Gasgoo Munich-"I don't believe AI will directly trigger mass layoffs, but many companies are repackaging business-driven headcount optimization as an AI-fueled transformation." Liu Songbo, vice dean of the School of Labor and Human Resources at Renmin University of China, recently discussed talent shifts in the auto industry amid the AI wave, calling the practice "AI-washing layoffs."
In February, Gallup surveyed 23,000 American workers; 660 of them remained unemployed after losing their jobs. Among this group, only about 1% pointed to AI as the direct cause of their unemployment. Far more blamed organizational restructuring, cost-cutting, or the broader economic environment.
The auto industry has been pushing AI transformation while simultaneously adjusting headcount in recent years. Because the two trends overlap, they are easily conflated. That overlap is the crux of Liu's argument: AI will jolt employment, but the underlying business shifts within companies deserve just as much attention.
AI accounts for only a sliver of corporate layoffs
Job losses tied to AI are already a matter of public record.
In early 2026, fintech firm Block announced layoffs of roughly 4,000 employees—over 40% of its workforce. In an open letter, the CEO noted that AI tools had reshaped how the company is built and runs, allowing smaller teams to accomplish more than before.
Oracle's case is even more explicit. In fiscal 2026, its headcount fell from roughly 162,000 to 141,000—a drop of about 21,000 in a single year. In its annual report, Oracle stated outright that adopting and deploying AI in operations has led to, and may continue to lead to, workforce reductions. At the same time, the company cited management shifts, product changes, strategic adjustments, and mergers as contributing factors.
Image source: Generated by Doubao
The auto industry is seeing similar shifts in roles.
In May 2026, General Motors cut about 600 IT staff—more than 10% of its IT department. The company said it was transforming its IT organization to better position itself for the future. Yet insiders say GM is simultaneously hiring IT talent with stronger AI capabilities.
This means a single company can eliminate some roles while adding others. The real shift is in the structure of jobs.
That dynamic tends to emerge whenever AI enters an organization. As AI boosts efficiency in certain tasks, demand for some roles drops—yet companies simultaneously add new technical capabilities. Most layoffs, however, are tied to the company's operating conditions. The auto industry is squarely in that environment right now.
According to the National Bureau of Statistics, auto manufacturing profits totaled 195.35 billion yuan in the first half of 2026, down 19.5% year on year. Shrinking margins force companies to rethink costs and resource allocation. Product lines, R&D projects, production scales, and staffing all feel the impact.
That pressure is already playing out in real moves. Domestic automakers rarely publicize "layoff" figures; instead, they idle inefficient capacity, scale back operations, reshuffle staff internally, or offer voluntary separation packages.
GAC Honda, for instance, closed its aging gasoline-car plant in Huangpu, Guangzhou. For the thousands of affected workers, it relied mainly on internal transfers to its new-energy vehicle plants and voluntary separation payouts. Meanwhile, the Ministry of Industry and Information Technology is pushing out outdated capacity; several legacy automakers in financial distress have idled production bases or exited the market by shedding dormant lines.
Notably, the industry is splitting sharply: as some companies pull back on capacity and headcount, leading players like BYD and Geely are still mass-hiring fresh graduates and expanding their talent reserves.
Image source: Volkswagen Group
Overseas automakers show a similar pattern. Global carmakers like Mercedes and Stellantis are also cutting jobs, while top-tier suppliers such as Bosch, ZF, and Continental are all streamlining organizations and trimming staff.
Based on available public information, the driving forces behind this round of capacity and staffing adjustments—both at home and abroad—center on market competition, price wars, and profit pressure. No automaker has officially cited AI as a reason for cutting staff or closing plants.
Liu's remarks target precisely this reality. He isn't denying AI's impact on employment; rather, he's warning companies against lumping all business-driven headcount optimization under the "AI layoffs" label. Li Linfeng, CEO of Haiwei Technology, also stated that AI and layoffs are not correlated.
The real question worth exploring is what specific tasks are actually changing once AI enters these companies.
As AI moves in, workflows shift first
When AI enters a company, the most visible changes are the individual tools.
Li Ning, vice general manager of the quality department at Changan Automobile, previously noted that the company uses an AI assistant to review supplier rectification reports. Tasks that used to take a human 30 minutes now wrap up in roughly two minutes.
Such tasks are tailor-made for AI. Rectification reports are packed with fixed formats and repetitive information. AI can rapidly parse the text, flag issues, compare pre- and post-rectification details, and hand off only the items requiring human review to engineers. Machines process the information; humans judge the results. The old workflow gets dismantled and rebuilt.
This is far closer to what actually happens inside companies than the narrative of "AI replacing a job." A quality engineer's role hasn't vanished into thin air, but a chunk of their time is now freed from organizing documents, hunting for information, and making repetitive comparisons. Similar shifts are rippling through R&D, procurement, customer service, finance, and IT.
Image source: Generated by Doubao
The problem comes next. If every department simply deploys a few AI tools, the company ends up with a batch of more efficient "point solutions." The original cross-departmental processes remain unchanged, data stays siloed, and staff divisions barely shift. AI has merely sped up a few isolated steps.
Deloitte Consulting has spotted this pattern. A report from the firm notes that over 80% of mainstream automakers have launched generative AI pilots, yet only about 15% have scaled those applications. Deloitte dubs this the "pilot dilemma"—the core issue being that AI apps are scattered across different scenarios and lack integration with existing business processes.
Take Changan's AI practice: the supplier rectification report is just an entry point. Once AI reads the report, subsequent steps—issue extraction, rectification tracking, quality data accumulation—should all flow into the same pipeline. AI handles the standardized information processing, freeing engineers to spend more time on problem assessment, supplier communication, and final decisions.
Change just one step, and you're no longer merely transforming "report review"—you're reengineering an entire quality-management workflow.
One global auto parts supplier uses multiple AI agents to interpret ERP and IoT data, automatically executing parallel tasks such as cross-shift production scheduling, alternative material recommendations, and equipment maintenance. After each execution, the system self-optimizes its task decomposition and resource weighting based on feedback, creating a continuous learning loop.
Then there's Geely. It has deployed over 160 AI role assistants and more than 600 AI skills, executing tasks over 130 million times. The automaker estimates the annual direct and indirect economic benefit exceeds 1 billion yuan. This is a far cry from using a standalone AI assistant. The latter tackles a single task; the former starts taking over an entire process.
For automakers, this is where AI begins to reshape the organization. A business process that once required multiple roles to work in relay—someone gathering data, someone organizing materials, someone analyzing issues, someone initiating workflows, someone following up—can now be streamlined. Once AI steps in, standardized links get compressed or automated, and human work gradually concentrates on judgment, coordination, and decision-making.
Image source: Generated by Doubao
Wang Xianbin, vice president of the Gasgoo Auto Research Institute, argues that AI's reshaping of the auto supply chain isn't piecemeal—it spans the entire chain. From product definition and R&D design to manufacturing, supply chain management, and after-sales service, AI is permeating every link. Yet an "AI-defined vehicle" isn't just about layering AI tools onto each step; it requires rebuilding the entire business process and organizational structure.
Consequently, the shifts in staffing structures driven by AI will grow far more specific.
Workloads in some roles will shrink, responsibilities in others will pivot, and entirely new positions will emerge. Changan, for example, has adjusted its hiring and talent strategy this year. Public information shows the automaker has prioritized intelligence, new energy, and software as key talent areas, with its 2026 campus recruitment plan projected to expand by more than 30%.
Industry-level data reflects this shift as well.
Zhaopin's "2026 Human Resource Management Trends Report" shows a supply-demand ratio of 3.08 for AI engineers—meaning roughly three job openings for every applicant. The ratio for automotive design and manufacturing engineers stands at 2.38, also signaling a tight talent market. The report also notes a sharp uptick in hiring demand for roles related to large language models and autonomous driving.
Within the auto sector, this structural deficit is even starker. In 2025, the new-energy vehicle industry faced a talent gap of 1.03 million people. R&D talent for electric powertrains and intelligent driving is especially scarce: the supply-demand ratio for autonomous-driving engineers is just 0.38, meaning roughly three roles are competing for one qualified candidate.
Gasgoo's previous reporting on the issue echoes this: automakers' shortages no longer stop at traditional software engineers. Finding talent who understands vehicles, algorithms, and data—and can embed that technology into specific automotive scenarios—is growing ever harder.
So the true change AI brings ultimately boils down to something highly concrete: for any given task, who will do it, which steps they will handle, and how broad a scope one person can cover.
The skills individuals need to build are already shifting
AI's impact on individuals is already showing up in how companies hire and train.
Hiring standards are adjusting. Geely brought on over 4,000 fresh graduates in 2026. XPENG plans to raise the share of campus hires from 50% to 70%, reasoning that younger workers adopt new tools and learn faster. Changan and some other automakers have even spun off their industry-education integration operations into standalone entities, embedding real-world corporate needs into school curricula ahead of time.
Job requirements are evolving too. Li Zhele, secretary-general of the Auto Talent Professional Committee under the China Talent Research Association, has outlined how the auto talent competency model is changing: where problem-solving sufficed before, complex problem-solving is now the baseline; learning ability has given way to rapid learning; and imitation-driven innovation is increasingly replaced by the demand to innovate first.
An in-vehicle display engineer, for instance, used to focus on hardware, structure, and display tech. As smart cabins evolve, they now need to grasp chips, software, algorithms, and user experience. Someone handling overseas business must simultaneously navigate product, regulatory, channel, and local market dynamics.
This skill set defines the "all-around employee"—the versatile, cross-functional talent everyone talks about. In the AI era, such workers need enough logical rigor and knowledge breadth to act as business architects and coordinators, deploying AI agents to shoulder heavy execution work. Crucially, though, humans must still be able to judge the reliability of AI outputs.
Image source: Generated by Doubao
Haiwei Technology has reportedly baked these requirements into its roles. CEO Li Linfeng sketched a possible future work model: a standout employee could direct 30 to 50 AI agents, taking on the workload of an entire business unit. Such an employee must first grasp the "silicon-based" work logic, then possess sufficient logical ability and knowledge breadth.
This means AI proficiency is migrating from a specialized skill for tech roles to a baseline capability across a much wider range of positions.
但这并不意味着所有人都要变成AI工程师。
An R&D engineer can have AI generate multiple technical proposals, but the final choice still hinges on cost, performance, reliability, and project timelines. A product manager can use AI to sift through reams of user feedback, yet deciding which features to build still comes back to product positioning. An overseas business lead can let AI analyze local markets, but still must personally evaluate partners, channels, and the commercial environment.
Discussing AI's limits, Liu noted that the technology excels at processing standardized, procedural information and tasks. But value judgments, timing calls, and bottom-line decisions in complex scenarios still require a human—someone who can shoulder responsibility and pressure.
The auto industry happens to be full of such work. R&D staff must trade off technology against cost; quality teams must diagnose issues based on on-site conditions; managers must balance competing departmental interests; and once a company expands overseas, variables like regulations, culture, and business customs multiply.
Changan's experience in Thailand illustrates the real-world demand for such composite skills. Entering the local market, the company confronted more than just product and technology issues—Chinese and local employees also held different views on overtime culture.
The threshold for an all-around employee isn't "knowing everything." It's about ensuring your knowledge boundaries don't stay trapped inside one familiar niche. AI lowers the barrier to cross-disciplinary learning, and individuals need to gradually widen their own capability ranges.
Liu's advice to avoid "excessive anxiety," applied to the individual, means there's no need to rush into guessing whether AI will replace you. Instead, make AI a tool in your workflow while continuing to build your professional expertise and judgment.
The first step is simply to start using AI—and make it work for you.
