Summary of Key Points
The once-desired “golden jobs” in large companies, which attracted many professionals, have now become a source of anxiety due to the widespread adoption of AI. Initially, employees thought that AI would be a benefit that would reduce their workload. However, they have found that AI has turned into a binding constraint in performance evaluations: not only has their workload not decreased, but they have also been forced into an AI-driven “arms race,” worrying that their experience will be quickly replaced by AI. Moreover, the evaluation criteria have shifted from focusing on business capabilities to emphasizing the ability to use AI. As a result, employees are caught in a dual anxiety of having to compete with colleagues using AI while fearing being replaced by it.
Detailed Analysis
1. AI: From “Free Help” to a “Binding Constraint”
At first, large companies provided employees with a large amount of tokens (similar to “game coins” used for AI operations) to use in creating plans, code, and PPTs using AI models like OpenClaw. Employees saw this as an opportunity to benefit from the company’s resources. However, it soon became clear that the usage of AI was a critical metric for performance evaluation. Using too few tokens meant being seen as out of touch with the times, while using too many was considered a waste of resources. For example, an employee in Haidian mentioned that the company mandated the use of AI, and both the amount and method of use were strictly tracked by the system, directly affecting their scores. While AI was supposed to save time, it led to more intense workloads; what used to take two hours to complete now only takes 10 minutes with AI, but employees are required to produce three versions for comparison, resulting in no net increase in productivity.
2. Forced into an AI Arms Race
Even if a task does not require the use of AI, employees are still forced to demonstrate their proficiency in using it. A staff member from Xierqi Operations noted that although the processes were already well-established, everyone is now competing to create more detailed instructions and report on the “AI penetration rate” (the proportion of work done with AI). Some have gone so far as to use AI to write long novels just to earn enough tokens. Daily work has become a three-step process: spending half an hour creating prompts for AI, another half an hour converting AI-generated content into human-readable format (to avoid suspicion of AI involvement), and finally, writing a weekly report on how AI is enhancing productivity—less time is actually spent on the actual business tasks.
3. The More You Train AI, the More You Fear Being Replaced
There’s a concept called “AI distillation”: as employees train AI, the AI gradually learns their skills, and they eventually become its learning material. A programmer named Kuku observed that the quality of code written by new interns using AI is approaching his own level from three years ago. The real reason for this is not the interns’ skill improvement but the power of AI itself. Employees are most concerned that AI can quickly acquire 80% of the experience they have accumulated over years, and since the cost of training AI (in tokens) is lower than their salaries, the more diligently they train AI, the faster it learns—and the more they risk being replaced.
4. Shift in Evaluation Criteria
The definition of a “good employee” has changed significantly. Those who used to focus on business details are now at a disadvantage due to low AI usage rates, while those who spend time training AI and creating impressive PPTs and reports receive higher scores. The focus is no longer on the progress of the business itself but on how much work involves AI. There’s even an unwritten rule that the “AI penetration rate” can be inflated by exaggeration: AI-generated content cannot be used directly to show workload, nor can it be completely avoided; employees must “perform” as if they are fully leveraging AI, taking credit for its benefits.
5. AI Dependency: Saving Time but Losing Thinking Skills
Long-term use of AI can lead to a lack of critical thinking. When AI handles tasks and generates ideas, employees save time but also give up the opportunity to think independently. For newcomers, this is like self-deception; it may seem efficient, but it does not lead to genuine skill development. For those who merely put on a show of effort, it creates a habit of “performing without thinking,” similar to creating VLOGs that appear impressive but lack substance.
In Conclusion
The anxiety among employees in large companies regarding AI is essentially about tools becoming a form of bondage. AI was intended to improve efficiency, but it has turned into another competitive arena that perpetuates the fear of being replaced. How about you? Does your job involve the use of AI? Is it a help or a source of additional stress?