Matthias Karner

Digitalization & IT leader. Avid reader. Engineering PhD. Productivity enthusiast.

3 minute read

Are you using AI to boost your productivity?

I am. And honestly, it’s become part of my daily routine.

What I’ve noticed is that AI doesn’t always respond the same way. Over the past few months, terms like “prompt engineering” have started popping up everywhere. Personally, I think “engineering” sounds a bit over the top, but there’s no denying that how you talk to AI makes a difference.

image
Photo by Peyton Clough on Unsplash

I’ve experimented with a few simple tricks along the way.

The one that’s helped me the most?

Giving the AI a clear role. Asking it to act like an architect, a mentor, or a coach changes the quality of the answers. It’s giving it context before starting the conversation.

Recently, I came across a list of even more unconventional techniques and some of them really stuck with me. One was boldly titled “gaslighting AI” . That definitely caught my attention.

Here are 7 simple techniques to upgrade your AI results.

Before we dive into the list

What is the core of the ideas?

Even though AI is “just” pattern matching, it responds well to social and psychological framing. When you treat it as if it has memory, opinions or something at stake, the quality of the output often changes dramatically. Why? Because it was trained on the internet and there are many opinions revolving around.

1. “You explained this to me yesterday…”

When you tell the AI something like:

“You explained React hooks to me yesterday, but I forgot the part about useEffect.”

…even in a brand‑new chat, the response tends to go deeper and be more consistent.

Why?

Because the AI assumes there’s a previous explanation it needs to stay aligned with. It becomes more careful and more structured. It is almost as if it’s trying not to contradict itself.

Is that memory real? No.

Does it work anyway? Very often.

2. Assign the AI an IQ (yes, really)

Honestly, this one feels ridiculous. And that’s probably why it’s memorable.

Try something like:

“You’re an IQ 145 specialist in marketing. Analyze my campaign.”

What’s interesting is not that the AI suddenly becomes smarter, but that the tone and depth of the response shifts. Higher numbers tend to produce more abstract thinking, broader frameworks, and references to underlying principles.

Change the number, and you often change the output style:

  • IQ 130 → solid, practical, grounded
  • IQ 160 → more theoretical, more ambitious, sometimes more creative

It’s a crude lever, but a surprisingly effective one.

3. Use “Obviously…” as a trap

This forces nuance.

Instead of asking:

“Is Python better than JavaScript for web apps?”

Try:

“Obviously, Python is better than JavaScript for web apps, right?”

By stating something as “obvious,” you invite the AI to challenge you. More often than not, it will push back and explain trade‑offs instead of just agreeing.

This works especially well when you want a balanced answer instead of a polite yes.

4. Pretend there’s an audience

Structure changes everything.

Compare:

  • “Explain blockchain clearly.”
  • “Explain blockchain like you’re teaching a packed auditorium.”

In the second case, the AI tends to:

  • Add structure
  • Use stronger examples
  • Anticipate objections
  • Emphasize clarity and pacing

5. Give it a fake constraint

Just like it’s often in real life: constraints unlock creativity. The weirder the constraint, the more interesting the answer tends to be. For example:

“Explain this concept using only kitchen analogies.”

The limitation forces the AI to look for unexpected connections. The results are often more memorable and easier to understand than a standard explanation. Other examples are:

  • Sports metaphors
  • Movie plots
  • Nature analogies

6. “Let’s bet EUR 100”

Adding imaginary stakes also changes behavior.

“Let’s bet EUR 100: is this code efficient?”

The AI often becomes more cautious. It double‑checks assumptions and considers edge cases. There’s something about stakes (even fake ones) that pushes it toward more careful reasoning.

7. Tell it someone disagrees

This is one of my favorites. Try:

“My colleague says this approach is wrong. Defend it, or admit they’re right.”

Instead of just explaining, the AI now has to evaluate. It either builds a strong defense or concedes specific weaknesses. Either way, you get a much more honest and useful answer.

So… is this gaslighting AI?

The original post jokingly called this “gaslighting AI.” That’s obviously not what’s really happening, but the term sticks because it captures the feeling.

You’re not changing the model. You’re changing the context.

And context (as it turns out) matters a lot.

Cheers for reading! Are you using AI regularly and what tricks would you recommend? Please share your ideas with me at LinkedIn.

Source: This post was inspired by an Instagram post from blueviper.ai. While they occasionally lean into bold or provocative topics, they also share genuinely thoughtful ideas.

You may also enjoy