At a glossy AI launch at Shanghai’s Expo Centre, Haier Group chairman and CEO Zhou Yunjie framed the company’s latest strategic pivot with a simple story: a year ago a Douyin user sketched a wish for a washing machine that could wash clothes, underwear and socks separately. Engineers answered within 24 hours, the user became a product ambassador, and Haier’s three‑drum “lazy” washing machine sold more than 80,000 preorders in its first week and has since cleared roughly 400,000 units globally. The anecdote has since become shorthand for Haier’s pitch that real customer signals, not tech spectacle, should drive product design and AI deployment.
Zhou used the stage to explain a personal doctrine he calls “three nots, three yeses”: don’t wait to be seen but go and see; don’t chase traffic but win hearts; don’t merely sell products but craft dreams. The story also propelled Zhou into the public spotlight after a contrasting image with fellow entrepreneur Lei Jun at the Two Sessions, and helped Haier cultivate what it describes as an IP ecosystem that now reaches some 190 million followers. That scale of direct customer contact is central to Haier’s argument that mass AI cannot eclipse the messy, human signals that point to real needs.
China’s domestic AI boom forms the wider backdrop. In the past year a cluster of large Chinese models and consumer apps stirred a wave of novel experiences — from chatbots to viral “养龙虾” games — that have excited consumers even as they have generated confusion and anxiety. Zhou’s public message taps into a broader unease about an accelerating supply of AI‑enabled products: if production and capabilities surge, will real human demands be heard, or simply buried in algorithmic noise and commodified output?
Haier’s operational response is straightforward and aggressive. The company says Zhou’s personal account has collected more than 9,000 product suggestions, 17 of which became formal projects, and that in 2025 the firm set about embedding AI across “every process and every role,” with ambitions to make it run like blood through operations. At the Two Sessions Zhou proposed two policy priorities: push embodied intelligence out of the lab into industrial frontlines, and codify an ethical baseline so AI development follows an explicit “intelligence for good” trajectory.
The roll‑out is visible in new consumer devices. Haier showcased an “AI Eye 2.0” family of appliances claimed to reach L4 intelligence across scenarios — youth lifestyles, childcare and pet care, eldercare and health management — with features such as thermostatic defrosting and directional airflows designed to respond to human comfort rather than raw algorithmic metrics. The message aligns with other high‑profile Chinese executives: Lei Jun has speculated AI could shorten working hours, while Dong Mingzhu insists AI must complement rather than displace people. The consensus among these voices is that the technology’s legitimacy depends on tangible benefits to everyday life.
That emphasis on human anchoring is both a strategic strength and a source of new challenges. Haier’s crowdsourced innovation model shortens the loop between complaint and product, accelerating time‑to‑market and creating highly visible successes. But scaling that approach risks large volumes of low‑value input, raises questions about data governance and user privacy, and tests quality control when engineering cycles compress. It also illustrates a political dimension: corporate calls for embodied intelligence and ethical guardrails dovetail with broader state aims to industrialize AI while containing social risk.
Whether Haier’s example is a model for consumer tech globally depends on execution. Rooting AI in authentic, often small, human needs can reduce the risk of “AI for show” and generate profitable, sticky products. Yet the harder test will be sustaining that discipline as companies chase scale, export markets and the productivity gains that corporate and state leaders expect. The story of a sketch on Douyin becoming a mass seller captures more than a marketing triumph; it is a live experiment in reshaping how industrial R&D, platform engagement and algorithmic design interlock in the age of pervasive AI.