Where did the entry-level jobs go?初级岗位都去哪了？

For decades, the deal was simple: study hard, graduate, take a junior role doing the routine work nobody senior wanted, and climb from there. That first rung of the ladder — the report-drafting, data-cleaning, ticket-answering, code-fixing entry-level job — is exactly what generative AI does best. And around the world, the data now shows the first rung getting thinner.

几十年来,职场的规则很简单：好好读书,顺利毕业,进公司做一份"资深员工不想做的杂活"的初级工作,然后一步步往上爬。而职业阶梯的第一级——写报告、清洗数据、回复客户、修补代码的初级岗位——恰恰是生成式 AI 最擅长的事。如今,全球的数据都指向同一个事实：阶梯的第一级正在变窄。

What the global data shows全球数据怎么说

The pattern isn't limited to the United States. Consulting firm EY reported Indian IT services companies cutting entry-level roles by 20–25% through automation and AI. Job platforms recorded a roughly 35% decline in junior tech postings across major European markets in 2024. And the World Economic Forum's Future of Jobs Report 2025 estimates that 39% of workers' core skills will be transformed or outdated by 2030.

这个趋势并不局限于美国。咨询公司 EY 的报告显示,印度 IT 服务企业因自动化与 AI 削减了 20–25% 的初级岗位；欧洲主要市场的招聘平台在 2024 年记录到初级科技岗位约 35% 的下滑；世界经济论坛《2025 未来就业报告》则估计,到 2030 年,39% 的劳动者核心技能将被改写或过时。

The crucial detail most headlines miss大多数新闻标题漏掉的关键细节

Here's what matters more than the scary numbers: the declines are not evenly distributed. The Stanford research found job losses concentrated in roles where AI can fully automate the work — while roles where AI augments human effort have stayed stable or grown. Machine learning and AI-related postings have surged even as junior generalist roles shrank. Economists like Acemoglu and Restrepo have long argued that technology destroys some tasks while creating others — and the new tasks tend to reward judgment, communication, and the ability to direct the technology itself.

比吓人的数字更重要的是这个事实：这场下滑并不是均匀发生的。斯坦福的研究发现,岗位流失集中在 AI 能够完全替代工作内容的职位上——而那些 AI 只是辅助人类的职位,不但稳定,甚至还在增长。初级通用岗位收缩的同时,机器学习与 AI 相关岗位的招聘量却在激增。经济学家 Acemoglu 与 Restrepo 早就指出：技术在消灭一部分任务的同时也在创造新任务——而新任务往往更看重判断力、沟通力,以及驾驭技术本身的能力。

The question for the Class of 2026 isn't "will AI take my job?" It's "am I aiming at a job AI replaces, or a job AI amplifies?"对 2026 届毕业生来说,真正的问题不是"AI 会不会抢走我的工作",而是"我瞄准的,是会被 AI 替代的岗位,还是会被 AI 放大的岗位？"

And Malaysia?那马来西亚呢？

Malaysia enters this shift from a position of apparent strength — national unemployment at a decade low of 3.0%, graduate unemployment at just 3.2%. But as we explored in our Barometer research note, over a third of employed Malaysian graduates already work below their qualification level. If the global squeeze on entry-level knowledge work reaches Malaysia at scale, it will land on a graduate population that already has thin margins.

马来西亚是带着一份看似漂亮的成绩单走进这场变革的——全国失业率 3.0% 创十年新低,毕业生失业率仅 3.2%。但正如我们在「晴雨表」研究札记中所写：超过三分之一的在职毕业生,工作已经低于其学历水平。如果全球对初级知识型岗位的挤压大规模传导到马来西亚,它将落在一个本就没有多少缓冲空间的毕业生群体身上。

Whether that's already happening — which industries, which majors, how fast — is exactly what Insightra's research programme is built to track. We'll map AI's impact on Malaysian entry-level hiring using local data, not imported headlines.

这一切是否已经在发生——影响到哪些行业、哪些专业、速度有多快——正是 Insightra 研究项目要持续追踪的问题。我们将用本地数据,而不是照搬海外头条,来绘制 AI 对马来西亚初级岗位招聘的真实影响图。

What a fresh graduate can actually do应届毕业生现在能做什么

• Move toward augmentation. Target roles where AI is a tool in your hands, not a replacement for them — and learn the tools of your field before your first interview, not after.
• Build judgment, not just output. AI produces drafts in seconds; the value shifts to knowing what good looks like — checking, choosing, and improving.
• Make yourself legible. With fewer junior openings, a portfolio of real projects beats a CGPA on a filtered-out CV.
• Look at the builders. Younger companies in AI and emerging tech hire for skills and hunger over pedigree — often the most open door for a fresh graduate.
• 往"AI 辅助型"岗位靠。选择那些 AI 是你手中工具、而不是你的替代品的职位——并在第一场面试之前,就把你所在领域的 AI 工具学起来。
• 修炼判断力,而不只是产出。AI 几秒钟就能生成初稿,价值正在转移到"知道什么才是好的"——会检查、会选择、会改进。
• 让自己被看见。初级岗位变少时,一份真实项目作品集,胜过一张躺在被筛掉的简历上的 CGPA。
• 看看那些"造未来"的公司。AI 与新兴科技领域的年轻公司,招人更看技能与热情,而非出身——对应届生来说,那往往是最敞开的一扇门。