Summary of Key Points
While AI has led to layoffs in some companies, it has also created numerous new job opportunities. Among these, botsitting (the role of an AI "nanny") is a promising new profession. This role involves teaching colleagues how to use AI, checking the quality of AI-generated content, and integrating AI into business processes. Botsitting requires relatively low qualifications and is suitable for recent graduates, as they are "native to the AI era." It is likely to become a standard job position, similar to data annotation, due to the enduring needs of companies for tailored solutions and accountability.
1. AI Doesn't Only Lay Off Workers; It Also Creates Over 200,000 New Jobs
Recently, the Ministry of Human Resources and Social Security organized an online recruitment event for internet companies, which revealed that more than 5,000 internet firms offered over 200,000 job positions this summer. Major companies such as JD.com, Tencent, ByteDance, and Meituan contributed to this increase, with roles in cutting-edge areas like AI algorithms, large-scale model applications, and high-performance computing.
In addition to the increase in job numbers, new job types have emerged, including algorithm engineers who write AI code and prompt engineers who teach AI how to perform tasks. Another emerging role is botsitting, which essentially involves taking care of AI systems. For example, Microsoft's Copilot team has positions for "AI trainers" and "digital adoption specialists" that perform these duties.
2. What Does Botsitting Really Involve?
The core of botsitting is to manage and support AI systems, focusing on three main tasks:
1. Feeding Information: Since AI doesn't know about your company's product lines or project context, you need to provide it with relevant information.
2. Checking for Errors: AI can sometimes generate incorrect data or misunderstand requirements, so you need to verify the content carefully.
3. Integrating into Processes: You must format the AI-generated output to meet the company's business needs and integrate it into existing work processes.
For instance, when using AI to write a market analysis report, you would need to provide information about the company's products, verify competitor data, and format the report. This entire process can take 1-2 hours.
Studies show that 87% of white-collar workers can save 13 hours per week using AI, but they also spend 6.4 hours on botsitting, effectively neutralizing the time savings. Another challenge is the "contextual burden": when switching to a new AI tool, you have to re-feed it with the same company context, and adding 10% more information can increase the workload by 25%. The more you rely on AI, the more time you spend on managing it.
3. Why Are Recent Graduates Particularly Suitable for Botsitting?
There are three main reasons why recent graduates are well-suited for this role:
1. Low Entry Bar and Wide Range of Experience: You don't need to understand algorithms or code; knowing how to identify errors in AI-generated content (such as inconsistencies or vague terminology) is sufficient. Graduates can quickly learn about different business areas within the company.
2. Experience with AI: Many recent graduates have used tools like ChatGPT/Claude for their studies, so they are familiar with AI's limitations and know when to trust its output and when to verify it. They may even have developed their own techniques for improving AI results.
3. Moderate Workload and Career Opportunities: The weekly workload of 6.4 hours is not too heavy for graduates, and there are opportunities for promotion. For example, at Scale AI, outstanding trainers can be promoted to quality analysts or project managers with salaries starting at $10-20 per hour, potentially leading to annual incomes of £40,000-£60,000 (approximately RMB 350,000-530,000).
4. Will Botsitting Exist Forever?
Yes, it is likely to become a standard job position, just like data annotation has. The reasons include:
1. Company Diversity: Each company has unique needs and cultural norms that are not reflected in public data, so someone must translate this "hidden knowledge" for AI systems.
2. Accountability: AI-generated content must be verified by humans; for example, lawyers using AI to create fake legal cases can face penalties. Regulatory bodies require human oversight of AI-generated advice.
3. Economic Trends: The demand for AI professionals (such as trainers and evaluators) is growing rapidly, especially in non-technical roles. According to the World Economic Forum, this trend will continue.
In summary, botsitting is not a temporary job; it represents a genuine need in the AI era. Just as data annotation has become a significant market, botsitting is likely to become an essential career option for many graduates in the future.