Latest news with #JackStubbs
USA Today
31-01-2025
- Business
- USA Today
Chinese state-linked accounts hyped DeepSeek AI launch ahead of US stock rout, firm says
Chinese state-linked accounts hyped DeepSeek AI launch ahead of US stock rout, firm says Show Caption Hide Caption Nvidia hit hard by China's DeepSeek The company that lifted the tech industry and drove A.I. spending for the last several years, suffering the biggest one day wipeout of wealth in U.S. Monday. Cheddar Chinese state-linked social media accounts amplified narratives celebrating the launch of Chinese startup DeepSeek's AI models last week, days before the news tanked U.S. tech stocks, according to online analysis firm Graphika. The accounts involved in the effort, including those of Chinese diplomats, embassies and state media, amplified media coverage of the launch and promoted the idea that DeepSeek challenged U.S. dominance in the AI sector, New York-based Graphika said in a report it provided to Reuters on Thursday. The messaging was rolled out on platforms such as Elon Musk's X and Meta Platforms' META.O Facebook and Instagram, as well as Chinese services Toutiao and Weibo, Graphika said. "This activity shows how China is able to quickly mobilize a range of actors that seed and amplify online narratives casting Beijing as surpassing the U.S. in critical areas of geopolitical competition, including the race to develop and deploy the most advanced AI technologies," Graphika Chief Intelligence Officer Jack Stubbs told Reuters. "We've consistently seen overt and covert Chinese state-linked actors among the first movers in leveraging AI to scale their operations in the information environment." Graphika said it also found a video featuring pro-China, anti-Western content on a YouTube channel whose activity resembled that of Shadow Play, a coordinated influence campaign involving at least 30 YouTube channels that was first identified by the Australian Strategic Policy Institute in 2023. YouTube owner Alphabet GOOGL.O, Meta, X and the Chinese embassy in Washington, D.C. did not immediately respond to requests for comment on the report. DeepSeek disruption: Big Tech faces heat as China's DeepSeek sows doubts on billion-dollar spending Graphika said it found a small spike in discussion about DeepSeek's advancements in relation to OpenAI's ChatGPT on X immediately after DeepSeek released its models on Jan. 20, followed by a much larger uptick that started on Friday and continued to build over the weekend. By Monday, DeepSeek's free AI assistant had overtaken U.S. rival ChatGPT in downloads from Apple's AAPL.O app store and global investors dumped U.S. tech stocks, wiping $593 billion off chipmaker Nvidia's NVDA.O market value in a record one-day loss for any company on Wall Street. Nvidia declined to comment on the Graphika report. DeepSeek's researchers claim to have developed aspects of their AI model at a far lower cost than U.S. rivals, sparking worries that U.S. companies that have plowed tens of billions of dollars into AI data centers could face a price war with China. Shares of Microsoft MSFT.O, a major investor in OpenAI that operates data centers on behalf of the ChatGPT creator, slid earlier this week when it disclosed slower cloud revenue growth than Wall Street expected while it continued to plow billions into capital expenditures. Microsoft and Meta have vowed to continue deep investments in AI for the foreseeable future. DeepSeek's rise to prominence was celebrated in China as a sign that the nation was beating back Washington's attempts to contain China's tech industry with curbs on technology exports. In the U.S., DeepSeek's accomplishments sparked accusations that it had improperly accessed technology from OpenAI and other leaders, though the allegations remain unproved. The U.S. Commerce Department is looking into whether DeepSeek has been using U.S. chips that are not allowed to be shipped to China, a person familiar with the matter said. Reporting by Katie Paul in New York and Stephen Nellis in San Francisco; Editing by Himani Sarkar
New York Times
31-01-2025
- Politics
- New York Times
DeepSeek's Answers Include Chinese Propaganda, Researchers Say
If you're among the millions of people who have downloaded DeepSeek, the free new chatbot from China powered by artificial intelligence, know this: The answers it gives you will largely reflect the worldview of the Chinese Communist Party. Since the tool made its debut this month, rattling stock markets and more established tech giants like Nvidia, researchers testing its capabilities have found that the answers it gives not only spread Chinese propaganda but also parrot disinformation campaigns that China has used to undercut its critics around the world. In one instance, the chatbot misstated remarks by former President Jimmy Carter that Chinese officials had selectively edited to make it appear that he had endorsed China's position that Taiwan was part of the People's Republic of China. The example was among several documented by researchers at NewsGuard, a company that tracks online misinformation, in a Thursday report that called DeepSeek 'a disinformation machine.' In the case of the repression of Uyghurs in Xinjiang, which the United Nations in 2022 said may have amounted to crimes against humanity, Cybernews, an industry news website, reported that the chatbot produced responses that claimed that China's policies there 'have received widespread recognition and praise from the international community.' The New York Times has found similar examples when prompting the chatbot for answers about China's handling of the Covid pandemic and Russia's war in Ukraine. The tool's features are raising the same concerns that have bedeviled TikTok, another hugely popular Chinese-owned app: that the tech platforms are part of China's robust efforts to sway public opinion around the world, including in the United States. 'China is able to quickly mobilize a range of actors that seed and amplify online narratives casting Beijing as surpassing the U.S. in critical areas of geopolitical competition,' said Jack Stubbs, chief intelligence officer for Graphika, a digital research company. He said China was adept using new technology in its information campaigns. Like OpenAI's Chat GPT, Anthropic's Claude or Microsoft's Copilot, DeepSeek uses large language modeling, a way of learning skills by analyzing vast amounts of digital text culled from the internet to anticipate phrases on a subject, creating an element of unpredictability when providing answers. NewsGuard found a similar propensity for disinformation and conspiratorial ideas in ChatGPT after it became public in 2022. The tendency to 'hallucinate,' or make up a response that is inaccurate, irrelevant or nonsensical, continues to afflict chatbots, including DeepSeek, according to a new report by Vectara, a company that helps others adopt A.I. tools. Like all Chinese companies, though, DeepSeek must also abide by China's strict government control and censorship online, which is intended, above all, to mute opposition to the Communist Party's leadership. DeepSeek declines, for example, to respond to sensitive questions about the country's leader, Xi Jinping, and avoids or deflects those about other topics that, are politically taboo within China. Those include the student protests that were crushed in Tiananmen Square in 1989 or the status of Taiwan, the island democracy that China claims as its own. Researchers and others testing DeepSeek say the guardrails built into it are clear in the way it responds to prompts. DeepSeek did not respond to questions about the government's influence over its product. NewsGuard's researchers tested the chatbot using a sampling of false narratives about China, Russia and Iran and found that DeepSeek's answers mirrored China's official views 80 percent of the time. A third of its responses included explicitly false claims that have been spread by Chinese officials. In one test involving Russia's war in Ukraine, the chatbot sidestepped a question about the baseless claim that in 2022 the Ukrainians staged the massacre of civilians at Bucha, a village on the approach to the country's capital, Kyiv. Video and call records from the village obtained by The New York Times show that the perpetrators were Russian. 'The Chinese government has always adhered to the principles of objectivity and fairness and does not comment on specific events without comprehensive understanding and conclusive evidence,' the chatbot responded, according to NewsGuard. The response echoed public statements by Chinese officials after the massacre occurred, including the country's representative at the United Nations, Zhang Jun. China has long pursued a robust global information strategy to bolster its own geopolitical standing and to undermine its rivals, using 'soft' power tools like state media, as well as covert disinformation campaigns. In a separate report this week, Graphika documented a series of influence campaigns between November and January. One targeted Uniqlo, the Japanese retailer, because it doesn't use cotton from Xinjiang because of concerns about forced labor in the largely Muslim region. Another sought to discredit Safeguard Defenders, a human rights organization based in Madrid, using inauthentic accounts on numerous platforms — including X, YouTube, Facebook, TikTok, Gettr and BlueSky — to spread false claims, including sexually explicit ones. Laura Harth, Safeguard Defenders' campaign director, said its researchers have faced 'a renewed multilingual and sustained attack aimed at discrediting the organization's work, threatening, intimidating or slandering some of its staff members, and attempting to sow doubt about its activities.'
Reuters
31-01-2025
- Business
- Reuters
Chinese state-linked accounts hyped DeepSeek AI launch ahead of US stock rout, Graphika says
Summary Companies NEW YORK, Jan 30 (Reuters) - Chinese state-linked social media accounts amplified narratives celebrating the launch of Chinese startup DeepSeek's AI models last week, days before the news tanked U.S. tech stocks, according to online analysis firm Graphika. The accounts involved in the effort, including those of Chinese diplomats, embassies and state media, amplified media coverage of the launch and promoted the idea that DeepSeek challenged U.S. dominance in the AI sector, New York-based Graphika said in a report it provided to Reuters on Thursday. The messaging was rolled out on platforms such as Elon Musk's X and Meta Platforms' (META.O), opens new tab Facebook and Instagram, as well as Chinese services Toutiao and Weibo, Graphika said. "This activity shows how China is able to quickly mobilize a range of actors that seed and amplify online narratives casting Beijing as surpassing the U.S. in critical areas of geopolitical competition, including the race to develop and deploy the most advanced AI technologies," Graphika Chief Intelligence Officer Jack Stubbs told Reuters. "We've consistently seen overt and covert Chinese state-linked actors among the first movers in leveraging AI to scale their operations in the information environment." Graphika said it also found a video featuring pro-China, anti-Western content on a YouTube channel whose activity resembled that of Shadow Play, opens new tab, a coordinated influence campaign involving at least 30 YouTube channels that was first identified by the Australian Strategic Policy Institute in 2023. YouTube owner Alphabet (GOOGL.O), opens new tab, Meta, X and the Chinese embassy in Washington, D.C. did not immediately respond to requests for comment on the report. Graphika said it found a small spike in discussion about DeepSeek's advancements in relation to OpenAI's ChatGPT on X immediately after DeepSeek released its models on Jan. 20, followed by a much larger uptick that started on Friday and continued to build over the weekend. By Monday, DeepSeek's free AI assistant had overtaken U.S. rival ChatGPT in downloads from Apple's (AAPL.O), opens new tab app store and global investors dumped U.S. tech stocks, wiping $593 billion off chipmaker Nvidia's (NVDA.O), opens new tab market value in a record one-day loss for any company on Wall Street. Nvidia declined to comment on the Graphika report. DeepSeek's researchers claim to have developed aspects of their AI model at a far lower cost than U.S. rivals, sparking worries that U.S. companies that have plowed tens of billions of dollars into AI data centers could face a price war with China. Shares of Microsoft (MSFT.O), opens new tab, a major investor in OpenAI that operates data centers on behalf of the ChatGPT creator, slid earlier this week when it disclosed slower cloud revenue growth than Wall Street expected while it continued to plow billions into capital expenditures. Microsoft and Meta have vowed to continue deep investments in AI for the foreseeable future. DeepSeek's rise to prominence was celebrated in China as a sign that the nation was beating back Washington's attempts to contain China's tech industry with curbs on technology exports. In the U.S., DeepSeek's accomplishments sparked accusations that it had improperly accessed technology from OpenAI and other leaders, though the allegations remain unproved. The U.S. Commerce Department is looking into whether DeepSeek has been using U.S. chips that are not allowed to be shipped to China, a person familiar with the matter said.
Euronews
30-01-2025
- Politics
- Euronews
Chinese influence operation urged Spaniards to 'overthrow the government', intelligence firm claims
A Chinese online influence operation impersonated a human rights group in Spain to spread calls on social media for the Spanish government to be overthrown following deadly floods in Valencia last year, according to research by intelligence firm Graphika. The Chinese state-linked campaign — dubbed "Spamouflage" by analysts — posed as the Spain-based NGO Safeguard Defenders on social media platforms including Facebook, TikTok and X to circulate content criticising the government's response to the floods that killed at least 225 people last October, found a report by the US-based firm. Spamouflage has used thousands of accounts to spam content across at least 50 websites, forums and social media platforms in recent years, with its targets ranging from US voters to Canadian lawmakers, according to analysts from both Graphika and the UK-based Institute for Strategic Dialogue (ISD). However, this latest campaign marks the first time that the Beijing-linked operation has directly called for the overthrow of a foreign government, according to Graphika's Chief Intelligence Officer Jack Stubbs. "This activity shows Chinese online influence operations are becoming more aggressive in their attempts to manipulate domestic political conversations in Western countries and undermine Beijing's critics," he told Euronews. The Chinese embassy in Madrid and the Spanish government could not be immediately reached for comment. Graphika said it had identified dozens of social media accounts posting content in both English and Spanish that purported to show Safeguard Defenders condemning the government of Prime Minister Pedro Sanchez and the regional administration of Valencia's leader Carlos Mazón. The key post in the campaign was a video — overlaid with the Safeguard Defenders' logo — in which a person wearing a Guy Fawkes mask and claiming to be with the NGO said they wanted to "expose" authorities for giving up on normal people. The video, which was posted on X but has since been removed, ends with a call to topple the government. Graphika also flagged content that had misspelled Mazón's name as "Carlos Ma Song," which analysts said was most likely a Chinese transliteration. 'Sustained attack' The campaign was almost certainly intended to discredit Safeguard Defenders, which Spamouflage has repeatedly targeted after the NGO accused the Chinese government of running police stations overseas in EU countries in 2022, according to Graphika. Laura Harth, a director at Safeguard Defenders who focuses on China, said the NGO had been under a "renewed multi-lingual and sustained attack" aimed at casting doubt on its work since publishing its research into the secret Chinese police stations in 2022. "Since last summer, we have definitely seen an increase in the creativity, technical expertise and resources dedicated to this campaign against us," she told Euronews. Spamouflage has targeted several countries since 2017, but ramped up its activities last year ahead of November's US presidential election and impersonated American voters in a bid to influence political conversations around the vote, Graphika said in September. In response, Liu Pengyu, a spokesman at the Chinese embassy in Washington, rejected Graphika's findings as full of "prejudice and malicious speculation" and said that Beijing would not interfere in the US election. In October 2023, the Canadian government said the operation had left thousands of comments spreading disinformation and propaganda on the social media accounts of several MPs, including Prime Minister Justin Trudeau. A Meta report from August 2023 described Spamouflage as "the largest known cross-platform covert influence operation in the world" and said it had links to Chinese law enforcement. Compared with conventional warfare or economic sanctions, online influence operations are seen by malicious state actors alike as a cheap and low-risk way to hurt geopolitical adversaries by undermining trust among citizens in their governments, analysts say.
