BEIJING/SINGAPORE/NEW YORK — State-linked Chinese entities are using cloud services provided by Amazon or its rivals to access advanced U.S. chips and artificial intelligence capabilities that they cannot acquire otherwise, recent public tender documents showed. The U.S. government has restricted the export of high-end AI chips to China over the past two years, citing the need to limit the Chinese military's capabilities. Providing access to such chips or advanced AI models through the cloud, however, is not a violation of U.S. regulations since only exports or transfers of a commodity, software or technology are regulated. A Reuters review of more than 50 tender documents posted over the past year on publicly available Chinese databases showed that at least 11 Chinese entities have sought access to restricted U.S. technologies or cloud services. Among those, four explicitly named Amazon Web Services, or AWS, as a cloud service provider, although they accessed the services through Chinese intermediary companies rather than from AWS directly. The tender documents, which Reuters is the first to report on, show the breadth of strategies Chinese entities are employing to secure advanced computing power and access generative AI models. They also underscore how U.S. companies are capitalizing on China's growing demand for computing power. "AWS complies with all applicable U.S. laws, including trade laws, regarding the provision of AWS services inside and outside of China," a spokesperson for Amazon's cloud business said. AWS controls nearly a third of the global cloud infrastructure market, according to research firm Canalys. In China, AWS is the sixth-largest cloud service provider, according to research firm IDC. Shenzhen University spent $27,996 (200,000 yuan) on an AWS account to gain access to cloud servers powered by Nvidia A100 and H100 chips for an unspecified project, according to a March tender document. It got this service via an intermediary, Yunda Technology Ltd Co, the document showed. Exports to China of the two Nvidia chips that are used to power large-language models, or LLM, such as OpenAI's ChatGPT, are banned by the United States. Shenzhen University and Yunda Technology did not respond to requests for comment. Nvidia declined to comment on Shenzhen University's spending or on any of the other Chinese entities' deals. Zhejiang Lab, a research institute developing its own LLM, called GeoGPT, said in a tender document in April that it intended to spend 184,000 yuan to purchase AWS cloud computing services as its AI model could not get enough computing power from homegrown Alibaba. A spokesperson for Zhejiang Lab said that it did not follow through with the purchase but did not respond to questions about the reasoning behind this decision or how it met its LLM's computing power requirements. Alibaba's cloud unit, Alicloud, did not respond to a request for comment. Reuters could not establish whether the purchase went ahead. Moving to tighten access The U.S. government is now trying to tighten regulations to restrict access through the cloud. "This loophole has been a concern of mine for years, and we are long overdue to address it," Michael McCaul, chair of the U.S. House of Representatives Foreign Affairs Committee, told Reuters in a statement, referring to the remote access of advanced U.S. computing through the cloud by foreign entities. Legislation was introduced in Congress in April to empower the Commerce Department to regulate remote access of U.S. technology, but it is not clear if and when it will be passed. A department spokesperson said it was working closely with Congress and "seeking additional resources to strengthen our existing controls that restrict PRC companies from accessing advanced AI chips through remote access to cloud computing capability." The Commerce Department also proposed a rule in January that would require U.S. cloud computing services to verify large AI model users and report to regulators when they use U.S. cloud computing services to train large AI models capable of "malicious cyber-enabled activity." The rule, which has not been finalized, would also enable the Commerce secretary to impose prohibitions on customers. "We are aware the Commerce Department is considering new regulations, and we comply with all applicable laws in the countries in which we operate," the AWS spokesperson said. Cloud demand in China The Chinese entities are also seeking access to Microsoft's cloud services. In April, Sichuan University said in a tender document it was building a generative AI platform and purchasing 40 million Microsoft Azure OpenAI tokens to support the delivery of this project. The university's procurement document in May showed that Sichuan Province Xuedong Technology Co Ltd supplied the tokens. Microsoft did not respond to requests for comment. Sichuan University and Sichuan Province Xuedong Technology did not respond to requests for comment on the purchase. OpenAI said in a statement that its own services are not supported in China and that Azure OpenAI operates under Microsoft's policies. It did not comment on the tenders. The University of Science and Technology of China's Suzhou Institute of Advanced Research said in a tender document in March that it wanted to rent 500 cloud servers, each powered by eight Nvidia A100 chips, for an unspecified purpose. The tender was fulfilled by Hefei Advanced Computing Center Operation Management Co Ltd, a procurement document showed in April, but the document did not name the cloud service provider. Reuters could not determine its identity. The University of Science and Technology of China, or USTC, was added to a U.S. export control list known as the “Entity List” in May for acquiring U.S. technology for quantum computing that could help China's military, and for involvement in its nuclear program development. USTC and Hefei Advanced Computing Center did not respond to requests for comment. Beyond restricted AI chips Amazon has offered Chinese organizations access not only to advanced AI chips but also to advanced AI models such as Anthropic's Claude, which they cannot otherwise access, according to public posts, tenders and marketing materials reviewed by Reuters. "Bedrock provides a selection of leading LLMs, including prominent closed-source models such as Anthropic's Claude 3," Chu Ruisong, president of AWS Greater China, told a generative AI-themed conference in Shanghai in May, referring to its cloud platform. In various Chinese-language posts for AWS developers and clients, Amazon highlighted the opportunity to try out "world-class AI models" and mentioned Chinese gaming firm Source Technology as one of its clients using Claude. Amazon has dedicated sales teams serving Chinese clients domestically and overseas, according to two former company executives. After Reuters contacted Amazon for comment, it updated dozens of posts on its Chinese-language channels with a note to say some of its services were not available in its China cloud regions. It also removed several promotional posts, including the one about Source Technology. Amazon did not give a reason for removing the posts and did not answer a Reuters query about that. "Amazon Bedrock customers are subject to Anthropic's end user license agreement, which prohibits access to Claude in China both via Amazon's Bedrock API [application programming interface] and via Anthropic's own API," the AWS spokesperson said. Anthropic said it does not support or allow customers or end-users within China to access Claude. "However, subsidiaries or product divisions of Chinese-headquartered companies may use Claude if the subsidiary itself is located in a supported region outside of China," an Anthropic spokesperson said. Source Technology did not respond to a request for comment.

Abuja, Nigeria — A U.S. State Department official was in Nigeria this week to meet with local and regional authorities about the responsible use of artificial intelligence in military applications. Mallory Stewart, assistant secretary of state for the Bureau of Arms Control, Deterrence and Stability, said her two-day visit with Nigerian officials from the regional bloc ECOWAS was part of the United States’ commitment to deepen security cooperation in Africa. The U.S. government has been working with 55 nations, including African nations, “to agree upon responsible uses of AI in the military context, using AI in a manner consistent with international laws [and] recognizing inherent human bias,” Stewart told journalists Wednesday. “We've learned the hard way [that there is] inherent human bias built into the AI system ... leading to maybe misinformation being provided to the decisionmaker,” she said. The goal, she continued, "is to hear from as many countries as possible that are at the stage of working in artificial intelligence to their military to see how we can minimize the risks." Last year, the Global Terrorism Index report named sub-Saharan Africa an epicenter of terrorism, accounting for nearly 60% of terror-related deaths. It is unclear whether the terror groups are using AI. Nigerian authorities have been pushing for the integration of artificial intelligence in military operations, while acknowledging that adopting AI will require Africa-specific policies. Security analyst Kabiru Adamu of Beacon Consulting said the use of AI in military operations has advantages. "Given the position of the U.S. in terms of its military capacity and technological advancement, it will definitely be in the position to support Nigeria's desires, especially if it's able to contextualize some of the peculiarities within the Nigerian security space,” Adamu said. “We can't isolate ourselves from the international committee of nations. AI is embedded in security, so we have to do it. But we need to be cognizant of the supporting infrastructure for good technology. Power is one of them, culture." The founder of Global Sentinel online magazine, Senator Iroegbu, said that while AI has benefits, the technology still needs to be treated with caution. "It limits casualties in terms of the number of soldiers that will be deployed, so you conserve your boots,” Iroegbu said. “It helps penetrate rough terrains, gather more intelligence. It's good that there's growing awareness of the issue of artificial intelligence, but Nigeria needs to first of all try to define its own policy and strategy with regards to artificial intelligence. More sensitization needs to be done, and more policy aspect of it needs to be developed.” In June, African ministers unanimously endorsed landmark continental AI strategy to advance Africa's digital future and development aspirations. And last week, the African Union approved the adoption of AI in public and private sectors in member states, including Nigeria.

Nairobi, Kenya — Overcrowding in African hospitals is blamed on the scarcity of health facilities and doctors, especially in rural areas. According to the United Nations, there is only one doctor for every 5,000 people in Africa, a continent that bears 25% of the global disease burden. But with the number of mobile phone users on the rise, some technological innovations are helping to bridge the doctor-patient gap and expand health care coverage.  Yaw Asamoah is head of MedPharma Care in Ghana. The company has developed an app that allows patients to connect face-to-face with doctors and pharmacies online so they can get medicine in their homes. He says the system improves patients' experiences when they seek health care services. "That's where MedPharma care comes in to see how we can digitize the whole idea of health care bringing telemedicine — making it possible for people either [to] have e-consultation, e-prescription, get their medicine delivered to them wherever they are, either at the office or at home... do their diagnostic remotely,” Asamoah said. The World Health Organization says 57 countries are suffering from a critical shortage of health personnel, 36 of them in Africa. The 2001 Abuja Declaration requires that African Union countries allocate 15% of their annual budgets to health, a requirement most governments have yet to fulfill. Funding and infrastructure issues have blocked millions of Africans' access to quality health care, but experts say digital tools could improve access to services in hard-to-reach areas that lack doctors. Mountaga Keita is a Guinean-born businessman who invented three portable diagnostic terminals which can monitor a patient’s temperature, blood pressure, heart function and conduct ultrasounds. "The benefit of that is the ease it brings to doctors and patients instead of clogging hospitals,” Keita said. “Now the doctors or nurses can get to the patient collected data and send the data in a very secure manner to the hospitals, and people can analyze and bring it back to the patient." Keita has so far deployed 40 kits to different hospitals in Guinea. According Keita, the diagnostic terminals have attracted the attention of other countries like Gabon, which has requested six machines. He is in talks with the governments of Burkina Faso, Ivory Coast, and Senegal to supply the kits there. Keita said his technology can help solve the doctor-to-patient ratio problem and save patients money.    "With this kind of technology, all the vital signs of a patient, forward it in a very secure manner, encrypted to a specialist who is in Tunisia, who is in Kenya, who is in Tokyo, Paris to interpret and bring the result,” he said. “Then we know if we are supposed to spend that 45,000 euros to evacuate ... or if we can locally cure the person." Since the COVID-19 pandemic, telemedicine has grown expansively and gained attention in Africa. Asamoah said telemedicine provides access to many doctors who specialize in different diseases, easing the burden on health care facilities.  "In a normal circumstance, if you went to a clinic in Ghana, you wanted to talk to a specialized consultant, you might probably not get either because they don't have, they haven't booked you, or they are not available,” he said. “But telemedicine can make it possible for you to make your appointment and talk to any doctor." McKinsey & Company, a global management consulting firm, analyzed the impact of digital health tools in Kenya, Nigeria, and South Africa and found that the tools could reduce the continent’s total health care costs by 15% by 2030.

Some bosses might want all their workers back in the office, but a clear majority of U.S. workers crave the flexibility of remote or hybrid jobs. And studies have found that such work, if managed well, will not harm a company's culture or capability to innovate. But some companies concerned about productivity are using software to monitor employees working from home. Maxim Adams has the story. Camera: Aleksandr Bergan

Manzini, Eswatini — Eswatini has launched an initiative to achieve sustainable development by harnessing the power of nuclear technology in such sectors as agriculture, health and energy planning. The plan was developed with the support of the International Atomic Energy Agency. The aim of the Country Program Framework, or CPF, launched two weeks ago by Eswatini Minister of Natural Resources and Energy Prince Lonkhokhela, is to leverage nuclear technology for social and economic development. Its key focus areas are energy security, food security and human health, aligning with the country’s National Development Plan and the United Nations Sustainable Development Cooperation Framework. Bongekile Matsenjwa, a chemical engineer and engineering manager for the Eswatini National Petroleum Company, believes the partnership between Eswatini and the International Atomic Energy Agency, or IAEA, can help the country make well-informed decisions about its energy future. "Access to clean, affordable and safe, reliable energy is an important ingredient for the sustainable development of the country,” he said. “I believe that this partnership can help Eswatini to make knowledgeable decisions on energy supply options with the help of energy planning so the country ... can independently chart our national energy future.” Sonia Paiva, a sustainable agriculture expert and advocate for nuclear technology, who was a panelist at the COP28 U.N. Climate Change Conference, believes Eswatini’s focus on nuclear technology is happening at the perfect moment, as the country has already established policies around the topic and is now moving toward implementation. “The whole world is looking to see how we can make our planet a better place to live in,” she said. In addition to its potential benefits in agriculture and energy, Dr. Mduduzi Mbuyisa, a medical doctor, believes this technology has immense potential to improve the health care system in Eswatini. “Nuclear medicine has a potential to ensure our diagnostic capabilities such that it helps us to take clearer pictures and help us in advanced imaging because we [are] using what we call PET or SPECT, which help to improve the care and overall health care system," he said. It will also ... help develop new skills and open up new career opportunities.” Eswatini's venture into nuclear technology is part of a larger trend of African countries seeking to harness the benefits of this technology. Against the backdrop of rising energy demands and climate change concerns, nuclear energy is increasingly seen as a potential solution.

In Nairobi, a new restaurant is generating business and buzz – not just because of the food, but because of the staff. Robots serving dishes is the main attraction of diners who flock to the Robot Cafe. Juma Majanga reports. Camera: Jimmy Makhulo.

islamabad — From sharing memes to sealing deals, millions of Pakistanis are struggling to communicate digitally as internet and data services have slowed down across much of the country.  Officials are blaming internet service providers for the slowdown, but media reports indicate the problem may stem from the deployment of a nationwide internet firewall aimed at controlling online content and traffic.  Crippling economy  The problem, which began several weeks ago, has worsened in recent days, frustrating freelancers like Moadood Ahmad, who is seeing a drop in income.  "If I don't show as available on Upwork or Fiverr, then new clients can't even approach me. Older clients are also disturbed," Ahmad told VOA. The Lahore-based digital marketing services provider says he has made virtually no money in the last two weeks.  According to DataReportal.com, 111 million Pakistanis in the nation of more than 240 million have access to internet. The country has nearly 189 million active cellular connections. State Bank of Pakistan put the country's IT exports at $3.2 billion in the financial year that ended in July 2024.  The spokesperson's office of the Pakistan Telecommunication Authority (PTA) told VOA it has only heard about issues with internet speed and web access through media. Speaking to VOA in late July, a PTA spokesperson blamed the slowdown at that time on a possible technical glitch.  During a hearing Thursday, the secretary for the Ministry of Information Technology and Telecommunication told a Senate committee that it was gathering data about the situation from mobile operators.  But the slump in services is not limited to mobile data users. In a statement to the media Thursday, an alliance of internet service providers (ISPs) said internet speeds had plummeted by 30% to 40%, "crippling [the] digital economy."  "Many are leaving the smaller ISPs because they can't sustain the poor service quality anymore. If this continues, we will see a mass exodus of businesses from Pakistan," said the Wireless and Internet Service Providers Association of Pakistan (WISPAP).  Speaking to VOA, WISPAP Chairman Shahzad Arshad said his alliance members were inundated with customer complaints.  "The authorities should at least tell us a timeframe that we can give to the customer," Arshad said.  Without mentioning the firewall, WISPAP's statement blamed increased "security and surveillance" for the decline in service quality.  "While the government remains steadfast in its commitment to enhancing security, the question remains: at what cost to the nation's digital economy?" the statement asked.  Controlling access  A source familiar with the nationwide firewall told VOA the system — acquired from China and installed through the Ministry of Defense — is based at a cable landing station in Pakistan, the place where the undersea internet data cable meets a country's internet system.     The firewall, also placed on the systems used by data providers, can give Pakistani authorities information about an individual user's online activities and where they are operating from, allowing for targeted monitoring, the source said.  Efforts to deploy the firewall on servers that major international companies use to store content close to end users are also underway. This can give authorities deeper access to an individual's data and control of their online activities.  Digital rights activists reject the government's claim that the firewall is only a cybersecurity tool. "To me, this is about political control. This desperation to do whatever they can," said Farieha Aziz, co-founder of the digital rights group Bolo Bhi, an Urdu name that means "speak up."  Government officials "are still not transparent and open about the capability [of the firewall], what kind of monitoring is happening," said Aziz. "What are they attempting?"  The country's latest efforts to control internet traffic and user activity come as the powerful military frequently complains of rising "digital terrorism," a term it uses for those who criticize or mock the armed forces on social media.  Digital rights activist Aziz also criticized telecom companies and internet service providers for not being transparent with users about the reasons for service disruptions.  Digital freedom is limited in Pakistan. Global rights watchdog Freedom House describes Pakistan as "not free," with a low score of only 26 out of 100 on its internet freedom index.  Pakistanis trying to bypass poor connectivity and speed by using VPNs are not faring much better either.  "Even those who use VPNs, and use our internet services, are unable to connect," WISPAP's Shahzad confirmed to VOA.  A PTA official told the Standing Committee on Cabinet Sec­retariat, earlier this month that the regulator was working on a plan to approve certain virtual private networks. All other VPNs would be blocked.  Senator Palwasha Khan, chairperson of the Senate's standing committee on IT, told VOA she expected internet speed to normalize in Pakistan within days. She said the committee was not informed if the problem was because of the firewall, but said she supported the censorship tool.  "I do agree that if some security measures are taken to control the chaos on social media, I don't have any problem with that," said Khan, a senator from the Pakistan People's Party, which is part of the ruling alliance.  Struggling to share voice notes and visual content with clients via a popular messaging app, freelancer Moadood Ahmad told VOA he regretted moving back to Pakistan from the United Arab Emirates.  "I am thinking I made a mistake," Ahmad said. "I should go back." 

NEW YORK — Google said Wednesday that an Iranian group linked to the country's Revolutionary Guard had tried to infiltrate the personal email accounts of roughly a dozen people linked to President Joe Biden and former President Donald Trump since May.  The tech company's threat intelligence arm said the group was still actively targeting people associated with Biden, Trump and Vice President Kamala Harris, who replaced Biden as the Democratic candidate last month when he dropped out of the presidential race. It said those targeted have included current and former government officials, as well as presidential campaign affiliates.  The new report from Google's Threat Analysis Group affirmed and expanded upon a Microsoft report released Friday that revealed a suspected Iranian cyber intrusion in this year's U.S. presidential election. It shed light on how foreign adversaries are increasing their efforts to disrupt the election that is now less than three months away.  Google's report said its threat researchers detected and disrupted a "small but steady cadence" of the Iranian attackers using email credential phishing, a type of cyberattack where the attacker poses as a trusted sender to try to get an email recipient to share login details. John Hultquist, chief analyst for the company's threat intelligence arm, said the company sends suspected targets of these attacks a Gmail pop-up that warns them that a government-backed attacker might be trying to steal their passwords.  The report said Google observed the group gaining access to one high-profile political consultant's personal Gmail account. Google reported the incident to the FBI in July. Microsoft's Friday report had shared similar information, noting that the email account of a former senior adviser to a presidential campaign had been compromised and weaponized to send a phishing email to a high-ranking campaign official.  The group is familiar to Google's threat intelligence arm and other researchers, and this isn't the first time it has tried to interfere in U.S. elections, Hultquist said. The report noted that the same Iranian group targeted both the Biden and Trump campaigns with phishing attacks during the 2020 cycle, as early as June of that year.  The group also has been prolific in other cyber espionage activity, particularly in the Middle East, the report said. In recent months, as the Israel-Hamas war has fueled tension in the region, that activity has included email phishing campaigns targeted at Israeli diplomats, academics, nongovernmental organizations and military affiliates.  Trump's campaign said Saturday that it had been hacked and that sensitive internal documents had been stolen and distributed. It said Iranian actors were to blame.  The same day, Politico revealed it had received leaked internal Trump campaign documents by email, though it wasn't clear whether the leaked documents were related to the suspected Iranian cyber activity. The Washington Post and The New York Times also received the documents.  While the Trump campaign hasn't provided specific evidence linking Iran to the hack, both Trump and his longtime friend and former adviser Roger Stone have said they were contacted by Microsoft related to suspected cyber intrusions. Stone's email was compromised by hackers targeting Trump's campaign, a person familiar with the matter said.  Google and Microsoft wouldn't identify the people targeted in the Iranian intrusion attempts or confirm that Stone was among them. Google did confirm that the Iranian group in its report, which it calls APT42, is the same as the one in Microsoft's research. Microsoft refers to the group as Mint Sandstorm.  Harris' campaign has declined to say whether it has identified any state-based intrusion attempts but has said it vigilantly monitors cyber threats and isn't aware of any security breaches of its systems.  The FBI on Monday confirmed that it's investigating the intrusion of the Trump campaign. Two people familiar with the matter said the FBI also is investigating attempts to gain access to the Biden-Harris campaign.  The reports of Iranian hacking come as U.S. intelligence officials have warned of persistent and mounting efforts from both Russia and Iran to influence the U.S. election through their online activity. Beyond these hacking incidents, groups linked to the countries have used fake news websites and social media accounts to churn out content that appears intended to sway voters' opinions.  While neither Microsoft nor Google specified Iran's intentions in the U.S. presidential race, U.S. officials have previously hinted that Iran particularly opposes Trump. U.S. officials also have expressed alarm about Tehran's efforts to seek retaliation for a 2020 strike on an Iranian general that was ordered by Trump.  Iran's mission to the United Nations, when asked about the claim of the Trump campaign, denied being involved.  "We do not accord any credence to such reports," the mission told The Associated Press. "The Iranian government neither possesses nor harbors any intent or motive to interfere in the United States presidential election."  The mission did not immediately respond to a request for comment Wednesday about Google's report.

WELLINGTON, New Zealand — The New Zealand government said Tuesday that it would introduce new legislation to make it easier for companies and researchers to develop and commercialize products using gene technologies such as gene editing.  Science, Innovation and Technology Minister Judith Collins said in a statement that rules and time-consuming processes have made research outside the lab almost impossible.  "These changes will bring New Zealand up to global best practice and ensure we can capitalize on the benefits," she said.  Current regulations mean that genetically modified organisms (GMOs) cannot be released out of containment without going through a complex and vigorous process and it is difficult to meet the set standard. Furthermore, gene editing is considered the same as genetic modification even when it doesn't involve the introduction of foreign DNA.  Under the new law, low-risk gene editing techniques that produce changes indistinguishable from conventional breeding will be exempted from regulation, local authorities will no longer be able to prevent the use of GMOs in their regions and there will be a new regulator of the industry.  "This is a major milestone in modernizing gene technology laws to enable us to improve health outcomes, adapt to climate change, deliver massive economic gains and improve the lives of New Zealanders," Collins said.  The government hopes to have the legislation passed and the regulator in operation by the end of 2025.

beijing — Engineers sent China's biggest-yet cargo drone on a test run over the weekend while a helicopter taxi took to the skies on a soon-to-open 100-km route to Shanghai, laying new milestones for the country's expanding low-altitude economy. Packing a payload capacity of 2 metric tons, the twin-engine aircraft took off on Sunday on an inaugural flight, state media said, citing developer Sichuan Tengden Sci-tech Innovation Co., for a trip of about 20 minutes in southwestern Sichuan province. China's civilian drone makers are testing larger payloads as the government pushes to build a low-altitude economy, with the country’s aviation regulator envisioning a $279-billion industry by 2030, for a four-fold expansion from 2023. The Tengden-built drone, with a wingspan of 16.1 meters and a height of 4.6 meters, is slightly larger than the world's most popular light aircraft, the four-seat Cessna 172. The trial run followed the maiden flight in June of a cargo drone developed by state-owned Aviation Industry Corp of China (AVIC), the leading aerospace enterprise. The AVIC's HH-100 has a payload capacity of 700 kilograms and a flight radius of 520 km. Next year, AVIC plans to test its biggest cargo drone, the TP2000, which can carry up to 2 tons of cargo and fly four times farther than the HH-100. China has already begun commercial deliveries by drone. In May, cargo drone firm Phoenix Wings, part of delivery giant SF Express, started delivering fresh fruit from the island province of Hainan to southern Guangdong, using Fengzhou-90 drones developed by SF, a unit of S.F. Holding 002352.SZ. Cargo drones promise shorter delivery times and lower transport costs, Chinese industry insiders say, while widening deliveries to sites lacking conventional aviation facilities, such as rooftop spaces in heavily built-up cities. They could also ferry people on taxi services.

WUHAN, China — Liu Yi is among China's 7 million ride-hailing drivers. A 36-year-old Wuhan resident, he started driving part-time this year when construction work slowed in the face of a nationwide glut of unsold apartments. Now he predicts another crisis as he stands next to his car watching neighbors order driverless taxis. "Everyone will go hungry," he said of Wuhan drivers competing against robotaxis from Apollo Go, a subsidiary of technology giant Baidu 9888.HK. Baidu and the Ministry of Industry and Information Technology declined comment. Ride-hailing and taxi drivers are among the first workers globally to face the threat of job loss from artificial intelligence as thousands of robotaxis hit Chinese streets, economists and industry experts said. Self-driving technology remains experimental but China has moved aggressively to green-light trials compared with the U.S which is quick to launch investigations and suspend approvals after accidents. At least 19 Chinese cities are running robotaxi and robobus tests, disclosure showed. Seven have approved tests without human-driver monitors by at least five industry leaders: Apollo Go, Pony.ai, WeRide, AutoX and SAIC Motor 600104.SS. Apollo Go has said it plans to deploy 1,000 in Wuhan by year-end and operate in 100 cities by 2030. Pony.ai, backed by Japan's Toyota Motor 7203.T, operates 300 robotaxis and plans 1,000 more by 2026. Its vice president has said robotaxis could take five years to become sustainably profitable, at which point they will expand "exponentially." WeRide is known for autonomous taxis, vans, buses and street sweepers. AutoX, backed by e-commerce leader Alibaba Group 9988.HK, operates in cities including Beijing and Shanghai. SAIC has been operating robotaxis since the end of 2021. "We've seen an acceleration in China. There's certainly now a rapid pace of permits being issued," said Boston Consulting Group managing director Augustin Wegscheider. "The U.S. has been a lot more gradual." Alphabet's GOOGL.O Waymo is the only U.S. firm operating uncrewed robotaxis that collect fares. It has over 1,000 cars in San Francisco, Los Angeles and Phoenix but could grow to "thousands," said a person with knowledge of its operations. Cruise, backed by General Motors GM.N, restarted testing in April after one of its vehicles hit a pedestrian last year. Cruise said it operates in three cities with safety its core mission. Waymo did not respond to a request for comment. "There's a clear contrast between U.S. and China" with robotaxi developers facing far more scrutiny and higher hurdles in the U.S., said former Waymo CEO John Krafcik. Robotaxis spark safety concerns in China, too, but fleets proliferate as authorities approve testing to support economic goals. Last year, President Xi Jinping called for "new productive forces," setting off regional competition. Beijing announced testing in limited areas in June and Guangzhou said this month it would open roads citywide to self-driving trials. Some Chinese firms have sought to test autonomous cars in the U.S. but the White House is set to ban vehicles with China-developed systems, said people briefed on the matter. Boston Consulting's Wegscheider compared China's push to develop autonomous vehicles to its support of electric vehicles. "Once they commit," he said, "they move pretty fast." 'Stupid radishes' China has 7 million registered ride-hailing drivers versus 4.4 million two years ago, official data showed. With ride-hailing providing last-resort jobs during economic slowdown, the side effects of robotaxis could prompt the government to tap the brakes, economists said. In July, discussion of job loss from robotaxis soared to the top of social media searches with hashtags including, "Are driverless cars stealing taxi drivers' livelihoods?" In Wuhan, Liu and other ride-hailing drivers call Apollo Go vehicles "stupid radishes" - a pun on the brand's name in local dialect - saying they cause traffic jams. Liu worries, too, about the impending introduction of Tesla's TSLA.O "Full Self-Driving" system - which still requires human drivers - and the automaker's robotaxi ambitions. "I'm afraid that after the radishes come," he said, "Tesla will come." Wuhan driver Wang Guoqiang, 63, sees a threat to workers who can least afford disruption. "Ride-hailing is work for the lowest class," he said, as he watched an Apollo Go vehicle park in front of his taxi. "If you kill off this industry, what is left for them to do?" Baidu declined to comment on the drivers' concerns and referred Reuters to comments in May by Chen Zhuo, Apollo Go's general manager. Chen said the firm would become "the world's first commercially profitable" autonomous-driving platform. Apollo Go loses almost $11,000 a car annually in Wuhan, Haitong International Securities estimated. A lower-cost model could enable per-vehicle annual profit of nearly $16,000, the securities firm said. By contrast, a ride-hailing car earns about $15,000 total for the driver and platform. 'Already at the forefront' Automating jobs could benefit China in the long run given a shrinking population, economists said. "In the short run, there must be a balance in speed between the creation of new jobs and the destruction of old jobs," said Tang Yao, associate professor of applied economics at Peking University. "We do not necessarily need to push at the fastest speed, as we are already at the forefront." Eastern Pioneer Driving School 603377.SS has more than halved its instructor number since 2019 to about 900. Instead, it has teachers at a Beijing control center remotely monitoring students in 610 cars equipped with computer instruction tools. Computers score students on every wheel turn and brake tap, and virtual reality simulators coach them on navigating winding roads. Massive screens provide real-time analysis of driver tasks, such as one student's 82% parallel-parking pass rate. Zhang Yang, the school's intelligent-training director, said the machines have done well. "The efficiency, pass rate and safety awareness have greatly improved."

Flying cars, long the dream of futurists, may finally be here. From California, Matt Dibble has our story about the rise of electric aircraft.

United Nations — U.N. member states on Thursday approved a treaty targeting cybercrime, the body's first such text, despite fierce opposition from human rights activists who have warned of potential surveillance dangers. After three years of negotiations and a final two-week session in New York, members approved the United Nations Convention Against Cybercrime by consensus, and it will now be submitted to the General Assembly for formal adoption. "I consider the documents ... adopted. Thank you very much, bravo to all!" Algerian diplomat Faouzia Boumaiza Mebarki, chairwoman of the treaty drafting committee, said to applause. The committee was set up, despite U.S. and European opposition, following an initial move in 2017 by Russia. The new treaty would enter into force once it has been ratified by 40 member nations and aims to "prevent and combat cybercrime more efficiently and effectively," notably regarding child sexual abuse imagery and money laundering. Hailing a "landmark convention," South Africa's delegate said, "the provisions of technical assistance and capacity building offer much needed support to countries with less developed cyber infrastructures." But the treaty's detractors -- an unusual alliance of human rights activists and big tech companies -- denounce it as being far too broad in scope, claiming it could amount to a global "surveillance" treaty and be used for repression. In particular, the text provides that a state may, in order to investigate any crime punishable by a minimum of four years' imprisonment under its domestic law, ask the authorities of another nation for any electronic evidence linked to the crime, and also request data from internet service providers. Warning of an "unprecedented multilateral tool for surveillance," Deborah Brown of Human Rights Watch told AFP the treaty "will be a disaster for human rights and is a dark moment for the UN." "This treaty is effectively a legal instrument of repression," she said. "It can be used to crack down on journalists, activists, LGBT people, free thinkers, and others across borders." Human rights clause Nick Ashton-Hart heads the Cybersecurity Tech Accord delegation to the treaty talks, representing more than 100 technology companies, including Microsoft and Meta. "Regretably," he said Thursday, the committee "adopted a convention without addressing many of the major flaws identified by civil society, the private sector, or even the U.N.'s own human rights body." "Wherever it is implemented the Convention will be harmful to the digital environment generally and human rights in particular," he told AFP, calling for nations not to sign or implement it. Some nations however are complaining the treaty actually includes too many human rights safeguards. A few days ago, Russia, which has historically supported the drafting process, complained the treaty had become "oversaturated with human rights safeguards," while accusing countries of pursuing "narrow self-serving goals under the banner of democratic values." During Thursday's session, Iran attempted to have several clauses with "inherent flaws" deleted. One paragraph in question stipulated that "nothing in this Convention shall be interpreted as permitting suppression of human rights or fundamental freedoms," such as "freedoms of expression, conscience, opinion, religion or belief." The deletion request was rejected with 102 votes against, 23 in favor (including Russia, India, Sudan, Venezuela, Syria, North Korea and Libya) and 26 abstentions. Neither Iran nor any other country, however, chose to prevent approval by consensus. 

NEW YORK — Iran is ramping up online activity that appears intended to influence the upcoming U.S. election, in one case targeting a presidential campaign with an email phishing attack, Microsoft said Friday. Iranian actors also have spent recent months creating fake news sites and impersonating activists, laying the groundwork to stoke division and potentially sway American voters this fall, especially in swing states, the technology giant found. The findings in Microsoft’s newest threat intelligence report show how Iran, which has been active in recent U.S. campaign cycles, is evolving its tactics for another election that’s likely to have global implications. The report goes a step beyond anything U.S. intelligence officials have disclosed, giving specific examples of Iranian groups and the actions they have taken so far. Iran's United Nations mission denied it had plans to interfere or launch cyberattacks in the U.S. presidential election. The report doesn’t specify Iran’s intentions besides sowing chaos in the United States, though U.S. officials have previously hinted that Iran particularly opposes former President Donald Trump. U.S. officials also have expressed alarm about Tehran's efforts to seek retaliation for a 2020 strike on an Iranian general that was ordered by Trump. This week, the Justice Department unsealed criminal charges against a Pakistani man with ties to Iran who’s alleged to have hatched assassination plots targeting multiple officials, potentially including Trump. The report also reveals how Russia and China are exploiting U.S. political polarization to advance their own divisive messaging in a consequential election year. Microsoft’s report identified four examples of recent Iranian activity that the company expects to increase as November’s election draws closer. First, a group linked to Iran’s Revolutionary Guard in June targeted a high-ranking U.S. presidential campaign official with a phishing email, a form of cyberattack often used to gather sensitive information, according to the report, which didn't identify which campaign was targeted. The group concealed the email’s origins by sending it from the hacked email account of a former senior adviser, Microsoft said. Days later, the Iranian group tried to log into an account that belonged to a former presidential candidate but wasn’t successful, Microsoft’s report said. The company notified those who were targeted. In a separate example, an Iranian group has been creating websites that pose as U.S.-based news sites targeted to voters on opposite sides of the political spectrum, the report said. One fake news site that lends itself to a left-leaning audience insults Trump by calling him "raving mad" and suggests he uses drugs, the report said. Another site meant to appeal to Republican readers centers on LGBTQ issues and gender-affirming surgery. A third example Microsoft cited found that Iranian groups are impersonating U.S. activists, potentially laying the groundwork for influence operations closer to the election. Finally, another Iranian group in May compromised an account owned by a government employee in a swing state, the report said. It was unclear whether that cyberattack was related to election interference efforts. Iran's U.N. mission sent The Associated Press an emailed statement: "Iran has been the victim of numerous offensive cyber operations targeting its infrastructure, public service centers, and industries. Iran’s cyber capabilities are defensive and proportionate to the threats it faces. Iran has neither the intention nor plans to launch cyber attacks. The U.S. presidential election is an internal matter in which Iran does not interfere." The Microsoft report said that as Iran escalates its cyber influence, Russia-linked actors also have pivoted their influence campaigns to focus on the U.S. election, while actors linked to the Chinese Communist Party have taken advantage of pro-Palestinian university protests and other current events in the U.S. to try to raise U.S. political tensions. Microsoft said it has continued to monitor how foreign foes are using generative AI technology. The increasingly cheap and easy-to-access tools can generate lifelike fake images, photos and videos in seconds, prompting concern among some experts that they will be weaponized to mislead voters this election cycle. While many countries have experimented with AI in their influence operations, the company said, those efforts haven’t had much impact so far. The report said as a result, some actors have "pivoted back to techniques that have proven effective in the past — simple digital manipulations, mischaracterization of content, and use of trusted labels or logos atop false information." Microsoft’s report aligns with recent warnings from U.S. intelligence officials, who say America’s adversaries appear determined to seed the internet with false and incendiary claims ahead of November’s vote. Top intelligence officials said last month that Russia continues to pose the greatest threat when it comes to election disinformation, while there are indications that Iran is expanding its efforts and China is proceeding cautiously when it comes to 2024. Iran's efforts seem aimed at undermining candidates seen as being more likely to increase tension with Tehran, the officials said. That’s a description that fits Trump, whose administration ended a nuclear deal with Iran, reimposed sanctions and ordered the killing of the top Iranian general. The influence efforts also coincide with a time of high tensions between Iran and Israel, whose military the U.S. strongly supports. Director of National Intelligence Avril Haines said last month that the Iranian government has covertly supported American protests over Israel’s war against Hamas in Gaza. Groups linked to Iran have posed as online activists, encouraged protests and provided financial support to some protest groups, Haines said. America’s foes, Iran among them, have a long history of seeking to influence U.S. elections. In 2020, groups linked to Iran sent emails to Democratic voters in an apparent effort to intimidate them into voting for Trump, intelligence officials said.

CARACAS, Venezuela — President Nicolás Maduro said he has ordered a 10-day block on access to X in Venezuela, accusing the owner Elon Musk of using the social network to promote hatred after the country's disputed presidential election. Associated Press journalists in Caracas found that by Thursday night posts had stopped loading on X on two private telephone services and state-owned Movilnet. "Elon Musk is the owner of X and has violated all the rules of the social network itself," said Maduro in a speech following a march by pro-government groups. Maduro alleged Musk "has incited hatred." Maduro also accused the social network of being used by his opponents to create political unrest. Venezuela's president said he had signed a resolution "with the proposal made by CONATEL, the National Telecommunications Commission, which has decided to remove the social network X, formerly known as Twitter, from circulation in Venezuela for 10 days so that they can present their documents." Maduro did not provide more details about the process taken against X. X's press office did not immediately respond to an email from AP requesting comment. "X out for 10 days! Elon Musk out!" Maduro said. The president's announcement comes after Maduro and Musk exchanged accusations over Venezuela's disputed July 28 presidential election. Electoral authorities declared Maduro the winner but have yet to produce voting tallies. Meanwhile, the opposition claims to have collected records from more than 80% of the 30,000 electronic voting machines nationwide showing the winner was their candidate, Edmundo González. Musk used the social network to accuse the self-proclaimed socialist leader of a "great electoral fraud." "Shame on the dictator Maduro," Musk said on Monday in a post. Since the election, Maduro has expressed the need to "regulate" social networks in Venezuela. Maduro also denounced that the social platform was used by his adversaries to threaten the families of his followers and political allies, military personnel, police officers and to generate a state of anxiety in Venezuela.