Part 1 · Know Your AI Readiness
Ethics · Knowing the Risks
Unfair treatment is one risk among several. You paste a confident, well-written summary into a board paper. It reads perfectly. The single number that matters is invented, and the AI has no idea.
What can go wrong
An AI language model does one thing well: it predicts the most likely next words from patterns in its training data. It does not know what is true. Hold on to that one fact, because almost every risk you will meet flows straight from it.
- Bias: it learns from human data, so it can repeat human prejudice in decisions like hiring, lending, or who gets flagged for review.
- Mistakes: it can be fluently, confidently wrong, with no signal on its face that it is.
- Made-up answers: it can invent names, quotes, figures, and citations that never existed. This is the hallucination you met in the first Mindset lesson.
- Privacy: whatever you paste into a public AI tool may be stored and reused, so confidential or personal data does not belong there.
- Acting for you: hand a task to an AI agent that can send, book, or buy, and a wrong assumption executes before you can check it.
- It predicts without knowing: good at sounding right, blind to being right. This is where mistakes and made-up answers begin.
- It learns from us: trained on human data, so it can inherit and repeat human bias.
- It can act before you check: an agent with permission to act turns a wrong assumption into a sent email or a booked slot.
To reach the AI-Ready bar here, name the risks that apply to your own work, and know your one check for each.
Key Takeaway
Fluency is not accuracy. The more confident an AI sounds, the more deliberately you check it.
Check your understanding
Self-check only. Nothing is scored and nothing is recorded; open an answer to compare it with your own.
What makes risk from an AI agent different from risk from a chatbot that just answers questions?
- There is no real difference, both just produce text.
- An agent can take actions on your behalf, so a mistake can change the wrong file or share something it should not, not just give you a wrong answer to ignore.
- Agents are always safer because they follow instructions exactly.
- Chatbots are riskier because agents cannot make mistakes.
Show the answer
An agent can take actions on your behalf, so a mistake can change the wrong file or share something it should not, not just give you a wrong answer to ignore.
A colleague says, "I know AI has risks, but I could not really tell you what they are." Where are they, and what is the AI-Ready move?
- They are AI Ready, acknowledging risk exists is enough.
- They are below AI Ready; the move is to be able to name three or more concrete risks (like made-up answers, privacy, mistakes, and agents acting wrongly) and how each applies to their work.
- They are above AI Ready, staying humble about risk is the advanced position.
- It does not matter, since the tools handle their own safety.
Show the answer
They are below AI Ready; the move is to be able to name three or more concrete risks (like made-up answers, privacy, mistakes, and agents acting wrongly) and how each applies to their work.
Saved in this browser only. Carry progress between devices with the resume link on the course page.