Why AI Ads Have Become a Crash Zone for Major Brands in China
Updated
In the first half of 2026, more than ten well-known Chinese brands found themselves in the middle of public opinion storms because of AI-generated advertising. The problems cluster around three clear categories: vulgar or edge-pushing content, visual distortions that look unnatural or grotesque, and outright false claims. What started as a promise of lower costs and faster production has turned into a recurring source of brand damage.
One of the most visible recent cases involved Supor. The long-established home appliance brand saw several of its e-commerce accounts release batches of AI short videos. The plots included men in bathrooms, women being interrupted while bathing or using the toilet, and cleaners entering private spaces, all framed as product demonstrations. The tone and scenarios struck many viewers as crude and inappropriate. The company later took the videos offline and said it would strengthen its review process. The episode was not isolated.
A feminine care brand ran an AI video that joked about menstrual odor, drawing sharp criticism from the very audience it was trying to serve. Outdoor and transit ads featuring AI-generated models with six fingers or distorted limbs appeared in public spaces, including airports. In each case the visual or narrative failure was obvious once people looked closely, yet the material had already gone live.
The Scale of the Problem….
Industry data shows the AI marketing technology market in China reached roughly 490 billion yuan in 2025 and is projected to exceed 700 billion yuan in 2026. Growth of more than 30 percent year on year has been accompanied by a sharp rise in content volume. What used to take days or weeks of creative work can now be produced in minutes. The same speed that creates opportunity also multiplies the chance of mistakes reaching the public.
Reports indicate that 15 to 17 percent of advertising agencies have already encountered compliance issues linked to AI-generated content. The speed at which problematic tactics appear has accelerated from weekly to daily. Once a format proves effective at capturing attention, similar versions spread quickly across accounts and platforms.
Why the Crashes Keep Happening
Several structural factors explain the pattern.
First, cost incentives work against careful review. When a single ad costs only a few yuan to generate, many teams decide that human checking is too expensive relative to the production price. The result is a pipeline that runs from prompt to publication with minimal human intervention. Volume replaces judgment.
Second, the technology itself has clear limits. Current generative models can produce realistic-looking images and fluent scripts, but they still struggle with cultural nuance, social sensitivity, and physical consistency. Hands remain a common failure point. Body proportions and facial expressions can slip into the uncanny valley. More importantly, models optimized for engagement often favor content that is surprising, provocative, or emotionally charged. Without strong guardrails, the system drifts toward material that attracts clicks while risking offense.
Third, the common outsourcing model dilutes responsibility. Many brands hand daily operations of store accounts and short-video channels to external agencies. Contracts frequently focus on hard metrics such as GMV, follower growth, and conversion rates. Brand safety clauses, content standards, and clear liability for AI-generated material are often weak or missing. When something goes wrong, the brand and the operator point at each other. The public, however, holds the brand accountable.
Fourth, internal processes have not kept pace with production speed. Traditional advertising workflows included multiple layers of creative review, legal checks, and brand approval. AI has compressed the timeline so dramatically that those layers become bottlenecks. Rather than redesign the process, some teams simply remove the layers.
The Real Cost of Cheap Content
The financial and reputational arithmetic is straightforward and unforgiving. A few yuan of production cost can trigger millions or tens of millions in market value loss, not to mention long-term damage to consumer trust. Supor’s share price moved lower in the days surrounding the controversy. Other brands have faced waves of negative comments, calls for boycotts, and the need for public explanations.
Brand equity is not a free resource. Years of careful positioning can be undercut by a handful of poorly judged videos that circulate widely in a short time. Consumers notice when a brand that claims to understand them suddenly publishes material that feels tone-deaf or exploitative. The gap between the brand’s stated values and the AI output becomes the story.
What Brands Need to Change
The solution is not to abandon AI. The efficiency gains are real and will only increase. The requirement is to treat AI as a powerful production tool that still needs human oversight and clear rules.
Practical steps include:
- Building mandatory review gates for any AI-generated commercial content before it goes live, especially for sensitive categories such as personal care, health, and household products.
- Writing brand safety and AI content standards into agency contracts, with clear consequences for violations.
- Requiring visible labeling of AI-generated material where regulations demand it, and applying the same transparency even when not strictly required.
- Training teams to spot both technical failures (extra fingers, distorted limbs) and cultural or ethical risks.
- Measuring success beyond short-term engagement metrics. Long-term brand health indicators should sit alongside GMV and click-through rates.
Platforms and regulators are also tightening expectations. Rules requiring identification of AI-generated content are already in force in various forms. Enforcement actions against misleading or harmful material continue to expand. Brands that wait for regulators to force better processes will find themselves reacting under pressure rather than controlling the narrative.
A Broader Industry Lesson
The current wave of AI ad failures is a classic case of technology outrunning organizational capability. The tools moved faster than the governance structures around them. In the rush to cut costs and increase output, many teams treated AI as a complete replacement for human judgment instead of a force multiplier that still requires direction.
The brands that will navigate this period successfully are those that keep the efficiency of AI while restoring the discipline that traditional advertising once enforced more naturally. Speed without standards is not progress. It is simply a faster way to make expensive mistakes.
AI advertising does not have to be a disaster zone. It becomes one when the only question asked is how cheaply and quickly content can be produced. The better question is whether the content still represents the brand the company claims to be.
We can help you with your Digital Markteing in China.
This article was written by the editorial team at SEO Agency China, a Shanghai-based agency that helps foreign brands get found and generate leads in the Chinese market.
If you want a hand with this, read more about Baidu SEO and paid search and our digital marketing team in Shanghai.
Sources
- CNNIC Statistical Report on Internet Development – the reference count for Chinese internet and search users.
- CAICT white papers (中国信通院) – the state telecom research body publishing on AI and the digital economy.
- iResearch (艾瑞咨询) research reports – Chinese market and consumer studies.
- China Briefing business intelligence – practical guidance on operating a foreign business in China.