Part 1 · Know Your AI Readiness
Mindset · What AI Can and Cannot Do
You ask an AI tool to summarise a forty-page report. Ten seconds later a clean summary appears, better than you expected. Encouraged, you ask it for the exact figure on page thirty-one. It gives you a number, stated with total confidence. You open the page to check. The number is not there.
Prediction, not judgement
The Mindset pillar, the first cluster of Part 1, opens with the most basic question in AI readiness: what can these tools do, and where do they fall down? Today's AI runs on prediction. Give it a pattern (a well-formed email, the structure of a summary, the next likely word) and it reproduces that pattern at speed, at almost no cost. That is why drafting, summarising, and translating feel effortless when you hand them over.
What AI does not have is judgement. It cannot weigh what matters, decide what should exist, or reliably tell you when it is wrong. That confident wrong figure has a name: a hallucination. It is fluent output with no grounding in fact. The tool is not lying. It predicts what an answer looks like. Whether the answer is true is not part of that job.
The gap is measurable. In an October 2025 study by Scale AI and the Center for AI Safety, the Remote Labor Index, the best AI agent tested completed only two and a half percent of real-world freelance tasks from start to finish. Strong at producing a draft; nowhere near owning the work around it.
| AI does reliably | Still needs you |
|---|---|
| Draft and summarise text | Decide what matters |
| Translate and rephrase | Verify facts and figures |
| Surface patterns in data | Apply domain expertise |
| Suggest options quickly | Own the final call |
Being AI Ready here means looking at a task in your own field and saying, with reasons, whether AI can do it reliably, do it as a first draft you must check, or not touch at all.
Key Takeaway
Certainty in an AI's wording tells you nothing about whether it is right. This week, catch yourself before acting on an AI output and ask how you would know if it were wrong.
Check your understanding
Self-check only. Nothing is scored and nothing is recorded; open an answer to compare it with your own.
A colleague says, "AI can handle our quarterly report end to end now. I will just let it run." Based on an AI-Ready understanding of what today's tools reliably do, what is the most accurate response?
- Agree, modern AI handles complete reports without supervision.
- AI can accelerate parts of it (drafting, summarising, formatting) but it needs a person to check the facts and figures, because it confidently makes things up.
- Disagree. AI cannot be trusted with reports at all, so do it manually.
- It depends entirely on which brand of AI you use.
Show the answer
AI can accelerate parts of it (drafting, summarising, formatting) but it needs a person to check the facts and figures, because it confidently makes things up.
Which of these best demonstrates someone has reached AI Ready (L2) on understanding AI?
- They can describe the maths behind how a neural network is trained.
- They use AI tools every day without thinking about it.
- They can explain to a teammate, in plain language, which tasks in their field AI handles well and which it does not, and name how it commonly goes wrong.
- They believe AI will eventually be able to do everything.
Show the answer
They can explain to a teammate, in plain language, which tasks in their field AI handles well and which it does not, and name how it commonly goes wrong.
Saved in this browser only. Carry progress between devices with the resume link on the course page.