Your field-guide to AI — what it means for your job and what to do about it
AI is changing work right now. Not eventually. Now. This page gives you the facts – where things stand, who is affected, and what the common objections miss. Every claim links to its source.
By The Numbers
What Can AI Actually Do Right Now?
In plain English: AI tools can write emails and reports (not perfect, but a solid first draft), code software from plain English instructions, analyze spreadsheets just by asking questions, summarize a 200-page PDF in seconds, create images from text descriptions, and control your computer – navigating apps, filling forms, and executing multi-step workflows.
The major tools are ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google), and Microsoft Copilot (built into Office). AI is no longer expensive or experimental. It is cheap, fast, and being embedded into the tools you already use – whether you realize it or not.
More detail: adoption stats and business use
- 78% of organizations now use AI in at least one function – up 23 points in a single year (Stanford HAI)
- Worker access to AI rose 50% in 2025 alone (Deloitte)
- OpenAI revenue tripled to $20+ billion; Anthropic grew from $87M to $7 billion
- 52% of AI usage is augmentation (human + AI together), 45% is automation (AI alone) – Anthropic Economic Index
Who Is Most Affected?
One in four workers globally are in occupations with some degree of AI exposure. Most jobs will be transformed, not eliminated -- but the transformation is not optional.
Already seeing decline in job postings (Bloomberry, 180M postings analyzed):
- Graphic artists: -33% | Writers/copywriters: -28% | Photographers: -28%
- Journalists: -22% | Medical scribes: -20% | Compliance specialists: -29%
Entry-level workers are hit hardest. Postings dropped 35% since 2023. Workers age 22-25 saw a 14% drop in job-finding rates in AI-exposed fields.
6.1 million U.S. workers face both high exposure and low ability to adapt. 86% of them are women, concentrated in clerical and admin roles.
More detail: company layoffs citing AI
Company-level layoffs directly citing AI include Block (cut nearly in half), Salesforce (~5,000 support roles), Klarna (~700 service staff), and Microsoft (~6,000 programmers – CEO noted 30% of code was AI-written). U.S. programmer employment fell 27.5% between 2023-2025. However, “AI washing” is real – some companies blame AI for cuts driven by other factors. Oxford Economics and Sam Altman himself have flagged this.
More detail: gender, education, and who's protected
- 79% of employed U.S. women hold positions categorized as high-risk vs. 58% of men
- Paradoxically, the most exposed workers are among the most educated and highest-paid – 17.4% hold graduate degrees
- Manual, in-person, and blue-collar jobs still face less exposure unless robotics advance significantly
- Healthcare is a “change, not replacement” area – most health roles are more likely to benefit than to be displaced
What’s Coming (2027-2028)
The biggest shift: AI moves from a tool you ask questions to an agent that takes actions. Instead of “draft me an email,” you’ll say “manage my calendar this week – reschedule conflicts, send apologies, and brief me each morning.” The AI actually does it. (MIT Sloan explainer)
Growing fastest (WEF): Big Data Specialists (+113%), AI/ML Specialists (+82%), Software Developers, Security Specialists. AI has already created 1.3 million new roles.
Declining fastest (WEF): Postal clerks, bank tellers, data entry clerks, cashiers, admin assistants, bookkeeping clerks, graphic designers.
More detail: economic projections from Goldman, McKinsey, Gartner
- Goldman Sachs: AI will have measurable GDP impact starting 2027, automating ~25% of tasks in advanced economies
- McKinsey: GenAI could unlock $2.6-$4.4 trillion in annual value across customer ops, marketing, engineering, and R&D
- Gartner: 40% of enterprise apps will include AI agents by end of 2026; but 40%+ of agentic projects will be canceled by 2027 due to cost/unclear value
- An NBER survey of 6,000 executives found 90% report no impact yet, but project 1.4% productivity boost in the next 3 years. The gap between “what has happened” and “what is expected” is closing fast.
Common Objections
“AI slows productivity”
Some studies found exactly that. The Upwork study found 77% of employees said AI added to their workload. A METR study found experienced developers were 19% slower with AI. But a Harvard/BCG study found AI users did 25% more tasks, 25% faster, at 40% higher quality – when the task was in AI’s wheelhouse. A Stanford/MIT study found a 14% boost for customer service agents (35% for novices).
AI does not automatically make you faster. It makes you faster at the right tasks when you know how to use it. The gains are real -- but only with training and workflow redesign.
“AI will take all jobs”
Goldman Sachs estimates 6-7% of U.S. workers (~11 million) face long-term displacement. The WEF projects a net gain of 78 million jobs by 2030. But the jobs being created are not the same jobs being destroyed – different skills, different pay, different locations. The losers are disproportionately younger and less experienced.
“AI makes too many mistakes”
Hallucination rates for routine tasks have dropped from 21.8% to 0.7% in four years. But for complex reasoning, top models still hallucinate 33-51% of the time. Rule of thumb: use AI for drafts, verify anything important.
“It’s just hype”
OpenAI went from $200M to $20B revenue in two years. 88% of organizations use it. But 95% of organizations saw zero measurable return on AI investment. The most accurate framing: a real technological revolution wrapped in a hype bubble. The internet analogy fits – the dot-com bubble burst, but the internet was not hype.
“I tried it and it wasn’t useful”
Most people type something vague and get generic output. Effective use requires clearly stating what you want – who it’s for, what tone, what length, what purpose. If you tried it in 2023 and gave up, you tried a flip phone in the smartphone era. The tools have improved enormously.
What To Do About It
The single most important thing you can do is understand how AI affects your specific work – not AI in general, but the tools, the timeline, and the skills that matter for what you actually do every day.