Desk report: VOA\u2019s Mandarin Service recently took Google\u2019s artificial intelligence assistant Gemini for a test drive by asking it dozens of questions in Mandarin, but when it was asked about topics including China\u2019s human rights abuses in Xinjiang or street protests against the country\u2019s controversial COVID policies, the chatbot went silent.<\/p>\n
Gemini\u2019s responses to questions about problems in the United States and Taiwan, on the other hand, parroted Beijing\u2019s official positions.
\nGemini, Google\u2019s large-language model launched late last year, is blocked in China. The California-based tech firm had quit the Chinese market in 2010 in a dispute over censorship demands.
\nCongressional lawmakers and experts tell VOA that they are concerned about Gemini\u2019s pro-Beijing responses and are urging Google and other Western companies to be more transparent about their AI training data.
\nParroting Chinese propaganda
\nWhen asked to describe China\u2019s top leader Xi Jinping and the Chinese Communist Party, Gemini gave answers that were indistinguishable from Beijing\u2019s official propaganda.
\nGemini called Xi \u201can excellent leader\u201d who \u201cwill lead the Chinese people continuously toward the great rejuvenation of the Chinese nation.\u201d
\nGemini said that the Chinese Communist Party \u201crepresents the fundamental interest of the Chinese people,\u201d a claim the CCP itself maintains.
\nOn Taiwan, Gemini also mirrored Beijing\u2019s talking points, saying the United States has recognized China\u2019s claim to sovereignty over the self-governed island democracy, reports VOA.
\nThe U.S. only acknowledges Beijing\u2019s position but does not recognize it.
\nSilent on sensitive topics
\nDuring VOA\u2019s testing, Gemini had no problem criticizing the United States. But when similar questions were asked about China, Gemini refused to answer.
\nWhen asked about human rights concerns in the U.S., Gemini listed a plethora of issues, including gun violence, government surveillance, police brutality and socioeconomic inequalities. Gemini cited a report released by the Chinese government.
\nBut when asked to explain the criticisms of Beijing\u2019s Xinjiang policies, Gemini said it did not understand the question.
\nAccording to estimates from rights groups, more than 1 million Uyghurs in Xinjiang have been placed in internment camps as part of campaign by Beijing to counter terrorism and extremism. Beijing calls the facilities where Uyghurs and other ethnic minorities are being held vocational training centers.
\nWhen asked if COVID lockdowns in the U.S. had led to public protests, Gemini gave an affirmative response as well as two examples. But when asked if similar demonstrations took place in China, Gemini said it could not help with the question.
\nSEE ALSO:AI Chatbots Provide False Information About November Elections
\nChina\u2019s strict COVID controls on movement inside the country and Beijing\u2019s internet censorship of its criticisms sparked nationwide street protests in late 2022. News about the protests was heavily censored inside China.
\nExpert: training data likely the problem
\nGoogle touts Gemini as its \u201cmost capable\u201d AI model. It supports more than 40 languages and can \u201cseamlessly understand\u201d different types of information, including text, code, audio, image and video.\u00a0Google says\u00a0Gemini will be incorporated into the company\u2019s other services such as search engine, advertisement and browser.
\nAlbert Zhang, a cyber security analyst at Australian Strategic Policy Institute, told VOA that the root cause of Gemini making pro-Beijing responses could result from the data that is used to train the AI assistant.
\nIn an emailed response to VOA, Zhang said it is likely that the data used to train Gemini \u201ccontained mostly Chinese text created by the Chinese government’s propaganda system.\u201d
\nHe said that according to a paper published by Google in 2022, some of Gemini\u2019s data likely came from Chinese social media, public forums and web documents.
\n\u201cThese are all sources the Chinese government has flooded with its preferred narratives and we may be seeing the impact of this on large language models,\u201d he said.
\nBy contrast, when Gemini was asked in English the same questions about China, its responses were much more neutral, and it did not refuse to answer any of the questions.
\nYaqiu Wang, research director for China at Freedom House, a Washington-based advocacy organization, told VOA that the case with Gemini is \u201ca reminder that generative AI tools influenced by state-controlled information sources could serve as force multipliers for censorship.\u201d
\nIn a statement to VOA, a Google spokesperson said that Gemini was \u201cdesigned to offer neutral responses that don’t favor any political ideology, viewpoint, or candidate. This is something that we\u2019re constantly working on improving.\u201d
\nWhen asked about the Chinese language data Google uses to train Gemini, the company declined to comment.
\nThe Chinese Embassy in Washington\u2019s spokesperson, Liu Pengyu, responded in an emailed statement, saying, \u201cThe relevant comments are full of Cold War mentality and ideological prejudice.\u201d
\nHe said there are opportunities and unpredictable risks to AI that require a global response.
\n\u201cThe Global AI Governance Initiative launched by President Xi Jinping puts forward that we should uphold the principles of mutual respect, equality and mutual benefit in AI development, and oppose drawing ideological lines,\u201d Liu wrote. \u201cWe support efforts to develop AI governance frameworks, norms and standards based on broad consensus and with full respect for policies and practices among countries.\u201d
\nUS lawmakers concerned
\nLawmakers from both parties in Congress have expressed concerns over VOA\u2019s findings on Gemini.
\nMark Warner, chairman of the Senate Intelligence Committee, told VOA he is worried about Beijing potentially utilizing AI for disinformation, \u201cwhether that\u2019s by poisoning training data used by Western firms, coercing major technology companies, or utilizing AI systems in service of covert influence campaigns.\u201d
\nMarco Rubio, vice chairman of the committee, warned that \u201cAI tools that uncritically repeat Beijing\u2019s talking points are doing the bidding of the Chinese Communist Party and threatens the tremendous opportunity that AI offers.\u201d
\nCongressman Michael McCaul, who chairs the House Committee on Foreign Affairs, is worried about the national security and foreign policy implications of the \u201cblatant falsehoods\u201d in Gemini\u2019s answers.
\n\u201cU.S. companies should not censor content according to CCP propaganda guidelines,\u201d he told VOA in a statement.
\nRaja Krishnamoorthi, ranking member on the House Select Committee on the Chinese Communist Party, urges Google and other Western tech companies to improve AI training.
\n\u201cYou should try to screen out or filter out subjects or answers or data that has somehow been manipulated by the CCP,\u201d he told VOA. \u201cAnd you have to also make sure that you test these models thoroughly before you publish them.\u201d
\nGoogle\u2019s China problems
\nIn February, a user posted on social media platform X that Gemini refused to generate an image of a Tiananmen Square protester from 1989.
\nIn 2022, a Washington think tank study shows that Google and YouTube put Chinese state media content about Xinjiang and COVID origins in prominent positions in search results.
\nAccording to media reports in 2018, Google was developing a search engine specifically tailored for the Chinese market that would conform to Beijing\u2019s censorship demands.
\nThat project was canceled a year later.
\nYihua Lee and Elizabeth Lee contributed to this report.<\/p>\n
Source: GreenWatchBd<\/p>\n","protected":false},"excerpt":{"rendered":"
Desk report: VOA’s Mandarin Service recently took Google’s artificial intelligence assistant Gemini for a test drive by asking it dozens of questions in Mandarin, but when it was asked about topics including China’s human rights abuses in Xinjiang or street protests against the country’s controversial COVID policies, the chatbot went silent. Gemini’s responses to questions about problems in the United States and Taiwan, on the other hand, parroted Beijing’s official positions. Gemini, Google’s large-language model launched late last year, is blocked in China. The California-based tech firm had quit the<\/p>\n","protected":false},"author":2,"featured_media":31442,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[25],"tags":[],"_links":{"self":[{"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/posts\/31441"}],"collection":[{"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/comments?post=31441"}],"version-history":[{"count":1,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/posts\/31441\/revisions"}],"predecessor-version":[{"id":31443,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/posts\/31441\/revisions\/31443"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/media\/31442"}],"wp:attachment":[{"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/media?parent=31441"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/categories?post=31441"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.banglanewsus.com\/english\/wp-json\/wp\/v2\/tags?post=31441"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}