI entirely will also lose access to Google Search because of its integration with the technology. "We believe schools should have control over AI in their classrooms and in providing administrators with controls to help them manage their approach," Chris Phillips, Google's vice president and general manager of education, said in a blog post. Google also announced plans to improve its search services for educational purposes by introducing step-by-step guided learning features. These tools will help students work through problems rather than simply providing quick answers to their school assignments. Despite the changes, schools will not be able to completely avoid Google's AI technology while continuing to use its search services. Google is the leading provider of educational technology to schools across the United States. The company says around 150 million students and teachers worldwide use Google Classroom, while many students also rely on school-issued Chromebook computers. The company faced widespread criticism in August after enabling Gemini for many K-12 students. Before that change, the chatbot could only be activated for students aged 18 and above. Tuesday's announcement quickly drew criticism from educators and policymakers. Lina Khan, former chair of the US Federal Trade Commission, accused Google of exploiting its dominance in the search market to pressure schools into adopting AI. "Google is using its search monopoly to coerce schools into adopting AI," Khan wrote on social media, warning that allowing an unregulated dominant company to control an essential service could create serious risks. She added that as public opposition to AI in schools grows, Google is making it clear that institutions refusing to adopt the technology could face consequences. New York City and Los Angeles, which operate the country's two largest school districts, have temporarily restricted generative AI tools in elementary and middle schools while assessing their safety, educational impact and appropriate use. Although many students already rely on AI for schoolwork and homework, critics warn that the technology can make cheating easier and encourage students to seek instant answers instead of developing critical thinking skills. Experts have also questioned the potential effects of AI exposure on children's brain development. Josh Golin, executive director of children's advocacy group Fairplay, said Google's attempt to pressure schools into adopting its AI products could ultimately prove counterproductive. He argued that Google's presence in schools is less about improving education and more about building long-term customer loyalty to its products. "Because AI isn't safe for kids and undermines education and critical thinking, schools should turn off Google's AI tools the minute that option becomes available," Golin said in a statement. Rather than worrying about losing access to Google Search, Golin encouraged schools to consider abandoning Google's services altogether. He urged educational institutions to explore more privacy-focused search alternatives and encourage students to conduct deeper research beyond simply searching and clicking on online results.

The United Kingdom is likely to face artificial intelligence-driven cyberattacks on its critical infrastructure by 2028 as criminals and hostile states increasingly use the technology for malicious activities, Britain's cyber chief has warned. Richard Horne, head of the National Cyber Security Centre (NCSC), said AI could help hackers develop malware for ransomware attacks and identify and exploit long-standing security weaknesses on an unprecedented scale. In a speech marking the publication of the NCSC's annual cyber report on Tuesday, Horne warned that outdated technology and inadequate investment in cybersecurity could disrupt essential services and business operations. He cited recent attacks on refrigeration systems in French supermarkets, where hackers exploited weaknesses to raise temperatures and cause food to thaw. Similar vulnerable systems were also found in the UK, prompting NCSC experts to work through the night to alert affected organisations. The attacks did not involve AI, Horne said. Instead, hackers exploited outdated refrigeration technology connected directly to the internet and protected by default passwords, highlighting the risks posed by poorly secured operational systems. The report described Chinese cyberactivity as the most sophisticated and sustained cyber challenge facing the UK. It said Russia and China were expected to use AI routinely to advance their geopolitical objectives. Horne also said Russia was deliberately conducting cyber campaigns below the threshold of armed conflict, while Iran was targeting the UK and individuals considered threats to the Iranian government. The NCSC said AI was highly likely to increase the scale and efficiency of hostile cyber operations by both state actors and criminals. However, it is more likely to enhance existing methods than to uncover entirely new vulnerabilities. Horne warned that digital technology now controls essential physical processes, including electricity switching, water facility gates, dam operations, manufacturing lines, and heating and refrigeration systems. He said business leaders who had relied on outdated or inexpensive technology would have to explain the consequences if essential supplies were disrupted, businesses collapsed or public services failed because of inadequate cybersecurity investment. The warning comes after governments and companies last week rushed to introduce security updates following the AI-assisted discovery of vulnerabilities in PHP, a programming language used by around 70 percent of websites worldwide. Two critical flaws identified with AI assistance could allow attackers to access websites without authentication and potentially take remote control of web servers without passwords. Russia has already been accused of targeting critical infrastructure in Europe. The UK and several European Union member states said Russia's FSB security service was behind a December 2025 attack on Poland's energy grid. A successful attack could have left 500,000 people without electricity during winter. According to the NCSC, malicious cyberactivity costs UK organisations nearly $20 billion annually. Carmaker Jaguar Land Rover suffered more than $260 million in direct losses from a cyberattack in 2025, which also affected its suppliers and harmed the British economy. Anne Keast-Butler, director of GCHQ, Britain's signals intelligence agency under which the NCSC operates, urged all UK organisations to take immediate action to strengthen their cybersecurity. She said cybersecurity had become part of the country's frontline national defence and was essential to protecting customers, services and livelihoods. The NCSC advised organisations to adopt basic but effective safeguards, including keeping devices and software updated, restricting administrator access to systems and regularly backing up important files and data.

Apple Inc. has reported stronger-than-expected revenue for its fiscal fourth quarter, marking an encouraging start for new CEO John Ternus, who took over the company's leadership in September. The upbeat results came as several major technology companies, including Alphabet, Amazon and Intel, reported disappointing quarterly performances. The California-based technology giant earned $30.71 billion, or $2.10 per share, during the July-September quarter, up 12 percent from $27.47 billion, or $1.85 per share, in the same period last year. Apple said the latest earnings included 5 cents per share from tariff refunds. Revenue also increased 12 percent to $114.5 billion from $102.5 billion a year earlier. According to a FactSet poll, analysts had projected earnings of $1.99 per share on revenue of $113.4 billion. Jacob Bourne, an analyst at Emarketer, said Apple's strong performance came amid a recent sell-off in technology stocks, as investors questioned the huge costs and uncertain returns associated with the artificial intelligence boom pursued by other major technology companies. He said record iPhone revenue and strong growth in China demonstrated the resilience of Apple's business. While competitors face investor concerns over their massive AI investments, Bourne said Apple's relatively modest spending on AI infrastructure was helping it preserve cash flow. He added that partnerships with companies such as Alphabet could strengthen Apple's AI capabilities without exposing it to the same level of financial risk faced by its rivals. Despite the strong quarterly performance, Apple warned that its closely watched gross margin, which stood at 47.9 percent in the fourth quarter, would decline to between 45 percent and 46 percent in the December quarter. The company attributed the expected decline mainly to higher memory costs and expenses associated with the release of its AI-powered Siri. Apple shares were initially little changed after the earnings announcement but fell 2.6 percent to $329.60 during the earnings call. The stock had already closed down 1.8 percent at $338.32 in regular trading.

AI agents are gaining popularity beyond the technology sector, with companies including Meta, OpenAI and Anthropic competing to offer tools that can independently handle users' personal and professional tasks. Unlike traditional AI chatbots, these agents can proactively perform activities such as managing schedules, writing emails, booking travel and shopping. However, choosing the right tool and deciding how much personal information to share can be challenging. Experts recommend using a combination of agents rather than relying on one, as each has different strengths. Business consultant Jose Artigas suggests Claude Duet for starting a business, Dots for technical problem-solving and Instinct for routine tasks such as emails and travel bookings. He recommends Meta's Muse for beginners because of its user-friendly features. Experts also advise testing different agents and starting with repetitive online tasks that users want to delegate. Clara Gigling of Oudae Labs recommends providing clear instructions, specific goals and necessary context to achieve better results. Users should also consider their privacy before connecting agents to emails, calendars, identification details and financial accounts. While companies have introduced privacy policies explaining data handling, cloud-based operations and memory controls, Gigling urged users to remain cautious when sharing sensitive information.

A China-linked online influence campaign known as Spamouflage has resurfaced ahead of the US midterm elections, with suspected accounts impersonating American voters and sharing divisive posts about US politics, social issues and the electoral process, according to online analysts. In a blog post published Friday, Graphika, a New York-based firm that monitors online activity, said it had detected Spamouflage-related accounts on X using tactics similar to those identified before the 2024 US presidential election. The accounts have been spreading messages questioning the integrity of elections and the legitimacy of American democracy, the company said. US intelligence and national security officials have repeatedly warned that Russia, China and Iran operate online influence campaigns targeting American voters ahead of elections. Unlike Russia, which has been accused of favouring particular candidates, China's efforts have largely focused on weakening public trust in US elections, voting and democratic institutions. However, Liu Chang, a spokesperson for the Chinese Embassy in Washington, denied the allegations, saying China "has never and will never interfere in the U.S. elections." Darren Linvill, co-director of Clemson University's Watt Family Innovation Center Media Forensics Hub, said the campaign was unlikely to influence actual political debates. "Spamouflage is famously bad at mimicking real Americans," Linvill said, adding that the operation was not effective at shaping genuine political discussions. Instead, the network has been more successful in restricting the flow of information, Linvill said. He noted that accounts associated with the campaign frequently target and harass members of the Chinese diaspora, reflecting Beijing's broader efforts to suppress opinions it opposes. Graphika also identified an account that appeared to have replaced a previously removed Spamouflage-linked account. The account has shared divisive political messages, including criticism of Democratic Party leaders during the October 2025 US government shutdown and attacks on immigration policies. The firm further identified four accounts with a history of Spamouflage activity that now promote MAGA-style identities. Another account, which has more than 200,000 followers and identifies itself as a Philippine culture blog, has shifted to sharing pro-China messages and divisive content targeting US audiences. Graphika said it was unable to establish whether the latter account had been taken over by another operator or whether its original owner had sold it or otherwise lost control.