Anthropic, the company behind Claude, studied how people actually work with AI and gave the difference a name: the 4D AI Fluency framework. The good news for anyone who feels behind is that not one of these four skills is technical. You don't need to code. You don't need to understand how the models work. These are judgment skills, which means anyone willing to practice can build them.

Here's each one in plain English, with a simple way to practice it this week.

1. Delegation — knowing what to hand over

The first skill is deciding what to give AI and what to keep for yourself. Most people get this wrong in one of two directions: they either do everything themselves and never delegate, or they hand over things that need their own judgment. The skill is knowing where the line is.

Practice this week

Before your next task, ask AI: "Here's what I need to get done: [describe it]. Break this into the parts you could handle well, the parts I should do myself, and the parts we should do together. Explain why for each."

2. Description — telling it clearly what you want

AI can only be as good as what you ask for. Description is the skill of giving it the context, the goal, and a picture of what a good answer looks like — instead of a one-line question and a shrug. Better input, better everything.

Practice this week

Take a request you'd normally type in one line and expand it: "Here's the context: [background]. Here's my goal: [outcome]. Here's what a great answer looks like: [format, length, tone]. Here's what to avoid: [pitfalls]." Notice the jump in quality.

3. Discernment — judging whether the answer is any good

AI sounds confident even when it's wrong. Discernment is your ability to look at what it gives you and tell whether it's actually right, useful, and true. Your judgment is the filter between a good-sounding answer and a good answer.

Practice this week

After AI gives you something, don't just accept it. Ask: "What are the weakest parts of this answer? What might be wrong or missing? What would an expert push back on?" Then check those points yourself.

4. Diligence — owning what you do with it

The last skill is responsibility. Checking the facts, standing behind the work, using AI honestly and safely. This is what separates a professional from someone just pasting whatever came out and hoping it's fine.

Practice this week

Before you send or publish anything AI helped with, run a 30-second check: Are the facts and numbers verified? Would I be comfortable if someone knew AI helped? Is this genuinely mine now, or just copied? If yes to all three, ship it.

The real takeaway

You're not behind because you're not a "tech person." Every one of these is a judgment skill you already use in the rest of your life — you're just pointing it at AI now. You're only behind if you decide not to start. Pick one of the four above and practice it this week. That's how fluency actually gets built: one rep at a time.