Julia Heuritsch on AI, ownership, and working with ambiguity

Which AI hype do you currently find the most overrated?

Julia Heuritsch:
“The idea that AI will replace human thinking or simply do the work for us.

That framing confuses output with responsibility and mistakes acceleration for autonomy. AI is not a substitute for thinking; it accelerates cognition and iteration. It shortens the distance between hypothesis, refutation, and revision.

AI can generate content, but it cannot own intent, values, or consequences. When organisations chase replacement instead of ownership, they don’t become smarter — they amplify existing confusion, faster and at scale.”

What’s one moment when AI truly surprised you?

Julia:
“I was surprised when I realised AI works best not as a productivity hack, but as a mirror.

Whenever my intent was vague or my assumptions were shaky, the output fell apart immediately. When my thinking was clear and grounded, the quality improved dramatically.

AI doesn’t create clarity — it reveals whether clarity is already there. It surfaces flaws early and forces earlier contact with reality.”

Which decisions should never be handed over to AI?

Julia:
“Anything involving moral responsibility and relational impact — especially decisions affecting dignity, trust, or safety.

AI can inform, simulate, and expose risks, but it cannot own decisions or consequences. Responsibility needs a human face.

Delegation ends where accountability begins.”

What do people need to unlearn to work better with AI?

Julia:
“Three things.

First, the belief that effort equals value.
Output becoming easier doesn’t make it meaningless, but it requires new markers of quality: intent, coherence, and accountability.

Second, the idea that AI is something to be ashamed of or something to compete with.
Tools don’t erode integrity; unconscious use does.

Third, the need to be right.
AI often confirms your thinking, but just as frequently presents multiple plausible perspectives. When we push for absolute answers, AI surfaces context, trade-offs, and alternative framings.

What becomes visible is not who is right, but who can tolerate ambiguity and make a conscious choice within it. Many people resist AI not because it is wrong, but because it removes the comfort of certainty and demands ownership over judgment rather than correctness.”

If you could introduce one AI rule in the workplace, what would it be?

Julia:
“You may use AI for anything — as long as you can clearly explain your intent, your choices, and are willing to stand behind the outcome.

No hiding behind ‘the system said so.’ AI may accelerate cognition and iteration, but ownership remains human.”