United News of Bangladesh

Apple on Wednesday unveiled its latest iPhone lineup, including the much-anticipated foldable iPhone Duo, as new CEO John Ternus made his debut at the company’s annual product launch. Ternus, who succeeded Tim Cook as Apple CEO on Sept. 1, presented the new devices at Apple’s headquarters in Cupertino, California. After highlighting updates to the iPhone, Apple Watch and AirPods, Ternus told the audience at the Steve Jobs Theater that “actually there is one more thing,” echoing the trademark phrase associated with Apple co-founder Steve Jobs. The foldable Duo features the largest display ever offered on an iPhone while remaining pocket-sized, Ternus said. When unfolded, the device has a 7.6-inch display, which is 80 percent larger than that of the new iPhone 18 Pro. It also supports the Apple Pencil, a feature previously available only on iPads. Ternus said the Duo would “redefine the experience of using a foldable phone.” The device allows users to split the screen between two different apps or open two windows of the same website, such as when comparing products while shopping online. Starting at $1,999, the Duo is the most expensive iPhone Apple has introduced. Ben Wood, chief analyst at CCS Insight, said the foldable device was always expected to overshadow the other announcements at the event. The Duo brings a distinctive design to the iPhone lineup and marks Apple’s entry into the niche foldable smartphone market, he said. Although demand for the Duo is expected, Wood noted that foldable phones still account for less than 5 percent of the global smartphone market. New iPhone 18 models The iPhone 18 Pro and iPhone 18 Pro Max come with improvements to battery life, cameras and overall performance, Ternus said. The two models will start at $1,199 and $1,299, respectively. Preorders will begin Sept. 12, with the phones becoming available Sept. 18. The Duo will go on sale Oct. 23. The new models offer up to 45 hours of video playback on the Pro Max and support faster charging. Apple has also introduced new camera features, including what it calls “cinematic effects” for regular video, enhanced manual controls and “smart focus tracking,” which allows users to keep a moving subject in focus, such as a child playing soccer. The company is also highlighting new artificial intelligence capabilities for Siri, with an emphasis on privacy and security as Apple seeks to narrow the gap with rivals in AI. Samsung, Motorola and Google are among the major competitors that have already introduced foldable smartphones. Analysts said Apple’s entry reflects its traditional approach of waiting for emerging technologies to mature before entering a new category. “The foldable launch is straight out of Apple’s playbook,” said Forrester analyst Dipanjan Chatterjee, adding that Apple typically waits for products to overcome their initial shortcomings before entering and shaping a market. The key question, he said, is whether Apple can expand demand for what remains a niche product category. The iPhone 18 lineup is about $100 more expensive than earlier generations, partly because of a global shortage of memory chips. The iPhone 17 ranged from $799 to $1,999, depending on the model. Earlier this year, Apple announced advances in artificial intelligence, including upgrades to Siri, while emphasizing privacy and practical everyday applications as it works to catch up with competitors. Ternus takes center stage The launch event also provided Ternus with an opportunity to outline his vision for Apple’s future, Chatterjee said. Ternus inherits one of the most successful consumer technology companies in modern history, presenting both an advantage and a challenge, he said. Chatterjee said Ternus would need to build on Cook’s execution, scale and ability to monetize Apple’s products while preparing the company for a future in which smartphones may no longer be consumers’ primary gateway to the world. Wood, meanwhile, said the attention surrounding the Duo made Ternus’ formal arrival as CEO less prominent than expected. The event represented a smooth transition from Cook, and its timing alongside the launch of the iPhone Duo was likely deliberate, Wood said. Despite the intense interest in the foldable phone, he said Apple’s broader ecosystem of hardware, software and services would remain the main driver of the company’s future growth.

A new documentary on Elon Musk uses a computer-generated avatar and artificial intelligence-generated voice synchronization to recreate an interview with the billionaire entrepreneur, using statements he has made publicly in the past. Oscar-winning filmmaker Alex Gibney, speaking ahead of the film's world premiere at the Venice Film Festival on Tuesday, said the approach was deliberately ironic. He said it reflected Musk's belief that humanity may be living in a simulation while also highlighting what he described as the shifting and contradictory nature of Musk's views on technology. Gibney said the technique was intended to show the contrast between the reality of Musk's actions and the narrative surrounding him. The filmmakers consulted lawyers before proceeding and obtained approval to use the digital representation. Ashley St. Clair, a former conservative influencer who has a child with Musk, also attended the festival and features prominently in the documentary. She discusses her relationship with Musk, his personality and a nondisclosure agreement she said she was asked to sign in connection with financial support and paternity. St. Clair said she declined the agreement. “I was offered the opportunity not to speak for $40 million,” she said. Earlier this year, St. Clair sued Musk's artificial intelligence company, alleging that its Grok chatbot enabled users to create sexually exploitative deepfake images of her, causing her humiliation and emotional distress. The nearly four-hour documentary also examines Musk's acquisition of X, formerly Twitter, and the changes made to its algorithm. Gibney said he had not fully understood before making the film how Musk had altered the platform's algorithms to give greater prominence to hostile and inflammatory content. Wired reporter Zoë Schiffer, who also appears in the film, said the changes included suppressing mainstream journalism while boosting more provocative voices and comments. St. Clair said she believes such actions contribute to deliberate social division and could have serious consequences without regulation or containment. From Silicon Valley to politics“Musk” traces Musk's life from his difficult childhood, including an interview with his father, Errol Musk, through his early years in Silicon Valley and his later political activities. The documentary also covers his work with the Department of Government Efficiency, or DOGE, during Donald Trump's second administration and his involvement in right-wing political movements internationally. Gibney said the project was initially planned as a considerably shorter film but expanded as Musk became increasingly involved in Trump's political campaign. Among those interviewed for the documentary are Tesla co-founder Martin Eberhard, Musk's first wife Justine Musk, neurologist Sam Harris, technology journalist Kara Swisher and former USAID Administrator Samantha Power. The film, which is scheduled for a theatrical release in October, also raises concerns over the access and information-gathering powers exercised by young officials associated with DOGE across multiple US government agencies.

China’s exports rose 25 percent in August compared with the same month a year earlier, driven by strong overseas demand for automobiles and high-tech products, while the country’s already large trade surplus grew further, customs authorities said Tuesday. The figures were largely consistent with economists’ forecasts and were released ahead of an expected meeting between Chinese President Xi Jinping and US President Donald Trump later this month. Beijing has yet to announce the exact date of the meeting. Trade is expected to feature prominently in discussions between the two leaders. China’s imports from around the world increased 28.2 percent in August year-on-year, accelerating from a 27.5 percent rise in July. Export growth also picked up from 23.9 percent in July. As a result, China’s trade surplus increased to $119.1 billion in August from $112.5 billion the previous month. The growing surplus has drawn criticism from the United States and other economies. China recorded a record annual trade surplus of $1.2 trillion last year. Beijing, however, maintains that it does not deliberately pursue a larger trade surplus. Chinese exports to the United States were valued at $42.5 billion in August, marking a 34.4 percent increase from a year earlier. The sharp rise was partly attributed to a lower comparison base following a decline in exports last year after higher US tariffs were introduced. US exports to China stood at $13.3 billion, giving China a surplus of about $29.2 billion, according to Chinese customs figures. Exports to the European Union increased 6.6 percent, while shipments to Southeast Asia and Latin America rose 30.2 percent and 17.5 percent, respectively. Lynn Song, chief economist for Greater China at Dutch bank ING, said exports have continued to grow faster than imports and are likely to push China’s trade surplus to another record this year. China has also managed to withstand disruptions caused by the Iran war better than many other economies. Its expanding trade with Southeast Asia, Latin America and Africa has helped offset some of the effects of higher US tariffs. Customs data showed automobile exports increased 43 percent in August from a year earlier, while semiconductor shipments surged 129.8 percent. Chi Lo, a senior Asia-Pacific market strategist at BNP Paribas Asset Management, said China remains highly competitive in technology-related exports. He noted that strong shipments of electric vehicles, industrial equipment and semiconductors have contributed to the country's export performance. Lo said China has rapidly advanced toward higher-value industries and established itself as a significant force in artificial intelligence infrastructure and industrial automation. Despite strong external demand, China's domestic economy continues to face difficulties. Consumer spending and investment remain weak following years of problems in the property sector. On Sunday, Beijing announced plans to inject about $54 billion into state-owned banks and insurers as part of efforts to support economic growth. China’s heavy dependence on exports has also raised concerns among its major trading partners. At a recent meeting of senior financial officials in Asheville, North Carolina, 19 members of the Group of 20 agreed to work toward reducing global economic imbalances. China was the only G20 member that did not support the agreement after US Treasury Secretary Scott Bessent described China’s trade surplus as an obstacle to global economic growth. Lo said the strategic tensions between China and the United States are likely to continue. He noted that both countries have leverage over strategically important products, with Washington restricting exports of advanced technology to China while Beijing has limited supplies of rare earths to the US. China and the European Union are also expected to hold ministerial-level trade discussions later this year. The EU is seeking ways to narrow its trade deficit with China, which is estimated at around 1 billion euros per day. In July, the EU introduced measures aimed at supporting its steel industry and restricted tax-free imports of small Chinese e-commerce packages.

Researchers at the US Department of Energy’s Princeton Plasma Physics Laboratory (PPPL) and Princeton University have developed an artificial intelligence-based software framework capable of making rapid decisions to control fusion plasma at speeds beyond human response. The framework, called PACMAN, or Prediction And Control using MAchiNe learning, was successfully tested in five experiments on the DIII-D National Fusion Facility tokamak in San Diego, according to researchers. Their findings were published in the journal Nuclear Fusion. Fusion research seeks to replicate the process that powers the sun to produce a potentially abundant source of electricity. Tokamaks use powerful magnetic fields to confine extremely hot plasma, but keeping the plasma stable requires continuous adjustments to heating systems, magnets and gas injectors. Plasma instabilities can develop within milliseconds, making them too fast for human operators to address. Computer simulations can predict plasma behavior, but some simulations take days or months and therefore cannot provide real-time control during experiments. Machine-learning models, however, can predict plasma behavior within milliseconds, researchers said. PACMAN integrates multiple AI models into a single control system. It collects real-time measurements such as plasma temperature, density and magnetic signals, checks the data, and uses AI models to assess the plasma's current and predicted behavior. Controllers then use those predictions to determine necessary adjustments, such as changing heating power. Before commands are sent to the tokamak, PACMAN checks them against hardware safety limits and resolves conflicting instructions. During five experiments at DIII-D, the system demonstrated several capabilities. It allowed a reinforcement-learning model to control plasma heating, predicted bursts of energy from the plasma edge, detected and controlled waves caused by fast particles, adjusted plasma density and rotation, and predicted a potentially disruptive instability known as a tearing mode. In one experiment, the AI predicted the tearing mode about 200 milliseconds before it was expected to occur, allowing researchers to adjust the plasma to prevent the instability rather than suppressing it after it began. PACMAN also coordinated all six of DIII-D's gyrotrons, which heat plasma using powerful microwave beams. The system simultaneously adjusted their power and mirror positions to achieve targets set by researchers. Researchers said the framework could significantly speed up the development and testing of new AI models. While integrating the first model took months, adding a second model took only a few days, allowing researchers to test, retrain and replace models more quickly. The researchers stressed that PACMAN is not designed to eliminate human oversight. Human operators continue to set the control objectives and parameters, while the framework enforces hardware safety limits regardless of AI recommendations. Egemen Kolemen, an associate professor at Princeton University and PPPL, said PACMAN's modular design could allow AI algorithms to be added, replaced or operated simultaneously without requiring changes to the rest of the system. The researchers believe the framework could eventually be adapted for tokamaks of different designs and sizes, potentially providing a common platform for AI-based plasma control across future fusion facilities. Source: Science Daily  

Sudan's prolonged war and worsening electricity crisis are driving more residents to turn to solar power, but the high cost of installing solar systems is putting the alternative energy source beyond the reach of many households. For 31-year-old housewife Randa al-Mansour, solar panels installed on the roof of her building have provided a lifeline through repeated power outages in Omdurman, just outside Khartoum. She said the frequent blackouts have made daily life particularly difficult for her three children, who often return home from school because there is no electricity to run classrooms or fans. Sudan's military and the paramilitary Rapid Support Forces have been fighting since April 2023. The conflict has killed at least 59,000 people and displaced around 13 million, according to figures cited by The Associated Press. The war has also severely damaged the country's already fragile electricity infrastructure. The UN Development Programme estimates that damage to the power grid has reached about $3 billion. As outages become increasingly frequent, advertisements for solar panels, batteries and inverters have become common across the capital. But the cost remains a major barrier. Mohey Eddin Abu-Bark, who runs a solar systems business in Khartoum, said a typical 10-kilowatt household system costs nearly $5,000. “For a low-income family, installing a solar panel system is a huge sacrifice but an inevitable choice,” he said. Many Sudanese families cannot afford such an investment. Manahel Madany, a 43-year-old civil servant with diabetes, said power cuts and extreme heat have repeatedly ruined her insulin supply because she cannot afford a solar system to keep her refrigerator running. She said she is sometimes forced to discard her insulin every few days after it spoils. The electricity crisis is also affecting the availability and storage of medicines. Madany said many pharmacies cannot afford solar systems needed to maintain proper storage conditions, contributing to shortages of essential drugs. The Sudan Doctors' Network has warned that prolonged power outages could have serious consequences for hospitals and patients, particularly those requiring continuous treatment such as kidney patients. According to the World Bank, about 30% of Sudan's population has no access to electricity, with rural communities particularly affected. Even in urban areas connected to the national grid, outages have reportedly lasted as long as 16 hours a day in recent months. Electricity generation has fallen sharply since the war began. Two electricity officials told AP that production has dropped from around 4,400 megawatts before the conflict to about 1,100 megawatts. They said several power plants have been damaged by drone attacks attributed to the RSF. Nearly half of the country's electricity engineers and technicians have also left Sudan since 2023, according to a survey by a local labour union. The officials spoke anonymously because they were not authorised to speak publicly. Sudan has strong potential for solar power, with the country receiving more than 4,000 hours of sunshine annually in some areas, according to the United Nations. However, the high initial cost of imported solar equipment, combined with the depreciation of Sudan's currency, continues to make the transition to solar energy difficult for millions of people.

The humble television remote control has transformed the way people watch TV, allowing viewers to switch channels without leaving the comfort of their couches and even influencing how television programs are produced. For people under 40, it may be hard to imagine a time when changing channels required getting up and turning a dial on the television. But that changed in 1950, when Zenith Electronics introduced the first TV remote, known as the “Lazy Bone.” The device was not wireless, meaning viewers still had to deal with a cable that could easily become a tripping hazard. Five years later, Zenith introduced the first wireless remote, the “Flash-Matic,” which used light signals to change channels. However, sunlight could trigger unwanted channel changes. In 1956, the company introduced the “Space Command,” which relied on sound waves. The clicking sound made when its buttons were pressed helped give rise to the term “clicker.” Infrared technology became common in remotes during the 1980s, just as the expansion of cable television dramatically increased the number of available channels and made remote controls almost essential. David Bianculli, a television critic for NPR's “Fresh Air” and a professor at Rowan University, said losing a remote in the cushions of a sofa could become a major crisis for viewers. The remote also changed television programming. Before viewers could easily switch channels, broadcasters placed great importance on maintaining a smooth sequence of programs to keep audiences from tuning elsewhere. Networks such as CBS and NBC developed carefully planned schedules to retain viewers. CBS's Saturday night lineup in the 1970s and NBC's “Must-See TV” Thursday nights became examples of successful programming strategies. The rise of cable television also created the “channel surfer,” a viewer who rapidly moved through numerous channels looking for something interesting. As a result, eye-catching programming became increasingly important. Despite its convenience, the remote has its drawbacks. It can encourage arguments over who controls the television, accumulate in multiple versions for different devices and inevitably stop working when batteries run out. With modern appliances increasingly capable of responding to voice commands, the traditional remote could eventually become less important. Until then, the search for the missing “clicker” remains a familiar household ritual. The late Zenith scientist Eugene Polley, who helped pioneer wireless remote technology alongside Robert Adler, once humorously compared the remote's importance with other major inventions, saying it was nearly as important as sex. Whatever its ultimate place in history, the remote has certainly given television viewers one enduring form of power: control over the clicker.

Elon Musk sent dozens of self-driving Cybercabs without steering wheels or brake pedals onto the streets of Austin, Texas, on Thursday, betting that Americans will overcome their fears of rides with “no control” and hop inside. The rollout of the gold-colored Teslas, which give passengers no way to take control in an emergency, comes as the company prepares to launch the service in other cities. Tesla shares have suffered amid a slump in vehicle sales, but they rose more than 5% Thursday on hopes that the futuristic-looking taxis will catch on quickly. “No steering wheel, no pedals,” Tesla teased on X before the launch. Musk followed with a post showing a giant Cybercab floating above the Austin skyline and later wrote, “A Storm of Cybercabs.” An invitation-only launch event was expected to be held later Thursday. It is unclear how soon Musk intends to roll out the Cybercab service, but he needs to move quickly. Tesla trails self-driving taxi leader Waymo in the number of cabs deployed and trips completed. To catch up, it must demonstrate that its camera-only system can safely navigate streets and avoid pedestrians. By contrast, Waymo and another rival, Amazon’s Zoox, supplement cameras with radar and a laser-based technology called lidar. Even if the technology works well, Americans still need to be convinced. A Pew Research Center survey conducted in February found that seven in 10 U.S. adults were “not too” or “not at all” comfortable riding in a driverless car. Tesla stock was battered last year after Musk took over President Donald Trump’s government cost-cutting campaign, dubbed DOGE, and embraced extreme-right political candidates, sparking protests at Tesla showrooms and boycotts by car buyers in several countries. Tesla posted a second consecutive annual decline in vehicle sales last year. Its profits plunged, and it lost its crown as the world’s bestselling electric vehicle maker to China’s BYD. Austin has had a self-driving Tesla “robotaxi” service since June last year, but those cars are equipped with steering wheels and brake pedals. The service, which has since expanded to five other cities in Texas and Florida, initially carried passengers with safety drivers aboard to take over in case of problems. Tesla has more than 200 “unsupervised” robotaxis in those cities, meaning they operate without safety drivers aboard, according to the monitoring site RobotaxiTracker. Waymo has more than 4,000 such vehicles in 14 cities. In the Pew survey, 16% of U.S. adults said they would be “somewhat” comfortable riding in a driverless car, while 7% said they would be “extremely” or “very” comfortable. Separate Gallup polling conducted in 2025 found that skepticism about the safety of driverless cars had risen over the previous several years. More Americans said all or mostly human-operated cars were the safest option than in a 2018 poll. It is unclear whether Musk is the right person to coax people into cars with no human controls. Another Pew poll conducted in January found that nearly six in 10 U.S. adults had a “very” or “mostly” unfavorable view of him. Musk eventually hopes to send full self-driving software to hundreds of thousands of Teslas through a software update. That would allow Tesla owners to turn their cars into taxis for hire when they are not using them. But regulators in the U.S. and overseas have signaled that he still has some convincing to do. Tesla has not received approval for drivers to use even its partial self-driving software in most European Union countries. In the U.S., where the software has been downloaded and used, regulators have opened several investigations into its safety. One probe is examining the software’s role in several Tesla crashes in fog, sun glare and other low-visibility conditions, including one in which a pedestrian was killed. Another is investigating dozens of incidents in which Teslas using partial self-driving software ran red lights or drove on the wrong side of the road, sometimes crashing into other vehicles and causing injuries. A third is examining why Tesla apparently violated regulatory requirements by failing to report crashes promptly.  

A federal judge on Wednesday ordered Google to make changes to the system underpinning its monopoly in digital advertising, but stopped short of ordering the breakup sought by the US government. The initial two-page ruling by US District Judge Leonie Brinkema in Virginia marks the second time in a year that Google has avoided a Justice Department proposal aimed at dismantling parts of its internet business. Meanwhile, practices that courts have found to be anti-competitive have helped enrich Google's parent company, Alphabet Inc., which has a market value of $4.11 trillion. Brinkema's full opinion, including details of the remedies, will remain sealed for 14 days to allow both sides to review it and suggest any necessary redactions. “We're very pleased the Court rejected the DOJ's proposal to break apart tools that help small businesses reach new customers and grow,” said Lee-Anne Mulholland, Google's vice president for regulatory affairs. After a judge ruled in 2024 that Google's widely used search engine constituted an illegal monopoly, the Justice Department sought remedies that would have required the company to sell its popular Chrome web browser. That proposal was rejected last September by another federal judge overseeing the case. Brinkema has now similarly rejected the government's argument that Google should be required to sell parts of the technology behind a complex advertising system that generates a significant share of the company's nearly $400 billion in annual advertising revenue. Critics of major technology companies condemned the decision. “It takes an Olympic level of mental gymnastics to find that Google is operating an illegal monopoly and then decide to do nothing about it,” said Sacha Haworth, executive director of The Tech Oversight Project. “With Big Tech continuing to suffocate new and innovative businesses from gaining traction, Judge Brinkema, like (US District Judge Amit) Mehta before her, is sending the wrong message at the wrong time,” Haworth said. Brinkema's approach contrasts with the position strongly opposed by the Justice Department last November, when she heard closing arguments in the trial's remedies phase. The ad-tech case began with the government's antitrust lawsuit filed in 2023, which resulted last year in a ruling that Google had illegally maintained a monopoly. Justice Department lawyers warned Brinkema in court filings that allowing Google to retain all of its advertising technology would still enable the company to “manipulate computer algorithms that are the engine of its monopolies in ways too difficult to detect.” Google argued in response that breaking up the complex network would be unnecessarily harsh. The system distributes advertisements that online publishers rely on to help finance digital services used by millions of consumers. The advertising exchange technology at the center of the case processes an estimated 55 million requests per second, according to figures submitted by Google in court filings. “This is technology that absolutely has to keep working for consumers,” Google said ahead of Brinkema's ruling. In her decision, Brinkema said she had agreed to most of the remedies proposed by both sides. Although the changes could potentially reduce Google's revenue or slow its future growth, the ruling is likely to be viewed by many investors as little more than a speed bump for a company that has recently benefited from the rapid growth of artificial intelligence technology. A similar investor reaction followed the penalties imposed by US District Judge Amit Mehta in the search monopoly case. Since Mehta's ruling in Washington, D.C., Alphabet's stock price has climbed 45%, creating an additional $1.3 trillion in shareholder wealth.

India is rapidly expanding its renewable energy capacity, but a growing amount of solar power is going unused because the country's electricity grid and storage systems are struggling to keep pace. Over the past 15 months, India curtailed nearly 11 terawatt-hours of solar power — enough electricity to supply around 10 million homes — according to government data and research by energy think tank Ember. The unused solar power comes at a time when India's electricity demand is rising sharply, driven by extreme heat, increased use of air conditioners, irrigation needs and the growing electrification of transport and other sectors. India now has more than 300 gigawatts of clean energy capacity, accounting for over half of its total installed electricity capacity. However, coal still generates most of the country's electricity. One major problem is known as curtailment, which happens when solar or wind farms are forced to reduce or stop generating electricity because the grid cannot safely absorb the available power. Solar generation usually peaks in the afternoon, but coal-fired power plants cannot always reduce their output quickly enough to make room for renewable electricity. “We're in a stage where some of the biggest hurdles in renewables are starting to hit us,” said Neshwin Rodrigues, an energy analyst at Ember. Another challenge is the concentration of solar projects in western India. Gujarat and Rajasthan together account for nearly half of the country's solar capacity, putting heavy pressure on transmission lines in the region. Experts say India needs to expand its power transmission network, develop more battery storage and make its electricity system more flexible. “Without enough storage, India risks keeping coal plants running even when cheap renewable power is available,” said Vibhuti Garg, South Asia director at the Institute for Energy Economics and Financial Analysis. Building new solar and wind projects can take around two years, but major transmission lines often require at least three years to complete. India has achieved about 80% of its annual transmission construction targets over the past five years, according to Ember. India aims to reach 500 gigawatts of clean electricity capacity by 2030 and expects non-fossil sources to account for nearly 70% of its installed power capacity by 2036. Energy experts say battery storage could help address the problem by storing excess solar power generated during the day and supplying it during the evening, when electricity demand remains high. India currently has about 3 gigawatts of battery storage and 7.4 gigawatts of operational pumped-storage capacity. The government expects the country's storage requirement to reach around 74 gigawatts by 2032. “If we try to increase the solar installations without solving for energy storage, it is only going to lead to curtailment,” said Ankit Mittal, CEO of battery storage company Ingro Energy. Industry experts say the rapid growth in electricity demand has caught many by surprise as transport, industries and other sectors increasingly shift to electricity and more data centres are built. The challenge for India now is not only to add more renewable energy projects, but also to build the transmission and storage systems needed to ensure that clean electricity can be used when and where it is needed.

Social media and technology companies are increasing their efforts to identify which of their users are children, as pressure from governments, regulators and the public grows over online safety. The issue has become a key part of Meta’s recently announced $18 billion legal settlement over allegations that its Facebook and Instagram platforms harmed children’s mental health by deliberately designing features that kept young users engaged. The settlement, involving 48 US states, the District of Columbia and US territories, places strong emphasis on improving age assurance — the process of determining whether a user is a child or an adult. The challenge is simple but difficult: safety measures designed for children are of limited use if online platforms cannot accurately identify which users are underage. Other major technology companies, including Google, TikTok and Roblox, are also strengthening their systems for identifying younger users. Under the settlement, Meta will have one year to improve its age-checking system using its own technology as well as third-party tools. The company will also face regular independent audits to assess how effectively the system works. AI increasingly used to estimate users' agesIn the physical world, people often use government-issued identification cards to prove their age. Online, however, many users are uncomfortable sharing documents such as passports or driving licences with large technology companies. Meta currently accepts identification documents for age checks and says it deletes them within 30 days. But companies are increasingly turning to artificial intelligence to estimate a user's age. Unlike traditional age verification, AI cannot confirm a person's actual date of birth. Instead, it examines different clues to estimate whether someone is likely to be a child or an adult. Google began using an AI-based age assurance system for YouTube last year, analysing users' viewing behaviour to help distinguish adults from minors. Meta also uses AI to identify potentially underage accounts. On Instagram, the technology can analyse clues such as posts about birthdays or school, the type of content users like or comment on, their online activity, friendships and the times of day they use the platform. Meta has also said its AI can examine photos for general visual clues, such as physical features that may help estimate a person's age. The company says it does not use the system to identify the specific person in an image. Under the settlement, Meta has also agreed to set targets to reduce cases where minors are incorrectly identified as adults. If an account is found to belong to a child under 13 and is removed, Meta will also review the ages of the user's friends. Face scans also being usedSome technology companies are using video selfies and facial analysis to estimate users' ages. Identity verification firms such as Yoti and Persona provide such technology to companies including Meta and Roblox. Roblox, which allows children under 13 to use its platform, asks users to submit a video selfie if they want to access certain chat features. The company says the videos are deleted after the age-checking process is completed. However, facial age estimation has raised concerns over accuracy. Some users have reported incorrect results, while experts have questioned whether the technology works equally well across different groups, including women and people from different racial or ethnic backgrounds. Roblox has said its system can generally estimate the ages of users between about 5 and 25 within one or two years. Debate continues over who should verify users' agesMeta and some other companies have argued that app store operators, such as Apple and Google, should be responsible for verifying users' ages before allowing them to download apps. Apple and Google have opposed the idea, arguing that such a system could fail to cover people using desktop computers, shared family devices or apps already installed on a device. Privacy remains a major concernWhile age checks are intended to protect children from harmful or inappropriate online content, critics warn that widespread age verification could threaten users' privacy and anonymity. Digital rights groups have raised concerns that people may eventually have to prove their age to access a wide range of online services, including news websites, health information, entertainment platforms and other forms of online content. Critics also argue that collecting identity documents, facial scans or other personal information could create security risks if such data is leaked, hacked or misused. The debate highlights a growing challenge for technology companies and governments: how to better protect children online without creating a system that requires everyone to give up large amounts of personal information to use the internet.