Are you against using AI with junior creatives?

No. I use AI in my own work, led by judgment, and I want my team fluent in it. What I am against is handing a junior the tool as a substitute for the thinking. Use AI to skip the drudgery that teaches nothing, then put the person on the decision that does. The tool should raise the floor of what they work on, not remove the work.

Doesn't automating repetitive tasks free people for better work?

It can, but only if you actually redirect the reclaimed time into harder thinking. The D&AD report found many shops bank the efficiency and never reinvest it. Freed time that goes nowhere is just a smaller team doing thinner work. The time dividend only pays off if a leader spends it on purpose.

How do you train judgment if AI does the grunt work?

You give people real decisions with real consequences early, and you let them be wrong where it is safe. Judgment is not a lecture. It is reps on work that ships, with someone experienced reviewing the calls. I would rather a junior own a small honest piece end to end than assist on a big one they never had to decide.

AI Took the Work That Trained Us by Tyler Garner

There is a new report making the rounds this week, and it names something I have been chewing on for a while. D&AD published its AI and Creativity Report, built on interviews with 197 creative leaders, and one finding cuts deeper than the rest. The entry-level work is disappearing. The repetitive, unglamorous tasks that AI now handles in an afternoon were the exact work that used to train the next generation of creative leaders. Kill that work and you do not just lose a few jobs. You lose the pipeline that makes anyone worth promoting.

What did the D&AD report actually find?

It found that the creative industry is spending a shortcut it does not fully understand. Across the leaders they talked to, AI is trending toward average output, junior roles are vanishing, and most shops are pocketing the speed without reinvesting it in deeper thinking. The CEO, David Patton, framed the whole thing as a choice: value judgment over shortcuts, or lose the rigor that separates competent from exceptional.

I read that and nodded. The report is not anti-AI. Neither am I. It is a warning about what you quietly give up when efficiency becomes the only thing you measure. The cheapest thing to automate is often the thing that was teaching someone.

Why does entry-level work matter if AI can do it faster?

Because that work was never only about the output. It was about the reps. I spent years earlier in my career on tasks a model could now do in minutes: resizing, cleaning up files, cutting rough assemblies, laying out the version nobody would ever see. Boring, yes. But every one of those passes was a small lesson in why a thing works or does not. You do not learn where the line should go by reading about it. You learn it by putting the line in the wrong place a few hundred times and feeling the difference.

Strip that out and you hand a junior the answer without the question. They can operate the tool. They cannot yet judge what it gives back, because judgment is the one thing the tool cannot hand them.

Can you train judgment without the boring work?

Not by watching. Judgment is trained, not born, and it is trained on real work with real stakes. You can move a junior past the drudgery that teaches nothing, the true busywork, and you should. But you cannot skip the part where they own a decision and live with the result. That is the whole mechanism. Someone makes a call, someone experienced reviews it, and the gap between the two becomes next week's instinct.

The trap in the report is subtle. If AI does the junior work, the junior never makes the small calls, so they never build the instinct, so there is no senior five years out. The pipeline does not collapse in a quarter. It collapses in a promotion cycle nobody sees coming.

What does this mean for how I build a team?

I keep my younger people on real decisions on purpose. Not assisting on a decision, owning one. A small piece, honest and complete, that ships with their name on the thinking. I would rather someone sweat a modest choice they had to defend than clean up a big one they never had to make. We use AI heavily. It handles the passes that were only ever labor. Then the person spends the time it saved on the part that is actually hard, which is deciding whether the work is any good and whether it tells the truth.

In attainable housing that standard is not optional. The work has to be right, not just fast. A render, a floor plan callout, a leasing email all have to survive a real person arriving at a real building. A junior who has only ever approved what a model returned has not learned to catch the gap between the picture and the place. A junior who has been made to decide has.

What should a creative leader do about it now?

Spend the time dividend on purpose. The efficiency is real, so decide out loud where the reclaimed hours go, or they will quietly turn into a thinner team doing shallower work. Put juniors on decisions early and review the decisions, not just the deliverables. Treat AI as the thing that raises the floor of what a young creative works on, not the thing that removes the work. And be honest that this costs you something now. Training people is slower than letting a model finish the file. It is also the only way you have a shop worth running in five years.

The shortcut is real and it is tempting. You can read the case in the D&AD report. I just think a leader who automates away the work that makes leaders is not being efficient. They are eating the seed corn and calling it a good year.