China’s factories and offices are racing to install AI systems that can do more than just answer questions. The country is betting big on agentic AI—software that acts on its own within workflows, like ordering parts or adjusting machinery without human input. But experts at the International Data Corporation’s China event this week say industries where mistakes cost lives—like healthcare, aerospace, and nuclear plants—won’t see these tools for years, if ever. The problem isn’t just the tech. It’s whether people trust it to handle critical decisions without crashing or causing disasters, even if the system itself seems perfect on paper.\n\nFor years, China has led the world in AI investment, driven by state policies pushing automation to cut costs and boost productivity. The government’s “Made in China 2025” plan, for example, sets aggressive targets for smart manufacturing. But when it comes to letting AI run things like surgery robots or airplane engine checks, engineers and regulators aren’t so eager. A 2023 survey by the China Academy of Information and Communications Technology found that 78% of healthcare professionals distrust AI systems to make real-time medical decisions, even if the error rate is lower than human doctors.\n\n## Why high-risk industries hesitate\nThe hesitation isn’t irrational. In 2022, a Tesla Full Self-Driving beta test in China reportedly caused a crash when the AI misread a road sign. While no one died, the incident became a case study in how quickly autonomous systems can fail in unpredictable ways. Aerospace is even more cautious. Comac, China’s state-owned aircraft maker, still requires two human pilots in every cockpit, even though AI could handle some flight tasks. A senior engineer at Comac told industry analysts last year that the company won’t certify AI for critical flight controls until it can prove it fails no more than once every 10 billion hours—a standard even human pilots struggle to meet.\n\nAI agents also face a trust gap with the public. In 2021, a Chinese hospital piloting an AI system to diagnose lung cancer saw patients refuse its recommendations 40% of the time, preferring human doctors instead. The issue? The AI’s decisions weren’t just technical; they felt impersonal. “Patients don’t care if the machine is 99% accurate if it can’t explain why it made a call,” said Dr. Wang Lei, a medical AI researcher at Peking Union Medical College. “Trust isn’t just about numbers. It’s about the story.”\n\n## China isn’t alone in its caution\nOther countries share China’s skepticism. Germany’s Fraunhofer Society recently delayed a project to automate nuclear plant inspections after regulators raised concerns about AI’s ability to handle rare but catastrophic failure modes. The U.S. FDA has approved only a handful of AI-driven medical devices—most of them for analyzing X-rays or MRIs, not making treatment decisions. Even tech giants like Baidu and Huawei are pivoting from flashy agentic demos to safer, smaller-scale AI tools for factories, where mistakes cost money but rarely lives.\n\n## What happens next?\nIndustry watchers expect China to keep pushing AI in lower-risk areas first. Factories making electronics or textiles are likely to see more autonomous agents over the next two years, especially as labor costs rise and supply chains stay volatile. But in sectors where a single error can kill dozens, progress will be slow. Regulators are drafting stricter rules for AI in healthcare and aviation, with some drafts reportedly requiring human oversight in all critical decision points.\n\nFor now, the AI boom in China feels more like a cautious upgrade than a revolution. The tech is here, and it’s powerful. But whether it’s ready to run the show in industries that can’t afford mistakes remains an open question.

What You Need to Know

  • Source: SCMP
  • Published: May 17, 2026 at 07:00 UTC
  • Category: World
  • Topics: #scmp · #asia · #china · #world-news · #machine-learning

Read the Full Story

This is a curated summary. For the complete article, original data, quotes and full analysis:

Read the full story on SCMP →

All reporting rights belong to the respective author(s) at SCMP. GlobalBR News summarizes publicly available content to help readers discover the most relevant global news.

Curated by GlobalBR News · May 17, 2026

🇧🇷 Resumo em Português

A China avança a passos largos no desenvolvimento de inteligência artificial, mas agora enfrenta um obstáculo crítico: a desconfiança em setores de alto risco, como saúde e aeronáutica, pode frear sua expansão. Especialistas alertam que a implementação de sistemas autônomos nesses campos, conhecidos como agentic AI, esbarra em questões de segurança e confiabilidade, colocando em xeque a ambição do país de liderar a próxima revolução tecnológica.

O problema ganha relevância global, sobretudo para o Brasil, que busca modernizar sua indústria e serviços de saúde com tecnologias inovadoras, muitas vezes importadas ou desenvolvidas em parceria com a China. A falta de transparência e os riscos de falhas em sistemas críticos — como diagnósticos médicos ou controle de voo — levantam dúvidas sobre a viabilidade de tais soluções no mercado brasileiro, onde a regulação ainda engatinha. Além disso, a dependência de tecnologias estrangeiras expõe vulnerabilidades estratégicas, especialmente em um cenário de tensões geopolíticas.

Enquanto Pequim tenta contornar esses desafios com regulamentações mais rígidas, o Brasil precisa acelerar suas próprias diretrizes para evitar ficar para trás ou, pior, dependente de soluções pouco confiáveis.

🇪🇸 Resumen en Español

China acelera su apuesta por la inteligencia artificial autónoma en sectores críticos como la sanidad y la aeronáutica, pero el escepticismo sobre su fiabilidad frena su despliegue. Expertos advierten que los sistemas de IA “agenticos” —capaces de tomar decisiones sin supervisión humana— chocan con la desconfianza de los reguladores y la sociedad, lo que podría ralentizar su implementación en áreas donde un error tiene consecuencias graves.

La ambición de Pekín por liderar esta tecnología choca con los desafíos éticos y técnicos, especialmente en un contexto donde la transparencia y la rendición de cuentas son reclamadas por la ciudadanía y los organismos internacionales. Para los hispanohablantes, este debate cobra relevancia al plantear si Europa y Latinoamérica deben acelerar sus propias normativas para evitar depender de modelos extranjeros en sectores estratégicos, o si, por el contrario, el retraso chino abre una ventana de oportunidad para desarrollar alternativas más seguras y alineadas con valores democráticos.