The AI Layoff Wave: Who's Being Replaced?
AI 轉型裁員潮：誰在被取代？
"We no longer need the mass of people we once did to maintain our velocity." — Jack Dorsey, Block CEO, February 2026
「我們不再需要過去那麼多人來維持我們的速度。」—— Jack Dorsey，Block 執行長，2026 年 2 月
1. The Numbers Don't Lie
1. 數字不會說謊
According to outplacement firm Challenger, Gray & Christmas, AI was directly cited as the reason behind approximately 55,000 U.S. layoffs in 2025 — more than 12 times the number attributed to AI just two years earlier. By late February 2026, over 37,000 tech workers had already been impacted.
This is no longer a hypothetical future scenario. CEOs are publicly stating, in memos and on social media, that AI is replacing human roles. The euphemisms are gone. The restructuring has begun.
根據美國再就業顧問公司 Challenger, Gray & Christmas 的數據，2025 年 AI 被直接列為裁員原因的案例，累計約 55,000 人——是兩年前的 12 倍以上。截至 2026 年 2 月底，又有超過 37,000 名科技從業者受到影響。
這不再是一個假設性的未來場景。CEO 們正在公開備忘錄和社群媒體上直言不諱地宣布：AI 正在取代人類職位。委婉說辭已經消失。結構性重組已經開始。
2. Companies That Explicitly Cited AI as the Reason for Layoffs
2. 明確以 AI 為由進行裁員的公司
The following companies have had their leadership publicly and explicitly state that AI capabilities were a primary driver of workforce reductions. These are not cases where AI was merely mentioned in passing — the connection was direct and documented.
以下公司的領導層已經公開且明確聲明，AI 能力是裁減人力的主要驅動因素。這些不是 AI 被順帶提及的案例——裁員與 AI 之間的關聯是直接且有文件記錄的。
Tier 1: AI Cited as the Primary Reason
第一級：AI 被列為首要原因
| Company | Employees Cut | % of Workforce | Date | Key Statement |
|---|---|---|---|---|
| Block Inc. | ~4,000 | ~40% | Feb 2026 | CEO Jack Dorsey: "We no longer need the mass of people we once did" |
| Salesforce | ~4,000 (support) + ~1,000 | ~44% of CS unit | 2025 + Feb 2026 | CEO Marc Benioff: "I need less heads" — AI agents now handle 50% of customer interactions |
| Chegg | ~580+ combined | ~67% combined | May + Oct 2025 | Blamed "new realities of AI" — stock lost 99% of value due to ChatGPT |
| Klarna | ~2,100+ (over 2 years) | ~40% | 2023–2025 | CEO: AI replaced workload of ~700 CS employees (later admitted quality suffered, began rehiring) |
| IBM | Hundreds (HR) | Varies | 2023–2025 | CEO Arvind Krishna: AskHR AI automates 94% of routine HR tasks |
| Duolingo | Contractors only | ~10% of contractors | 2023–2025 | CEO: phasing out contractors for work AI can handle; launched 148 AI-written courses |
| 公司 | 裁員人數 | 佔員工比例 | 時間 | 關鍵聲明 |
|---|---|---|---|---|
| Block Inc. | 約 4,000 人 | 約 40% | 2026 年 2 月 | CEO Jack Dorsey：「我們不再需要過去那麼多人來維持速度」 |
| Salesforce | 約 4,000（客服）+ 約 1,000 | 客服部門約 44% | 2025 年 + 2026 年 2 月 | CEO Marc Benioff：「我需要更少的人頭」—— AI 代理已處理 50% 的客戶互動 |
| Chegg | 合計約 580+ 人 | 合計約 67% | 2025 年 5 月 + 10 月 | 歸咎「AI 的新現實」——因 ChatGPT 導致股價跌去 99% |
| Klarna | 約 2,100+ 人（2 年內） | 約 40% | 2023–2025 年 | CEO：AI 取代了約 700 名客服人員的工作量（後來承認品質下降，開始重新招人） |
| IBM | 數百人（HR 部門） | 不定 | 2023–2025 年 | CEO Arvind Krishna：AskHR AI 自動化 94% 的例行 HR 任務 |
| Duolingo | 僅約聘人員 | 約聘的 10% | 2023–2025 年 | CEO：逐步淘汰 AI 可以完成的約聘工作；推出 148 門 AI 編寫的課程 |
Tier 2: AI Cited as a Significant Factor (alongside restructuring)
第二級：AI 被列為重要因素（與組織重整並行）
| Company | Employees Cut | % of Workforce | Date | AI Connection |
|---|---|---|---|---|
| ~700 | 15% | Jan 2026 | Explicitly "reallocating resources to AI-focused roles" | |
| HP Inc. | 4,000–6,000 | ~10% | Nov 2025 (by 2028) | CEO cited AI adoption as driver; $1B annual savings target |
| Intuit | ~1,800 | ~10% | Jul 2024 | Cited AI transformation, investing in GenAI "Intuit Assist" (planned to rehire ~1,800 in AI roles) |
| UPS | ~60,000–90,000 | ~10%+ | 2024–2026 | CEO Carol Tome cited ML and automation; "Network of the Future" initiative |
| SAP | ~8,000–10,000 | ~7–9% | Jan 2024 | Pivoting to "Business AI"; €2B restructuring budget |
| Dow Inc. | ~4,500 | ~13% | Jan 2026 | "Transform to Outperform" initiative citing AI and automation |
| 公司 | 裁員人數 | 佔員工比例 | 時間 | AI 關聯 |
|---|---|---|---|---|
| 約 700 人 | 15% | 2026 年 1 月 | 明確表示「將資源重新分配到 AI 相關職位」 | |
| HP Inc. | 4,000–6,000 人 | 約 10% | 2025 年 11 月（至 2028 年） | CEO 將 AI 採用列為驅動因素；目標年省 10 億美元 |
| Intuit | 約 1,800 人 | 約 10% | 2024 年 7 月 | 以 AI 轉型為由，投資 GenAI「Intuit Assist」（計劃重新招聘約 1,800 名 AI 相關職位） |
| UPS | 約 60,000–90,000 人 | 約 10% 以上 | 2024–2026 年 | CEO Carol Tome 提及機器學習與自動化；「未來網路」計畫 |
| SAP | 約 8,000–10,000 人 | 約 7–9% | 2024 年 1 月 | 轉向「商業 AI」；20 億歐元重整預算 |
| Dow Inc. | 約 4,500 人 | 約 13% | 2026 年 1 月 | 「Transform to Outperform」計畫，明確提及 AI 與自動化 |
Tier 3: AI as Part of Broader Strategy (partially explicit)
第三級：AI 作為更廣泛策略的一部分（部分明確）
| Company | Employees Cut | % of Workforce | Date | AI Connection |
|---|---|---|---|---|
| Amazon | ~30,000 (corporate) | ~10% of corp. staff | Oct 2025 – Jan 2026 | CEO Andy Jassy expects AI to "reduce total corporate workforce"; also cited anti-bureaucracy push |
| Microsoft | ~15,000+ | ~7% | 2025 | Traditional sales roles replaced with "AI Sales Solutions"; coincides with $80B AI investment |
| Shopify | Multiple small rounds | Not quantified | 2024–2026 | CEO Tobi Lutke: teams must prove AI can't do the job before hiring; no single AI-labeled mass layoff |
| Meta | ~600 (AI unit) + ~1,500 (Reality Labs) | Varies | 2025–2026 | Restructuring around AI priorities rather than replacing with AI |
| 公司 | 裁員人數 | 佔員工比例 | 時間 | AI 關聯 |
|---|---|---|---|---|
| Amazon | 約 30,000（企業端） | 企業端約 10% | 2025 年 10 月 – 2026 年 1 月 | CEO Andy Jassy 預期 AI 將「減少企業總人力」；同時也提到反官僚化 |
| Microsoft | 約 15,000+ 人 | 約 7% | 2025 年 | 傳統銷售職位被「AI 銷售方案」取代；同期投資 800 億美元於 AI |
| Shopify | 多輪小規模裁員 | 未量化 | 2024–2026 年 | CEO Tobi Lutke：團隊必須證明 AI 無法完成工作才能招人；無單次大規模 AI 標籤裁員 |
| Meta | 約 600（AI 部門）+ 約 1,500（Reality Labs） | 不定 | 2025–2026 年 | 圍繞 AI 優先順序的重整，而非直接以 AI 取代人力 |
3. Job Categories Most Affected (Ranked High to Low)
3. 被裁員的職業種類排行（由高到低）
Based on documented layoff data from the companies above, the following job categories have been most impacted by AI-driven workforce reductions:
根據上述公司的裁員文件資料，以下職業類別受到 AI 驅動人力縮減的影響最大：
Rank 1: Customer Service / Support — MOST AFFECTED
The single clearest, most documented category. Klarna replaced ~700 customer service roles with an AI chatbot. Salesforce cut its support division from ~9,000 to ~5,000. Block's massive 40% cut reportedly hit customer operations heavily. AI chatbots can handle routine inquiries 24/7 at a fraction of the cost, and companies are betting on them despite early quality concerns.
Evidence strength: Very High — Multiple companies with explicit statements and numbers.
第 1 名：客戶服務 / 技術支援 —— 衝擊最大
這是目前最清晰、最有文件記錄的受影響類別。Klarna 用 AI 聊天機器人取代了約 700 名客服人員的工作量。Salesforce 將客服部門從約 9,000 人砍到約 5,000 人。Block 大規模的 40% 裁員據報也重創了客戶營運部門。AI 聊天機器人可以 24 小時處理例行諮詢，成本只是人類的一小部分，儘管早期品質有疑慮，企業仍然全力押注。
證據強度：非常高 —— 多家公司有明確聲明和具體數字。
Rank 2: Content Creation / Translation / Writing
Duolingo phased out contract translators and content writers, replacing them with AI that produces courses at dramatically higher speed. Media companies have quietly cut editorial staff. Shopify CEO noted AI sped up translations by 100x. This category spans copywriters, translators, technical writers, and content moderators.
Evidence strength: High — Duolingo is the clearest case; broader media trend is well-documented.
第 2 名：內容創作 / 翻譯 / 文案
Duolingo 逐步淘汰了約聘翻譯和內容撰寫人員，改用 AI 以大幅提高的速度產出課程。媒體公司悄悄裁減了編輯人力。Shopify CEO 指出 AI 將翻譯速度提升了 100 倍。此類別涵蓋文案撰稿人、翻譯人員、技術文件撰寫者和內容審核員。
證據強度：高 —— Duolingo 是最明確的案例；更廣泛的媒體業趨勢也有充分記錄。
Rank 3: Data Entry / Administrative / Back-Office
IBM's AskHR AI automates 94% of routine HR administrative tasks. UPS eliminated thousands of management and administrative roles. These roles — filling forms, processing invoices, scheduling, routine reporting — are the easiest to automate because they follow predictable, rule-based patterns.
Evidence strength: High — IBM HR automation is explicitly documented with metrics.
第 3 名：資料輸入 / 行政 / 後勤
IBM 的 AskHR AI 自動化了 94% 的例行 HR 行政任務。UPS 裁減了數千個管理和行政職位。這些職位——填寫表單、處理發票、排程、例行報表——最容易被自動化，因為它們遵循可預測的、規則性的模式。
證據強度：高 —— IBM HR 自動化有明確的指標記錄。
Rank 4: Middle Management
Google cut 35% of managers (specifically those leading teams of 3 or fewer). Amazon's "anti-bureaucracy push" eliminated management layers. UPS cut ~14,000 management roles in 2025. AI tools that aggregate reports, track metrics, and coordinate workflows reduce the need for coordination-focused middle managers.
Evidence strength: High — Google and Amazon management cuts are well-documented.
第 4 名：中階管理層
Google 裁減了 35% 的管理人員（特別是帶領 3 人或更少團隊的主管）。Amazon 的「反官僚化」推動消除了管理層級。UPS 在 2025 年裁減了約 14,000 個管理職位。AI 工具可以彙總報告、追蹤指標、協調工作流程，降低了以協調為主的中階管理者的需求。
證據強度：高 —— Google 和 Amazon 的管理層縮減有充分記錄。
Rank 5: Software QA / Junior Engineering
Google reported that over 25% of new code is now AI-generated (as of Q3 2024, likely higher now). Companies are shedding software testing and entry-level coding positions. Junior developers and QA engineers face significant pressure as AI code generation tools improve. However, senior engineering roles remain in high demand.
Evidence strength: Moderate-High — Google's metric is documented; broader junior dev displacement is widely reported but harder to quantify precisely.
第 5 名：軟體測試 / 初階工程師
Google 報告其超過 25% 的新程式碼現在由 AI 生成（截至 2024 年 Q3，目前可能更高）。公司正在縮減軟體測試和初階程式開發職位。隨著 AI 程式碼生成工具的進步，初階開發者和 QA 工程師面臨巨大壓力。然而，資深工程師的需求仍然很高。
證據強度：中高 —— Google 的指標有記錄；更廣泛的初階開發者取代現象被廣泛報導，但難以精確量化。
Rank 6: Sales (Traditional / Non-AI)
Microsoft replaced traditional sales roles with "AI Sales Solutions" positions. Salesforce's AI agents handle an increasing share of sales interactions. The shift is from high-volume, script-based sales to AI-assisted or AI-driven pipelines.
Evidence strength: Moderate-High — Microsoft's role restructuring is documented.
第 6 名：傳統銷售
Microsoft 將傳統銷售職位替換為「AI 銷售方案」職位。Salesforce 的 AI 代理處理越來越多的銷售互動。趨勢正從高量、腳本式銷售轉向 AI 輔助或 AI 驅動的銷售流程。
證據強度：中高 —— Microsoft 的職位重組有文件記錄。
Rank 7: HR / Recruiting
IBM replaced hundreds of HR workers with its AskHR AI system. Automated screening, resume parsing, and routine HR inquiries are increasingly AI-handled. This category is emerging but accelerating.
Evidence strength: Moderate-High — IBM is the primary explicit case.
第 7 名：人資 / 招募
IBM 用 AskHR AI 系統取代了數百名 HR 工作者。自動化篩選、履歷解析和例行 HR 諮詢越來越多由 AI 處理。這個類別正在興起並加速發展。
證據強度：中高 —— IBM 是主要的明確案例。
Rank 8: Financial Analysis / Research (Entry-Level)
Anthropic CEO Dario Amodei specifically cited law, consulting, and finance as sectors where entry-level roles face significant AI displacement. Junior analysts performing research, data aggregation, and report drafting are most vulnerable. However, concrete layoff data specifically citing AI in this category is still limited.
Evidence strength: Moderate — Based primarily on CEO predictions and early signals rather than mass layoff data.
第 8 名：金融分析 / 研究（初階）
Anthropic CEO Dario Amodei 特別點名法律、顧問和金融業是初階職位面臨顯著 AI 取代風險的領域。負責研究、資料彙總和報告撰寫的初階分析師最為脆弱。然而，在這個類別中明確以 AI 為由的大規模裁員數據仍然有限。
證據強度：中等 —— 主要基於 CEO 預測和早期訊號，而非大規模裁員數據。
Summary Table
| Rank | Job Category | Evidence | Key Companies |
|---|---|---|---|
| 1 | Customer Service / Support | Very High | Klarna, Salesforce, Block |
| 2 | Content / Translation / Writing | High | Duolingo, Shopify, media |
| 3 | Data Entry / Admin / Back-Office | High | IBM, UPS |
| 4 | Middle Management | High | Google, Amazon, UPS |
| 5 | Software QA / Junior Engineering | Moderate-High | Google, various tech |
| 6 | Traditional Sales | Moderate-High | Microsoft, Salesforce |
| 7 | HR / Recruiting | Moderate-High | IBM |
| 8 | Financial Analysis (Entry-Level) | Moderate | Predictions, early signals |
總覽表
| 排名 | 職業類別 | 證據強度 | 主要公司 |
|---|---|---|---|
| 1 | 客戶服務 / 技術支援 | 非常高 | Klarna、Salesforce、Block |
| 2 | 內容創作 / 翻譯 / 文案 | 高 | Duolingo、Shopify、媒體業 |
| 3 | 資料輸入 / 行政 / 後勤 | 高 | IBM、UPS |
| 4 | 中階管理層 | 高 | Google、Amazon、UPS |
| 5 | 軟體測試 / 初階工程師 | 中高 | Google、各科技公司 |
| 6 | 傳統銷售 | 中高 | Microsoft、Salesforce |
| 7 | 人資 / 招募 | 中高 | IBM |
| 8 | 金融分析（初階） | 中等 | 預測與早期訊號 |
4. In Their Own Words: CEO Statements That Shook the Industry
4. 他們自己說的：震動產業的 CEO 聲明
Jack Dorsey — Block CEO (February 2026)
In a 626-word memo posted publicly on X:
"A significantly smaller team, using the tools we're building, can do more and do it better. And intelligence tool capabilities are compounding faster every week."
"I think most companies are late. Within the next year, I believe the majority of companies will reach the same conclusion and make similar structural changes. I'd rather get there honestly and on our own terms than be forced into it reactively."
Dorsey's memo was notable for its total lack of corporate euphemism. No "restructuring for growth." No "difficult decision." Just: AI can do the work, so we don't need 40% of our people.
Jack Dorsey —— Block 執行長（2026 年 2 月）
在 X 上公開發佈的 626 字備忘錄中：
「一個明顯更小的團隊，使用我們正在打造的工具，可以做更多、做更好。而智慧工具的能力每週都在複合式成長。」
「我認為大多數公司都太晚了。在未來一年內，我相信大多數公司都會得出相同的結論，做出類似的結構性變革。我寧願誠實地、按照我們自己的節奏到達那裡，而不是被迫做出反應。」
Dorsey 的備忘錄值得注意的是他完全沒有使用企業委婉語。沒有「為成長而重組」，沒有「艱難的決定」。就是直接說：AI 能做這些工作，所以我們不需要 40% 的人了。
Marc Benioff — Salesforce CEO (September 2025)
"I've reduced it from 9,000 heads to about 5,000, because I need less heads."
Benioff stated this matter-of-factly, noting that Salesforce's AI agents now handle 50% of customer interactions. The bluntness was remarkable — "I need less heads" became one of the most-quoted CEO statements of 2025.
Marc Benioff —— Salesforce 執行長（2025 年 9 月）
「我已經從 9,000 個人頭減少到大約 5,000 個，因為我需要更少的人頭。」
Benioff 語氣平淡地陳述了這個事實，指出 Salesforce 的 AI 代理現在處理 50% 的客戶互動。這種直白令人印象深刻——「我需要更少的人頭」成為 2025 年被引用最多的 CEO 聲明之一。
Sebastian Siemiatkowski — Klarna CEO (October 2025)
"I feel a lot of my tech bros are being slightly not to the point on this topic."
"To me, AI is capable of doing all our jobs, my own included."
"Yes, new jobs will be created, but in the shorter term, that doesn't help the Brussels translator. He's not going to become a YouTube influencer tomorrow."
Siemiatkowski was one of the first CEOs to accuse fellow tech leaders of "sugarcoating" AI's impact on jobs. He later warned that mass unemployment was a real possibility. Ironically, Klarna later admitted that aggressive AI-driven cuts had degraded service quality, and began rehiring humans.
Sebastian Siemiatkowski —— Klarna 執行長（2025 年 10 月）
「我覺得很多科技圈的兄弟們在這個話題上有點避重就輕。」
「對我來說，AI 有能力做我們所有人的工作，包括我自己的。」
「是的，新工作會被創造出來，但在短期內，這幫不了布魯塞爾的翻譯員。他不可能明天就變成 YouTube 網紅。」
Siemiatkowski 是最早指控其他科技領袖「粉飾」AI 對就業影響的 CEO 之一。他後來警告大規模失業是一個真實的可能性。諷刺的是，Klarna 後來承認激進的 AI 裁員已經降低了服務品質，並開始重新招聘人類員工。
Tobi Lutke — Shopify CEO (April 2025)
"Reflexive AI usage is now a baseline expectation at Shopify."
Teams "must demonstrate why they cannot get what they want done using AI" before asking for more employees.
Lutke didn't announce a mass layoff. He did something potentially more impactful: he changed the default. The burden of proof shifted from "justify why AI should do this" to "justify why a human should."
Tobi Lutke —— Shopify 執行長（2025 年 4 月）
「反射性地使用 AI 現在是 Shopify 的基本期望。」
團隊「必須證明為什麼他們無法使用 AI 完成想做的事」，才能要求更多員工。
Lutke 沒有宣布大規模裁員。他做了一件可能影響更深遠的事：改變了預設值。舉證責任從「證明為什麼 AI 應該做這件事」轉變為「證明為什麼人類應該做這件事」。
Arvind Krishna — IBM CEO (2025)
AI could "easily see 30% of workforce getting replaced" and confirmed AI replaced hundreds of HR employees.
But Krishna also noted that IBM's total headcount actually grew because savings from HR automation were reinvested in engineering and sales hiring. This is the "displacement, not destruction" argument — AI eliminates some roles but creates others.
Arvind Krishna —— IBM 執行長（2025 年）
AI 可以「輕易看到 30% 的勞動力被取代」，並確認 AI 已取代了數百名 HR 員工。
但 Krishna 也指出，IBM 的總人數實際上增加了，因為 HR 自動化節省的成本被重新投資於工程和銷售招聘。這就是「取代而非毀滅」的論點——AI 消除了某些職位，但創造了其他職位。
Dario Amodei — Anthropic CEO (May 2025)
"We, as the producers of this technology, have a duty and an obligation to be honest about what is coming."
Amodei predicted AI could eliminate 50% of entry-level white-collar jobs within 1–5 years and warned unemployment could spike to 10–20%. As the CEO of one of the leading AI companies, his candor was both praised and criticized.
Dario Amodei —— Anthropic 執行長（2025 年 5 月）
「我們作為這項技術的生產者，有責任和義務對即將到來的事情保持誠實。」
Amodei 預測 AI 可能在 1–5 年內消除 50% 的初階白領工作，並警告失業率可能飆升到 10–20%。作為領先 AI 公司之一的 CEO，他的坦誠既受到讚揚也遭到批評。
The Contrarian: Matt Garman — AWS CEO (December 2025)
Replacing junior employees with AI is "one of the dumbest ideas I've ever heard."
Garman argued that junior roles are essential for building the talent pipeline — you can't have senior engineers without giving juniors the chance to learn and grow. This was a notable pushback against the prevailing narrative.
反對者：Matt Garman —— AWS 執行長（2025 年 12 月）
用 AI 取代初階員工是「我聽過最蠢的主意之一」。
Garman 認為初階職位對建立人才管道至關重要——如果不給初階人員學習和成長的機會，就不可能有資深工程師。這是對主流論述的一個值得注意的反駁。
5. The Reality Check: Are Companies Actually Replacing Workers, or Just Cutting Costs?
5. 現實查核：企業真的在用 AI 取代員工，還是只是藉口砍成本？
It would be irresponsible to present the AI layoff narrative without noting the significant skepticism from economists and researchers:
The "AI as pretext" argument: According to Oxford Economics, most employers don't appear to be replacing significant numbers of workers with AI. Some economists believe companies may be using AI as a cover story for standard cost-cutting to please Wall Street. As one researcher noted: decision-makers "just need to cut costs and make it happen. That's it."
Potential vs. Performance: A Harvard Business Review analysis found that 60% of organizations reduced headcount in anticipation of AI's future impact, while just 2% said they've made large layoffs tied to actual AI implementation. In other words, companies are firing people based on what AI might do, not what it's currently doing.
The Klarna Cautionary Tale: Klarna is perhaps the most instructive case. CEO Siemiatkowski proudly announced that AI had replaced 700 customer service roles — only for the company to later admit that quality had suffered and it needed to rehire humans. This suggests the gap between AI's theoretical capability and real-world reliability is still significant.
Rehiring patterns: Multiple companies (Klarna, IBM, Intuit) have announced that while they cut certain roles, they're hiring in other areas — often at similar headcount levels. The net effect may be more of a rotation than a reduction.
如果只呈現 AI 裁員的敘事而不提及經濟學家和研究人員的重大質疑，那是不負責任的：
「AI 只是藉口」的論點： 根據 Oxford Economics 的研究，大多數僱主似乎並沒有用 AI 取代大量勞工。一些經濟學家認為，企業可能只是把 AI 當作標準降本的幌子，以取悅華爾街。正如一位研究人員指出的：決策者「只需要降低成本然後執行。就這樣。」
潛力 vs. 實際表現： Harvard Business Review 的分析發現，60% 的組織是基於對 AI 未來影響的預期而裁員，而只有 2% 表示他們是因為實際的 AI 實施而進行大規模裁員。換句話說，公司是根據 AI 可能做什麼來裁人，而不是根據它正在做什麼。
Klarna 的警示故事： Klarna 可能是最具啟發性的案例。CEO Siemiatkowski 驕傲地宣布 AI 已取代了 700 個客服職位——結果公司後來承認品質下降了，需要重新招聘人類。這說明 AI 的理論能力與實際可靠性之間的差距仍然很大。
重新招聘模式： 多家公司（Klarna、IBM、Intuit）在宣布裁減某些職位的同時，也在其他領域招聘——通常是相似的人數規模。淨效果可能更像是一種輪替而非縮減。
6. What This Means for Workers: A Pragmatic Assessment
6. 這對工作者意味著什麼：務實的評估
The bad news is real
- Routine, predictable work is genuinely at risk. If your job involves answering the same 50 types of questions, filling out forms, or translating standard documents, AI can already do this — imperfectly, but cheaply.
- The "prove a human should do this" standard is spreading. Shopify's policy may become the norm. The default is shifting.
- Entry-level positions face the most pressure. Ironically, the roles that train people for senior positions are the ones being cut first.
But the nuance matters
- AI is not replacing senior, judgment-heavy roles — yet. Complex decision-making, creative strategy, and novel problem-solving remain firmly human.
- Companies that cut too aggressively are paying the price. Klarna's quality degradation and rehiring is a warning sign.
- The 60/2 gap is telling. Most companies are cutting based on AI's potential, not its proven capabilities. This suggests some of these cuts may be premature.
- New roles are emerging. AI operations, prompt engineering, AI safety, model evaluation, and human-AI workflow design are growing fields.
The actionable takeaway
The safest position is not to panic, but to be realistic: learn to work with AI, and shift toward roles that require judgment, creativity, and human connection — the areas where AI still falls short, and where it may remain limited for years to come.
壞消息是真實的
- 例行性、可預測的工作確實面臨風險。 如果你的工作是回答同樣的 50 種問題、填寫表單、或翻譯標準文件，AI 已經能做到——雖然不完美，但便宜得多。
- 「證明人類應該做這件事」的標準正在擴散。 Shopify 的政策可能成為常態。預設值正在改變。
- 初階職位面臨最大壓力。 諷刺的是，訓練人們成為資深人才的職位，反而是最先被砍的。
但細微差別很重要
- AI 還沒有取代需要高度判斷力的資深職位。 複雜決策、創意策略和新穎問題解決仍然牢牢屬於人類。
- 裁員太激進的公司正在付出代價。 Klarna 的品質下降和重新招聘就是一個警訊。
- 60/2 的差距很有說服力。 大多數公司是基於 AI 的潛力而非已證實的能力來裁員。這說明部分裁員可能是過早的。
- 新職位正在出現。 AI 運維、提示工程、AI 安全、模型評估和人機工作流設計都是成長中的領域。
可行動的結論
最安全的立場不是恐慌，而是保持務實：學會與 AI 合作，並轉向需要判斷力、創造力和人際連結的職位——這些是 AI 仍然不足的領域，也可能在未來幾年持續受限的領域。
7. Timeline: The AI Layoff Wave at a Glance
7. 時間線：AI 裁員潮一覽
| Date | Company | Event |
|---|---|---|
| Jan 2024 | SAP | ~8,000–10,000 affected in AI pivot restructuring |
| Jan 2024 | UPS | ~12,000 management layoffs citing ML/automation |
| Jul 2024 | Intuit | ~1,800 cut for AI transformation (rehiring in AI roles) |
| End 2023 – Apr 2025 | Duolingo | Contract translators/writers phased out for AI |
| Apr 2025 | Shopify | CEO mandates "prove AI can't do it" before hiring |
| May 2025 | Chegg | 22% workforce cut — "new realities of AI" |
| May 2025 | Anthropic CEO | Warns 50% of entry-level white-collar jobs at risk |
| Sep 2025 | Salesforce | CEO confirms ~4,000 CS cuts — "I need less heads" |
| Oct 2025 | Chegg | Additional 45% cut — combined 67% total |
| Oct 2025 | Klarna CEO | Accuses tech leaders of "sugarcoating" AI job impact |
| Nov 2025 | HP Inc. | 4,000–6,000 to be cut by 2028 citing AI adoption |
| 2025 | Microsoft | ~15,000+ cut; sales roles replaced with AI solutions |
| Jan 2026 | Amazon | ~16,000 corporate layoffs (AI + anti-bureaucracy) |
| Jan 2026 | ~700 cut — resources reallocated to AI teams | |
| Jan 2026 | Dow Inc. | ~4,500 cut under AI/automation initiative |
| Feb 2026 | Block Inc. | ~4,000 cut (40%) — Dorsey's viral AI memo |
| Feb 2026 | Klarna | Admits AI quality issues, begins rehiring humans |
| 時間 | 公司 | 事件 |
|---|---|---|
| 2024 年 1 月 | SAP | 約 8,000–10,000 人受 AI 轉型重組影響 |
| 2024 年 1 月 | UPS | 約 12,000 名管理層裁員，提及 ML / 自動化 |
| 2024 年 7 月 | Intuit | 約 1,800 人因 AI 轉型被裁（同時招聘 AI 職位） |
| 2023 年底 – 2025 年 4 月 | Duolingo | 約聘翻譯 / 撰稿人員被 AI 逐步取代 |
| 2025 年 4 月 | Shopify | CEO 要求「先證明 AI 做不到」才能招人 |
| 2025 年 5 月 | Chegg | 裁減 22% 人力——「AI 的新現實」 |
| 2025 年 5 月 | Anthropic CEO | 警告 50% 初階白領工作面臨風險 |
| 2025 年 9 月 | Salesforce | CEO 確認裁減約 4,000 名客服——「我需要更少人頭」 |
| 2025 年 10 月 | Chegg | 再裁 45%——合計 67% |
| 2025 年 10 月 | Klarna CEO | 指控科技領袖「粉飾」AI 對就業的影響 |
| 2025 年 11 月 | HP Inc. | 至 2028 年裁減 4,000–6,000 人，提及 AI 採用 |
| 2025 年 | Microsoft | 裁減超過 15,000 人；銷售職位被 AI 方案取代 |
| 2026 年 1 月 | Amazon | 約 16,000 名企業端裁員（AI + 反官僚化） |
| 2026 年 1 月 | 裁減約 700 人——資源重新分配至 AI 團隊 | |
| 2026 年 1 月 | Dow Inc. | 約 4,500 人受 AI / 自動化計畫影響 |
| 2026 年 2 月 | Block Inc. | 裁減約 4,000 人（40%）——Dorsey 病毒式傳播的 AI 備忘錄 |
| 2026 年 2 月 | Klarna | 承認 AI 品質問題，開始重新招聘人類 |
Confidence Disclosure
This article applies a strict confidence threshold. Every company listed in Tier 1 and Tier 2 has been verified through multiple major news sources (CNBC, CNN, Fortune, TechCrunch, etc.) with direct CEO quotes or official company statements explicitly linking layoffs to AI. Companies in Tier 3 have weaker or more indirect AI attribution, which is noted accordingly.
The job category rankings are based on aggregated evidence from documented layoffs, not projections. Categories with lower evidence strength are explicitly marked.
Where information is uncertain or contested (such as the "AI as pretext" debate), both perspectives are presented.
信心度揭露
本文採用嚴格的信心度門檻。第一級和第二級列出的每家公司都經過多個主要新聞來源（CNBC、CNN、Fortune、TechCrunch 等）的交叉驗證，包含 CEO 直接引述或公司官方聲明明確將裁員與 AI 連結。第三級的公司 AI 歸因較弱或較間接，已在文中相應標註。
職業類別排名基於已記錄裁員的彙總證據，而非預測。證據強度較低的類別已明確標示。
在資訊不確定或有爭議的地方（例如「AI 只是藉口」的辯論），兩方觀點都有呈現。