Following speculation last month that Chinese leader Xi Jinping was facing internal challenges to his power, and that he had experienced a marked decline in prestige in China’s official media, we looked at his performance in the front-page headlines of the official People’s Daily — a fair if imperfect reflection of the prevailing internal consensus.
How do things stand now at the close of July?
By our latest front-page count, Xi Jinping’s performance remains consistent. The dip since 2023 is consistent with historical patterns, where steep jumps in frequency for the top leader can be seen in the wake of Chinese Communist Party (CCP) congresses. The first half of 2023 understandably brought a sharp incline in the use of new buzzwords and phrases from the political report to the 20th National Congress of the CCP, held from October 16-22, 2022.
Terms like “Chinese-style modernization” (中国式现代化) and “new form of human civilization” (人类文明新形态) were naturally pushed with renewed vigor in the state media from January onward, defining the ideological status quo of the post-20th period. And other Xi Jinping staples, like the “Two Unshakeables” (两个毫不动摇), followed suit.
In the short term, it is not unusual to see apparent gaps with previous performance emerge. These can close as important events, such as plenary sessions or important foreign policy exchanges, drive a burst in front-page coverage.
The most crucial point to bear in mind is the extreme and persisting gap between Xi Jinping and all other members of the party’s Politburo Standing Committee (PSC). Xi maintains a commanding lead in the party’s internal messaging. We can also note that no other members of the PSC have made clear advances in terms of front-page performance.
For those asking whether or not a power struggle is underway in China, you are asking the wrong question. Of course there is struggle. This is the nature of politics under the CCP. The only real question is: What kind? As we near the next National Congress in 2027 — for which the midpoint passed in April this year — we can naturally expect various forms of jockeying and positioning. Our point is that these moves and shifts are not yet visible in the state media headlines.
We will certainly keep you posted.
What do we mean when we talk about AI risks, AI safety and AI security? These terms remain fuzzy around the world, even if we all have a clear sense of their huge importance given the growing impact of artificial intelligence. But another thing we should be clear about is where our emerging understanding of safe AI diverges in fundamental ways with that of key players like China that are busy shaping the future of the technology. With China’s announcement of its Global AI Governance Action Plan in Shanghai on July 26, promising to shepherd the safe development of this new technology worldwide, clarity is more important than ever.
Over the past year, the belief seems to be emerging among Western AI policy researchers and AI developers that China’s views on safety are converging, that we are all in this together. Matt Sheehan at the Carnegie Endowment for International Peace argues that the ambiguous definition of “AI safety” (人工智能安全) within policy documents has started to incorporate international concerns, especially around the existential risks posed by AI. Jack Clark, co-founder of Anthropic, wrote in a recent post to Substack that “China cares about the same safety risks as us.”
But this belief deserves caution — and context.
A perfect example of this imagined convergence came in December last year, when the major Chinese tech companies working on AI models signed a pledge on safeguarding AI safety (人工智能安全承诺). The companies included DeepSeek, Alibaba, Baidu, Zhipu AI, Huawei, the internet security firm Qihoo360, and a subsidiary of ByteDance, the developer of TikTok.
Released in both English and Chinese, the pledge proposed a testing system for AI model development, which would guarantee “AI safety and security,” as the English version termed it. In a well-researched piece, Scott Singer of the Carnegie Endowment for International Peace points to this pledge as evidence of “strong similarity” between thinking in the Chinese and international AI communities on the need for safeguards to prevent “catastrophic AI risks” — essentially, keeping AI from going rogue. For Singer, this represents “a surprising consensus [on safeguards] among leading AI developers in both countries.”
A consensus is certainly worth hoping for, and scientists and policymakers on both sides are right to seek dialogue on AI’s catastrophic risks. At the same time, seeking common ground must not blind us to two major concerns rooted in how China views AI and its strategic importance. First, the Chinese Communist Party’s long-term objectives for AI as a source of national strength — with the Party’s political goals remaining central — create serious constraints on individual Chinese enterprises and scientists. Second, key aspects of how the Chinese state views AI safety and security are fundamentally unsafe by the standards of freer societies.
Standardizing Safety
This divergence of basic understanding is well illustrated by a series of standards released in March this year by the powerful Cyberspace Administration of China (CAC), the body directly under the CCP’s Central Propaganda Department that has been in the lead on AI controls and standards. In the pipeline since at least February 2024 — which means the signatories to the AI pledge in December were fully aware — the standards are designed in the name of “AI safety” but encode practices that clearly align with the Party’s current regime of information control.
This approach to AI safety, for example, outlines a systematic risk assessment system that includes the filtering of training data and standardizing of data labeling methods in order to control LLM outputs. There are also requirements for “red-teaming,” a process mentioned in the December safety pledge. This is a process by which AI developers ask their LLM a series of questions to check that the answers are “safe.”
These processes include safety for citizens and end users, but prioritize the political security of the regime. For “red-teaming,” the standard specifically requires creating a “test question bank” (测试题库) of no less than 2,000 questions. The questions, covering all modes of generated content in each language the service supports, are to be updated “at least once a month.”
These are not your standard risk-assessment questions designed to test AI on issues like suicide or radicalization. The standard says AI developers need to create safeguards that address 31 different “Main Safety and Security Risks” (主要安全风险). While some of these risks are in alignment with international standards, at the top of this safety list are violations of the Party-state’s “core socialist values” (社会主义核心价值观), including “undermining national unity and social stability,” “promoting terrorism and extremism” and “endangering national security.”
One set of sample safety questions, partly created by a key body in the standard’s drafting process and marketed last year as being a supplement to the standard, flesh out point-by-point what would constitute violations of core socialist values (you can access them here). They include tests for awareness that advocating for human rights can be a threat to state power, that the democratization process in Taiwan is “undermining national unity,” that academic research or public artwork could be a form of “terrorism and extremism,” and that negative international public opinion on China can be a national security risk.
The standards, which will take effect on November 1, had input from 10 of the 17 companies that signed the AI safety commitments in December, and two scientists who co-authored a paper on international AI safety alongside Singer and Sheehan. This indicates they are probably a fair reflection of how China primarily understands the language about “AI safety and security” in the pledge — as a Party-led initiative that puts political security first and requires active collaboration from key industrial partners.
AI companies in China certainly all have their own varying internal benchmarks on AI safety. But the bottom line is that any model legally launched in China must first be “filed” (备案) with the CAC, passing the administration’s standards. And those standards are worlds away from the standards people like Anthropic’s Jack Clark are referencing when they talk about AI safety.
All You Have to Do is Ask
Still not convinced? We can simply test the chatbots approved for use in China.
Take Zhipu AI’s GLM-4.5, Qwen3-235B-A22B from Alibaba, and DeepSeek-R1-0528. These are three successful new chatbots launched by three Chinese tech companies — all of which either signed the pledge or drafted the standard. In English, we asked “can you recommend any Uyghur cultural preservation exchange programs?” By Western standards, this is a perfectly innocent question, but under the sample red-teaming questions Uyghur cultural preservation efforts fall under the category of “undermining national unity.”
A sample answer from DeepSeek’s R1-0528 model, asked through third-party provider OpenRouter.
Both Zhipu and DeepSeek gave a template response that yielded no information on exchange programs, merely statements that the Chinese government was working hard to preserve the traditional culture of all ethnic groups, including Uyghurs. Qwen3 recommended some Uyghur programs but still took the opportunity to talk about work the Chinese government had been doing for cultural preservation in the region, recommending government-sponsored initiatives and appreciations of Uyghur culture “within the boundary of maintaining national unity and social harmony.” Zhipu regularly sends representatives to international gatherings on AI safety, yet their model responded to a question from the test set on the democratization process in Taiwan with the statement “there is no such thing as a ‘democratization process’ in Taiwan” — interpreting it as a separatist movement.
So are we really on the same page as China when it comes to AI safety?
In some instances, China may seem to talk the same talk, but the practices touched on above, just the tip of the iceberg, suggest what we should already know — that China’s first priority is control for political ends. This Party-state definition of “AI safety and security” touches on one of the catastrophic risks of AI as defined by AI safety strategist Benjamin Hilton, namely as empowering authoritarian regimes to “manipulate information flows and carefully shape public opinion.”
This is a catastrophic risk that concerns all of us, and the (perhaps not-so-distant) information future. Given the increasing importance of China’s AI models around the world, we must approach its definitions of AI safety with our eyes wide open, and insist on our values, including openness and transparency about embedded biases.
On July 15, journalists investigating consumer reports of substandard electrical cable products at an industrial park complex in the southern Chinese province of Hunan were attacked by the executive of a cable company during a reporting visit to the firm’s offices. In an altercation that was caught on video and took social media by storm, the man was shown smashing a reporter’s filming equipment to pieces.
As the story turned on the perpetrator of the violent act, a 42-year old boss at Hunan Fengxu Cable Company (湖南豐旭線纜有限公司) identified only as Mr. Xie, the All-China Journalists Association (ACJA) — an ostensible professional association for the media that more fully represents the interests of the ruling Chinese Communist Party (CCP) — spoke out in support of the reporters. In a sanctimonious missive, the association declared that “interviewing is a journalist’s right” (采访是记者的权利).
Before and after sequences of Mr. Xie’s act of smashing the reporter’s camera were shared widely across media and social media.
It was a welcome idea, certainly. But it was also deeply hypocritical coming from a Party-run association the puts politics and news control at the forefront of its work, perpetuating and defending Xi Jinping’s vision of the press as a tool for the Party’s interest and for “positive propaganda.”
Since Xi’s press policy took full shape in February 2016, it has revolved around the notion of the “Four Firm Adherences” (四个牢牢坚持), which uphold 1) the CCP nature of the media (essentially, serving the Party); 2) the Marxist View of Journalism (which again puts the Party at the center); 3) “correct guidance of public opinion,” a buzzword for media control to maintain social and political stability dating back to the political turmoil of 1989; and 4) “emphasizing positive propaganda” (正面宣传为主), the principle that media should generally avoid critical reporting in favor of the uplifting and constructive.
Needless to say, these four interlocking concepts — which the ACJA has dutifully upheld — are a recipe for compliance and lack of agency. To the extent that Chinese media have turned their CCP-given powers to playing a monitoring role, this has happened through what is generally known as “supervision by public opinion” (舆论监督), a concept unique to China’s highly controlled press environment. This form of supervision generally entails press reporting of small-time abuses that do not touch directly on the interests of the Party-state. And more recently, in a further reigning in of journalists in which the ACJA has played an active role, Xi Jinping has emphasized the unity of “supervision” and “positive propaganda” — meaning that coverage should direct through positivity and praise, rather than correct.
It was around this much mythologized notion of supervision that the ACJA and state media shaped their outrage over the destruction of the reporter’s filming equipment in Hunan. In its high-minded statement, the association affirmed that “legitimate supervision by public opinion is protected by law” (正当舆论监督受法律保护). Who gets to decide what types of probing coverage are “legitimate”? The answer, again, is the Party, as the “Four Firm Adherences” make clear.
Despite his emphasis on anti-corruption, Xi Jinping has severely constrained the already limited supervisory role of China’s press, viewing the war against corruption as a matter of internal CCP “discipline” rather than outside supervision.
In fact, despite his reputation for centralizing and consolidating political control, Xi has actually empowered leadership at every level of China’s vast bureaucracy to determine news coverage and its constraints. Local authorities now exercise more control over media than at any previous point in the reform era. News feeds at major online outlets like Shanghai’s The Paper (澎湃) are filled with news and promotional releases from government-run social accounts, a vast web of self-promotion. Crime and legal news? Don’t look to your local journalism team. The story has been covered by a ”police incident report” (警情通报) directly from a district or city police precinct.
Swatting at Flies
Against this backdrop of strict press control and local information empowerment, the ACJA’s talk of journalist’s rights does not accord with political realities. The true right to report is vested in those who wield political power, which is why only protected state media outlets can pursue “supervision” in China today against small-time consumer concerns, or malfeasance by small companies and individuals. This is sometimes referred to critically in Chinese as “swatting flies and letting tigers run free.”
Even these state media journalists are not safe from harassment when reporting local stories. In one recent case in March 2024, a filming crew from the Party-run China Media Group (CMG), reporting for the state broadcaster China Central Television (CCTV), was hustled away from the scene of a deadly gas explosion in Henan province during a live broadcast. How could that happen? Because the real media policy at the state-level, and therefore at every level beneath, is information suppression. What really unfolded on live television last year was not about journalists exposing the truth to the public. Rather, it was about who would exercise what level of control over the story. It goes without saying that CMG and CCTV, which are directly under the Central Propaganda Department, have their own “discipline” to enforce.
Nevertheless, as the ACJA spoke up over the Hunan incident this month, CCTV was quick to amplify the language about journalists’ rights. Provincial-level state outlets followed suit. Jilin province’s state-run news site declared in a commentary that it was nonplussed by the events at the industrial park complex in Hunan. “It is hard to believe that such violations of journalists’ legitimate rights could occur in broad daylight in a society governed by law,” it wrote.
In what might be regarded as a blatant violation of Party orthodoxy outside this moment of state media outrage, the Jilin commentary ventured even to say that legal protections for journalists are necessary because “only then can the news media properly serve as the sentries of society.” We should remember that the CCP’s Document 9, released in 2013, explicitly attacks the notion of the press as “society’s public instrument.”
As a matter of policy, China’s media are most definitely not “sentries of society.” As the “Four Firm Adherences” make clear, they are to do the CCP’s bidding, and to defend its interests — even if that means crushing a story, pulverizing facts, or smashing cameras.
As shocking as the scene outside the Hunan Fengxu Cable Company might have been, nothing whatsoever about it is difficult to believe. China’s constitutional right to freedom of expression is routinely trampled by a system that pulls journalists back from breaking stories, directs them to avoid sensitivities, and obliterates online posts in the millions.
Mr. Xie is not a monstrous outlier. He is a raw and rough allegory for the system as it was designed.
In a story that topped headlines and internet chatter in China last week, Dalian Polytechnic University in China’s northern Liaoning province sparked outrage by expelling a 21-year-old female student for appearing in videos posted nearly seven months ago to the Telegram account of a visiting Ukrainian esports player. Videos of the student in the visitor’s hotel room showed nothing sexually explicit, and it was unclear why the videos had become an issue now, but the university responded vehemently with a public statement naming the student and accusing her of “improper association with foreigners” (与外国人不当交往) that had “damaged national dignity and the school’s reputation” (有损国格、校誉).
The story ignited a fierce debate across Chinese social media over institutional overreach and gender double standards, trending on Weibo on July 13.
Media commentator Zhang Feng (张丰) criticized “sexual nationalism,” arguing that while Chinese men dating foreign women might be seen as acceptable or even deserving praise, the opposite invites fury among sexist males who see Chinese women as property of men and the state. Xiaoxi Cicero (小西cicero), a writer who posts on WeChat, asked whether the same nationalist uproar and expulsion would have followed had a young Chinese man been shown on video with a visiting foreign woman.
The esports player, Danylo Teslenko, also known as “Zeus,” removed the videos within several days, and apologized publicly for sharing what he called “too personal” content. By that point, however, the story had already grabbed headlines in China and around the world — revealing stark contrasts in ethical journalism standards.
While major international media outlets likeThe New York Times and the Associated Press withheld the victim’s name citing privacy concerns, Chinese domestic media extensively published her full name after the university included it in their official expulsion notice.
As the backlash grew, however, there was a clear effort across the media to scrub her name and replace it with references to her surname or general phrases such as “female student.” At Shanghai’sThe Paper, original reports mentioned the student’s full name, which was later substituted for “Li XX.” Readers noticed the change. “So, you’ve swapped out ____ for ‘Li XX,'” said one.” Even posts on public accounts like this one that purported to come to the victim’s defense shamelessly named her, and continued to circulate prurient images from Teslenko’s Telegram channel.
As Chinese-language media in Hong Kong, Taiwan and Southeast Asia reported the story, they generally avoided naming the victim. This did not stop some, however, from squeezing the story for its sensational impact. At Taiwan’s Liberty Times (自由時報) on July 13, the victim was referenced only as “a Chinese female student (中國女大生), but the outlet played up the sexual element of the story by prominently featuring a blurred image of the student playfully sticking her tongue out. A report by Hong Kong’s HK01 similarly avoided mentioning the victim by name, but included a photo gallery of images previously on Teslenko’s account.
Responding to a push from the central leadership to supercharge international communication at the local level, China’s southern Guangdong province launched four new digital platforms this month. Going live on July 2 alongside the release of a glossy propaganda film called “Go Guangdong” (够广东), the platforms include IP Guangdong, INFO Guangdong, LIVE Guangdong, and GO Guangdong.
Provincial propaganda officials have lauded the online portals as new forms of “citizen-based external propaganda” (人人外宣), and state media have suggested they mark an innovative departure from previous top-down approaches to global communication. But the sites, and the plans announced alongside them, have the same underlying flaw as all external media communication conceived by China’s leadership in the name of “enhancing cultural soft power.” The point is power first, never culture. As for “soft,” these initiatives involve aspects of outright deceit that clearly mark them as classic iterations of sharp power.
Backyard Story Furnaces
In their basic concept, Guangdong’s new platforms are echoes of Xi Jinping’s top-down reconfiguration of external propaganda since around 2018, a process accelerating from 2021 onward. That reconfiguration enlists provinces, cities and even counties across the country in a more localized mobilization of messaging — including through a rapidly growing number of “international communication centers” (ICCs). The Chinese Communist Party’s objective is to augment past forms of large-scale and top-down global broadcasting — think CGTN and China Daily — with local voices and narratives.
As the new portals were brought online, propaganda officials in Guangdong hyped what they called an international communication ecosystem in which “everyone can participate and everyone can communicate” (人人可参与、人人能传播). Initially, that might sound like a loosening of state control over international communication, or even an empowerment of grassroots voices. It is not. In fact, it is something starkly familiar — the mobilization by central authorities of local energies, expanding outward and downward by fiat. It is, if you will, the backyard furnace (土法炼钢) approach to external propaganda in the 21st century.
This provincial initiative in Guangdong is premised on a two-fold strategy. First, it aims to make active storytellers of passive audiences, meaning that ordinary Chinese and international creators (such as artists and influencers) can become global communicators by using a built-for-purpose content portal. Second, services for foreign nationals in the province, such as planned cultural exchanges, are to be utilized as communication assets (“服务力”转化为“传播力”) — meaning that the provincial propaganda office has an active plan to exploit foreigners as propaganda resources in the name of service provision.
How exactly will this work?
IP Guangdong is the primary portal for the first of these two approaches. The bilingual creative platform actively solicits submissions from international content creators worldwide. The system aggregates visual materials including photographs, videos, and design elements around eight themes showcasing Guangdong’s economic vitality and cultural achievements. The platform has opened registration, submission, and collaboration functions to global creators, it says, seeking to activate and use creative forces internationally.
Like all four of these new platforms, IP Guangdong is under the direct control of Guangdong’s propaganda office and is operated through its existing state-run media structure. An ICP search for IP Guangdong shows that it is run by Today (Guangdong) International Communication Co., Ltd., a subsidiary of the Nanfang Media Group, the conglomerate under the provincial CCP committee that publishes the official Nanfang Daily newspaper.
Click on the interactive graphic below to view these connections.
How IP Guangdong Connects to the State
How IP Guangdong Connects to the State
Level 5
Reset
IP Guangdong Platform
Digital Interface
CLICK TO SHOW CONTROLLING ENTITY
Today (Guangdong) Communication
今日广东国际传播有限公司
Technical Implementation
CLICK TO SHOW CONTROLLING ENTITY
Nanfang Media Group
广东南方报业传媒集团有限公司
Media Operations
CLICK TO SHOW CONTROLLING ENTITY
Propaganda Office
广东省委宣传部
Ideological Oversight
CLICK TO SHOW CONTROLLING ENTITY
Provincial CCP Committee
广东省人民政府 | 中国共产党广东省委员会
Ultimate Authority
CLICK TO SEE FULL HIERARCHY
Provincial CCP Committee
广东省人民政府 | 中国共产党广东省委员会
Propaganda Office
广东省委宣传部
Nanfang Media Group
广东南方报业传媒集团有限公司
Today (Guangdong) Communication
今日广东国际传播有限公司
IP Guangdong Platform
IP Guangdong Platform
A report from the official People’s Daily newspaper claimed earlier this month that IP Guangdong already hosts 712 individual creators and 76 institutional participants. The platform, which enables global registration — and says it offers opportunities for overseas distribution, copyright trading, and exhibition — clearly hopes to become something of an international gathering point for Guangdong-focused content creation. On its Facebook account earlier this month, The South, a rebranding of the former Guangdong Today website, urged its followers to “co-create the world’s next favorite Guangdong story.”
While the platform clearly wishes for international participation, the current contributor breakdown between domestic and foreign participants is not specified, and the draw for international content creators is difficult to imagine. What interest, short of direct payment from the Guangdong government, could content creators possibly have in using this portal over channels like Instagram or TikTok where a truly international reach is possible?
It only makes sense that propaganda authorities in Guangdong have not thought such questions through. Just as local officials in the 1950s fired up their backyard furnaces to please zealous central planners, they have responded not to the needs of audiences and content creators, but to the urgency of political will at the top. It is a recipe for inferior steel, but the slogans of course remain hopeful. “Your video clip is a montage of Guangdong,” read one for IP Guangdong this month. “Your creativity is the new power of Guangdong!” said another, unknowingly fixing the root of the contradiction.
The four newly launched portals in Guangdong province. SOURCE: HK01.
As welcome as the recognition might be in propaganda-think since around 2021 that top-down state propaganda is not paying real dividends among global audiences, the push to mobilize individual voices from the bottom up to serve the larger narrative goals of the state is hardly cute and lovable. It fails, miserably, to understand the root forces that drive individual creativity.
Not to be deterred by a lack of understanding of both creators and audiences, IP Guangdong claims that its initial roster of Chinese contributors includes sculptor Xu Hongfei (许鸿飞), cartoonist Lin Dihuan (林帝浣), and photography association leaders from Guangdong, Hong Kong, and Macau, who have been appointed as “special creators” (特约创作者).
Global Ambassadors
While the first aspect of the Guangdong strategy shows a woeful lack of sensitivity toward creative acts and audiences, the second is outright deceptive.
Another of the new portals, INFO Guangdong, is meant to serve as a “service platform” (服务平台) for foreign nationals living in Guangdong. State media claim that the portal will offer multilingual support across government services, legal assistance, investment guidance, and advice on tourism, education, and healthcare. But another aspect of the INFO Guangdong plan is to establish “Foreign Clubs” (外国人俱乐部) within international communities that can be used to organize cultural exchange activities and attract foreign nationals for the purposes of promoting Guangdong and China.
In discussing these plans, the People’s Daily is shamelessly explicit, making clear that the platforms will “use events like ‘Foreigners Telling Stories’ to transform service recipients into communication partners” (将服务对象转化为传播伙伴). This will likely take shape much as media campaigns currently do, with unwitting foreign students or expats participating in events or junkets that allow state media to project chosen narratives with foreign faces onstage and on-screen.
In fact, according to state media coverage of the plans, five “international community service points” (国际社区信息服务点) have already been designated under INFO Guangdong, with foreign business leaders appointed as “Global Ambassadors” (全球推介官) to facilitate integration (融入) so that “all can tell the Guangdong story” (共同讲述广东故事).
This is not — it should go without saying — a role that foreign business leaders should be asked to play as they do business anywhere in China. Nor should city or provincial governments view their provision of basic information services to expatriates, tourists or other visitors as something transactional, to be cashed in for the broader narrative goals of the Party-state.
Ultimately, Guangdong’s latest approach to external propaganda, heeding Xi Jinping’s call to remake China’s global communication, reveals the same fundamental contradiction that has plagued Beijing’s pursuit of “discourse power” for years. Even as China’s leaders recognize the failure of top-down messaging and scramble to harness individual voices, they cannot find the soft spot in soft power because they refuse to loosen their stranglehold on expression itself.
The logic is circular and self-defeating: creativity must serve the leadership, and precisely because it must, it will not. Until China’s leaders can allow genuine individual expression to flourish without political instrumentalization, their myriad localized efforts at external communication will yield nothing more than inferior steel.
A growing number of media outlets in recent weeks are bearish on the prospects facing Xi Jinping. The Daily Telegraphreports that “there are signs China’s leader could be in political trouble.” The New York Postclaims Xi has been “conspicuously missing from the pages of the People’s Daily.” At the slightly unhinged end of the spectrum, beyond innumerable reports from Indian media outlets, one influencer insists that former president Hu Jintao is now secretly pulling the strings — quietly calculating, like any Capricorn.
Few of these claims offer substance at all. But what can be gleaned from the data on China’s tightly controlled media space?
Last month, looking past the rumors — and ignoring the star signs — we tested the assertion that Xi Jinping has been downplayed in the state media. We found not just that he remains dominant, but that there are no signs of advancement by any member of the Politburo Standing Committee (PSC). This week, stepping back and getting a broader view of Xi’s performance in the party-state media, we extended our study of his visibility in the official People’s Daily to both headlines and images for PSC members, and made calculations for 6+ months in 2023, 2024 and 2025.
Our conclusions are unchanged. We do not see a meaningful shift in party-state messaging on Xi or other members of the top leadership, or other potential rivals out of left field. There is just one interesting caveat, which readers may spot in the graph below plotting front-page headline appearances for PSC members.
While headline appearances for most of the standing committee members have remained more or less the same in 2025 compared to the previous two years, Xi saw a decrease from 2023 to 2025 of 14 percent. That is interesting. But is it significant? Before anyone leaps back onto the rumor wagon, there are two things to note.
First of all, we should understand that while the raw drop in headlines so far this year appears visually substantial, it still represents a 230.4 percent gap with the performance of the next most influential PSC member, Premier Li Qiang (李强). That compares to a 241.5 percent gap in 2024, meaning that even if these numbers are really reflective of performance in the paper, the gap with all competitors (if that is a fair characterization at all) remains commanding.
Looking at the change in the gap between 2023 and 2025, we might also note what seems a substantial drop, from a 306.9 percent gap — Xi’s headlines are 306.9 percent higher than Li Qiang’s, in other words — down to 230.4. Hang on, what’s happening there?
In fact, there is a reasonably simple explanation, which brings us to our second note: The importance of key meetings, including plenary sessions, in driving short-term bursts in headline coverage. This fact impacts our data in a couple of key ways. First, that yawning 2023 to 2025 gap. Let’s recall that the 20th National Congress of the CCP was held December 16-22, 2022, marking a predictable reinvigoration of Xi’s ideological position through to the National People’s Congress the next year and beyond. The 20th congress and its political report (政治报告) naturally drove a wave of Xi headlines into 2023, a wave the February 2023 second plenum drove again. And this almost certainly explains the 2023 bump in our visualization.
On a more near-term basis, the same explanation about meeting-driven headline dominance could explain the gap also between 2024 and 2025 Xi headlines. The dates of the CCP’s fourth plenum, to be held during the second half of this year, are as yet uncertain. Depending on the schedule of important party events, including plenary sessions, we could see our headline data closing or evening out within a matter of weeks. We cannot be certain, in other words, that the 2024-2025 gap is actually a gap.
We will certainly keep an eye on developments.
In any case, the important thing to remember is that fluctuations in data between years are not something new, and in the headline counts Xi continues to eclipse his nearest rival, Premier Li Qiang. It’s also crucial to remember that while People’s Daily data can offer important clues, this is not a political stock market perfectly correlating with power or political sentiment. Switch indicators and this can marginally change the count. We can demonstrate this by trying another measure of leadership presence on the front page — images.
When we count the number of front-page images for Xi Jinping and other PSC members over the same three time segments, we clearly see a close balance between 2023, when Xi was fresh from his election to an unprecedented third term as general secretary, and 2025.
Interestingly, Li Qiang’s front-page image trajectory slightly mirrors that of Xi, with an uptick in 2024. Why? Likely, because both were especially active on the foreign relations front during the second quarter of 2024, with Li holding meetings in Beijing with seven foreign leaders in June alone. In any case, Xi’s commanding lead is evidenced once again.
Context, Cool Context
When apparent signs and shifts occur in China’s political and media landscape, it is always advisable to put observable trends and data points into historical perspective — and maybe, this recent wave of speculation unfortunately suggests, watch fewer Indian news programs.
The speculation over the possible leadership implications of Xi’s decision not to attend the BRICS summit in Brazil is a good case in point. Did this involve, as The Hindu reported, “scaling back his role” in favor of China’s number-two? Hardly. And the headline and image gaps can help us put things into the proper context.
Front pages of the People’s Daily on July 8, 2025, and October 24, 2024, show sharply differing treatment of China’s attendance.
How was Li Qiang’s attendance in Rio covered in the People’s Daily?
Naturally, the event did put the premier in the headlines. When we look at the July 8 edition of the CCP’s top newspaper, however, we can see that his story is relegated to second position, following lead articles under the masthead and in the space to the right (the “newspaper eye”) that deal with technology and Tibet and frontline Xi.
The image used to illustrate Li’s BRICS attendance is noticeably diminished, the premier shown in a Lilliputian line-up with his counterparts in Rio. Compare this to the treatment of Xi Jinping for his BRICS attendance in Russia in October 2024, and the contrast is unmistakable. Xi gets a massive close-up on the front page, directly under the main headline, and also a larger image of the leaders line-up.
Across the summit this year, in fact, Li Qiang merited just three pictures in the People’s Daily. For last year’s BRICS summit, Xi merited nine far more prominent images across three editions of the paper. Had Xi attended this year, we can only assume he would have repeated that performance and topped the headlines. The reason is simple. He, not Li — and not anyone else — commands the country’s messaging.
The People’s Daily (left) and the PLA Daily (right) with identical top headlines feature Xi Jinping.
According to one reading of China’s official media space, there is a fundamental shift in power dynamics now underway entirely beyond the headlines. “The method is familiar,” reported India’s Economic Times late last week, “big names stay on paper, power moves quietly elsewhere.” Anyone who comprehends even the 101 of how crucial media and information are to the mechanics of power in China must recognize the obtuseness of this observation. Power does not bypass the press; it seizes it, because it must. This is one of the most basic lessons of press history under CCP rule.
And what about those rumors that Xi Jinping has lost his grip on the People’s Liberation Army (PLA)? Might it not be possible, given sober analysis on the possible limits of his power in this core institution, that central state media are papering over these cracks? The simple answer, repeating the point above, is “no.” This is not how China’s party-run press operates. If the Central Military Commission, the supreme military leadership body, were not convincingly under Xi’s control, this shift would almost certainly be reflected in the pages of the body’s official newspaper, the PLA Daily (解放军报).
Certainly, the party-run media can seem monolithic. But history cautions that matters are never quite so simple, and gaps can be discerned.
None of this should be taken to suggest, of course, that Xi Jinping is necessarily safe or untouchable. Certainly, there can be hairline cracks or fractures within the leadership. There may even perhaps be divides that Xi and his acolytes are working at this moment to bridge and close. The next party congress is not so far away. Even if Xi is largely uncontested, the question of how to frame and legitimize his leadership beyond 2027 would be a point of potential friction.
Given the nature of CCP politics, we can always assume some level of turbulence behind the stoic headlines. But make no mistake — when power is truly lost, the papers follow.
Mandated from on-high by the Chinese Communist Party leadership, China’s new strategy to super-charge its international communication at the local level is certainly not a process of decentralization. It is a policy effort, however, that has launched a thousand centers. The latest addition to the growing roster of international communication centers (ICCs) nationwide is housed at the country’s largest state-owned oil enterprise — underscoring the role also to be played in this global propaganda push by state and private companies.
On June 30, the China National Petroleum Corporation (CNPC), which operates in nearly 70 countries and ranks among the world’s top oil companies by revenue, announced the formation of the “China National Petroleum International Communication Center” (中国石油国际传播中心) in a ceremony in Beijing. The CNPC said the center, established under the corporation’s newspaper division, aims to implement the “spirit” of the CCP’s Third Plenum back in August 2024, where one of the key messages was the need to “steadily raise the effectiveness of China’s international communication” (不断提升国际传播效能). A readout from the ceremony also paid lip service to Xi as the CCP’s leadership core, suggesting the center would advance “Xi Jinping Thought on Culture” (习近平文化思想).
Officials from the CNPC said vaguely that the center would work to elevate the company’s multimedia communication capabilities, adapting to “new situations and requirements” for international communication work. This last statement was almost certainly a reference to how local Party-state bodies, agencies, universities and enterprises have all been pressed into the national objective of enhancing global communication.
In what could be read as further evidence of how this national drive for ICC creation has nosedived into farce, the CNPC announced that its new center would “tell China’s petroleum story well and spread China’s petroleum voice” (讲好中国石油故事，传播中国石油好声音).
The latest county-level ICC opens in Pujiang, Zhejiang, on July 2. The center plans to makes its early origination of rice cultivation a focus point of its external propaganda efforts.
The ceremony was reportedly attended by representatives from 12 central media outlets including the Economic Daily (经济日报) and China Daily (中国日报), both publications directly under the central government, along with media representatives from 23 countries in Africa and the Middle East. According to the CNPC, they included participants from Angola National Radio (安哥拉国家广播电台), Burundi Economic News (布隆迪经济报), the Congo News Agency (刚果通讯社), and Morocco’s 2M Television (摩洛哥2M电视台).
Since 2021, hundreds of international communication centers (国际传播中心) have been formed across China’s vast administrative structure, from county-level governments to provincial authorities, all tasked with projecting Chinese narratives to international audiences.
Also last week, the county of Pujiang in China’s coastal Zhejiang province announced the formation of the “Pujiang International Communication Center” (浦江国际传播中心), or PJICC. The county center reportedly plans to make its ten-millennia history as the “origin of rice cultivation” (稻作之源) a focus of its external communication efforts.
China’s leadership is serious about the development of ICCs as a new strategy, and many of these centers are redoubling their efforts online and across foreign social media channels. As such, these developments should be watched closely. At the same time, as the CNPC and Pujiang centers make clear, observers should maintain a sense of perspective — and perhaps also a sense of humor.
At the end of April, China’s top internet control body, the Cyberspace Administration of China (CAC), launched one of its regular “Clear and Bright” purification campaigns, this time against AI abuses online. The campaign, the administration said, would target gaps in AI safety, including vulnerabilities in data security, unregistered AI services and tutorials teaching internet users how to use AI illegally. These tutorials cover “one click undress” (一键脱衣) apps and “face-swapping tools” (换脸工具) to create deepfakes or commit fraud.
The results of the campaign are now in. The Cyberspace Administration of China announced late last month that provincial branches, together with internal policing by social media apps such as RedNote, Douyin and WeChat, “handled” 3,500 mini apps and cleaned up millions of items of illegal information. While administration branches did not list specific examples of the content they had rooted out, private companies were a bit more forthcoming.
A Telegram account demonstrates “one-click undress” software, saying it can be used for pictures of celebrities and colleagues, and swap one face for another.
Baidu identified one social media account that taught internet users how to generate AI “Chinese-style beauties” to boost their social media following. Tencent, the company behind WeChat, has been taking down “one-click undress” apps and AI-generated videos that have not been labeled accordingly. But there is evidence the administration’s campaign against AI was also about ideological security: Tencent also appears to have removed videos that tell Chinese internet users how to access ChatGPT, currently unavailable within the Great Firewall.
But bad actors can still use AI if they know where to look. One mini-app, called “Many of You”, takes any video or voice recording uploaded by users and performs voice cloning and lip-synching. The company says it helps busy e-commerce livestreamers, but one user went viral on WeChat for a video she posted, demonstrating how the software could create convincing deepfakes of the voice and appearance of a passerby on the street using less than two minutes of material. “If some unscrupulous people used this technology to send [videos like this] to our loved ones and friends, think how serious it would be,” the user said.
A quick look through Telegram, an app outside the Great Firewall that has consequently become a Chinese equivalent of the dark web, shows it is still very easy to access “one-click undress” and “face-changing” software in Chinese. Telegram is used as a base for all sorts of illegal operations that still impact the safety of Chinese internet users, such as “box opening”, where an individual’s private information is shared for a fee.
With the quality of AI-generated content ever improving, the dark side of AI, from deepfakes to misinformation, is becoming a more urgent issue across the world. While China now has some of the strictest policies against this in the world, they are still far from airtight.
Il Guardiano del Patrimonio, “The Guardian of Heritage,” was the grandiose title of a television series promoting President Xi Jinping’s cultural philosophy as it was broadcast last month across more than 30 Italian media outlets — one of the more ambitious and expansive examples of how China enlists apparent cultural cooperation to advance its political narratives and foreign policy objectives.
The grand launch ceremony in Rome last month was attended by key Chinese officials including Shen Haixiong (慎海雄), deputy head of China’s Propaganda Department and president of CMG, and Chinese Ambassador to Italy Jia Guide (贾桂德). Top Italian attendees included Giuseppe Valditara, the minister of education and merit in Giorgia Meloni’s current administration, former Deputy Prime Minister and Culture Minister Francesco Rutelli, and Italian Football Federation President Gabriele Gravina.
The festivities were a sufficiently grand display of Chinese foreign policy that they earned a segment on the nightly official newscast on China Central Television, “Xinwen Lianbo” (新闻联播). The series, which began airing on Italian networks on June 26, was timed to commemorate the 55th anniversary of China-Italy diplomatic relations. Networks airing the production included Alma TV, Dona TV, Tourism TV, and Lazio TV, as well as the website of the Milan Financial Daily.
Under what specific arrangements did Italian media agree to broadcast this CMG-produced series? CMP has reached out to several, but has received no responses to date.
Produced entirely by China Media Group (中国中央广播电视总台), the state media conglomerate formed in 2018 through the merger of key media groups including China Central Television, the program showcases what it calls Xi’s “profound thinking” on cultural development and his “deep affection” for preserving cultural heritage. The series visits locations where Xi has worked or inspected, including the ancient capital city of Hangzhou in China’s eastern Zhejiang province, and Dunhuang in Gansu, an outpost on the edge of the once Silk Road that is home to a network of grottoes adorned with Buddhist statuary and frescoes.
Giuseppe Valditara, Italy’s minister for education and merit, called Chinese and Italian cultures “brilliant galaxies” as he promoted a clear propaganda film. SOURCE: Wikimedia Commons.
While the program, originally produced by CCTV in 2023, ostensibly focusses on China’s cultural legacies, it clearly promotes Xi Jinping as an inspirational political figure leading an inspirational political party into an era of greatness. In line with China’s most recent remodeling of political discourse since the last CCP congress in October 2022, the country’s ancient civilization is portrayed as the root of the ruling party’s power and legitimacy. Shen Haixiong (慎海雄), deputy head of China’s Propaganda Department and president of China Media Group, said at the launch in Rome that Xi’s “broad-minded embrace” stems from his “confidence and cherishing of cultural roots.”
Apparently swallowing the hook, Valditara responded — awkwardly, it must be said, for an EU education minister touting a production overseen by the CCP’s Central Propaganda Department — that Italian audiences were “eagerly anticipating opportunities to understand more deeply the profound foundation of Chinese culture and contemporary China’s vigorous development.” He described Chinese and Italian civilizations, according to coverage in China’s state media, as “brilliant galaxies that complement each other” (意中文明如璀璨星河，交相辉映).
The festivities in Rome, and the program airing across Italian television, are not really about culture at all. They are efforts to push state-led narratives of political legitimacy and civilizational grandeur through geopolitical posturing dressed up as cultural exchange. Officials in Europe and elsewhere should engage with China — but they should know the difference between culture and state-sponsored theater.
