China’s leader, Xi Jinping, has consolidated power, he has also sidelined women. The powerful Politburo, which usually has 24 members, has been exclusively male since 2022. No woman has ever made it to the more important Politburo Standing Committee. No wonder China has been tumbling down international rankings of gender equality. At the heart of Communist Party policies towards women is a crushing error. The party is worried about the country’s demography. But it imagines the solution lies in browbeating women and telling them, in traditional tones, to be “virtuous wives and good mothers”. The result is an approach to women that will not only fail to realise the party’s narrow aims, but will also make countless women miserable. Mao Zedong famously claimed that women hold up “half the sky”, but you would never know it today. The party’s disastrous one-child policy, in place until 2015, led to millions of female fetuses being aborted. The inevitable result is that millions of women are now missing. Their absence leads to other terrible outcomes. Counting men aged 23-37 and women aged 22-36, China has 22.5m more men than women. To be a catch, a man needs an education, a solid job, a flat and, in many regions, to pay a hefty fee to his bride’s family. Around a third of young male migrants who intend to marry think they have only a 50% or lower chance of doing so by the age of 30. As a result, resentments abound: many men, especially rural ones, are chauvinists. The sense of bitterness is likely to spread. Women make up more than half of those in higher education; men who lack education or economic prospects risk falling ever further behind. The Chinese government cannot suddenly conjure up millions of missing women to alleviate the situation. But it could introduce policies that boost the marriage rate among Chinese women while improving the choices available to them. One idea is to ensure that women have proper rights within marriages. The enforcement of laws on harassment, marital rape and domestic violence is woefully inadequate. More might marry if they were assured of fair treatment in cases of divorce, including payment of child support. If it were easier for wives to divorce bad husbands, more might remarry.
Fairer divorce settlements might also hasten the end of bride prices. Research suggests a young male migrant must save for six years to afford a bride price of 127,300 yuan ($18,780). A law bans extortionate bride prices already, but it is ambiguous and poorly enforced. Many parents still set great store by the price a daughter can command; the desperation of men to find a partner means that those with money will pay; female divorcees sometimes retain a chunk of it if marriages fail. But the payment both commodifies women and disadvantages impecunious men in rural areas in particular. More generally, the government could lower the hurdles to home ownership and work that stand in the way of couples. It should buy more of China’s vast tracts of unsold housing, converting it into affordable rental homes favoured by young families. Recent relaxations in the hukou system of internal passports, which make it easier for newcomers to register themselves and their families for schools and other public services in cities, are welcome. Traditionalist men, including those at the top of the Communist Party, may see some of these ideas as foreign impositions. If they stopped chiding Chinese women and tried asking them for their opinions, they might hear a different story. ■ Subscribers to The Economist can sign up to our Opinion newsletter, which brings together the best of our leaders, columns, guest essays and reader correspondence. This article was downloaded by zlibrary from https://www.economist.com//leaders/2026/07/16/chinas-rulers-have-a-woman-problem
How to make AI safe—and lessen dependence on America and China State-backed efforts to catch up with frontier models are doomed July 16th 2026 PROGRESS IS rapidly being made towards regulating the release of powerful new AI models in America. Its government is set to soon outline its plan, after taking suggestions from the industry. Speaking to The Economist this week Sir Demis Hassabis, the boss of Google DeepMind, pitched a hybrid public-private regulator for the industry, echoing our own thoughts. The regulator is modelled on the agency that regulates brokers and stock markets. Rules are needed because the newest frontier models possess some dangerous powers—from hacking critical digital infrastructure to writing recipes for new bioweapons. Without any organised way of dealing with such risks, the government has taken to making up the rules as it goes along in order to control access to Mythos and Sol, Anthropic and OpenAI’s latest offerings. Ideally, predictable rules would be agreed on internationally, especially between America and China, whose models surpass everyone else’s. But time is short. Sir Demis thinks that, if America moves unilaterally, the rest of the world will sign up to its system. He may be right. Yet safety is only one part of the problem. It is increasingly clear that America and China will also restrict access to their models for economic and strategic reasons. Under President Donald Trump, America has had no qualms about using its allies’ military dependence as negotiating leverage; it could one day try to exploit their AI dependence, too. China has treated other exports, such as rare earths, as sources of geopolitical advantage, and probably views AI the same way.
The control that these two governments exert over frontier AI puts other countries in a very tight spot, especially as the technology spreads. If America cuts off a country from its models and data centres, “Are your factories going to still run?” asks Arthur Mensch, head of Mistral AI, Europe’s best model maker, on this week’s “Inside Tech” show. Access to AI will also be increasingly important for security. What, then, can other countries do? A state-backed effort to catch up with America’s frontier models would be doomed. OpenAI and Anthropic have each raised well over $100bn in funding so far. Any effort that spends much less is not credible. Even if politicians find the cash, AI research is the sort of endeavour—involving expensive researchers and risky bets—in which governments have a feeble record. Who today remembers Quaero, the rival to Google that the French and German governments announced in 2005? Other countries can, though, build data centres. Most will need some of them, to run AI on sensitive data. Local data centres offer insurance against being suddenly cut off from processing power, or “compute”. They also make it easier to maintain a credible ability to switch to open-weight models, which anyone can run if access to frontier capabilities is cut off. Unfortunately, much of the world is far behind the data-centre buildouts in America and China. Usually, the biggest problems are regulation and access to power. By delaying revenues and locking up capital, procedural snafus can have enormous effects: a nine-month delay worsens the economics of a data centre as much as would doubling its lifetime electricity bill, according to the Carnegie Endowment, a think-tank. The average delay before being connected to the grid is three years in India and Britain and three-and-a-half in Germany and South Korea, compared with two years in America. Speeding up regulatory approval, offering fast-tracks through connection queues and allowing power generation that bypasses the grid would all help. There is no need for subsidies: AI companies are desperate for compute and will pay to get it. There is even the enticing possibility of agreements where AI developers promise to offer a country the same models as America in return for data centres. Simply becoming a big customer of America’s AI labs, by encouraging the diffusion of the technology, could itself build leverage. Model makers need revenue to justify their gigantic investments;
they will have a reason to lobby against American protectionism that hurts foreign buyers. Some countries are in the fortunate position of having their own choke points. Taiwan has chipmaking; the Netherlands has ASML; South Korea has memory technology. These specialisms should be prized. But even the lucky few cannot build fully sovereign AI capabilities. So countries must prepare to deal transactionally with the AI superpowers. Democratic America is a better partner than authoritarian China. Even so, every source of negotiating leverage will count. ■ Subscribers to The Economist can sign up to our Opinion newsletter, which brings together the best of our leaders, columns, guest essays and reader correspondence. This article was downloaded by zlibrary from https://www.economist.com//leaders/2026/07/15/how-to-make-ai-safe-and-lessen- dependence-on-america-and-china
America should stop making it so hard to have fun Bureaucracy is nightlife’s toughest bouncer July 16th 2026 In 1660 the laws of the Puritan Massachusetts Bay Colony declared that its nascent watchmen should “examine all night walkers after ten of the clock at night”. This year Boston’s largest police union offered up a modern equivalent, opposing the extension of bar hours to 3am on the grounds that: “Truth be told, not a lot of good happens after midnight.” For a country that prides itself on defending individual liberty, America has long been surprisingly prissy about nightlife and the boozy, exuberant and sometimes disorderly fun that comes with it. It should learn to lighten up. Rules about alcohol are often to blame. When Prohibition ended in 1933 a patchwork of laws emerged in its place, from caps on the number of liquor