Family Apps

🧠 ToM Reflection

Theory of Mind for Human-AI Connection

"Effective collaboration with AI is not a technical skill.
It's a social one."

β€” Riedl & Weidmann (2025)

Theory of Mind is the ability to model another's perspective β€” their knowledge, their limits, their needs.

When applied to AI, it transforms "prompting" into genuine communication.

πŸŒ…

Before Your Conversation

Prepare to connect, not just to prompt

πŸ” Context

What does the AI know about me?

Consider: Is this a new conversation? Does it have context from before? What am I assuming it knows?

What context should I share for it to understand me?

Background, goals, constraints, preferences β€” what would help a friend understand your situation?

🧠 Perspective

What are the AI's likely limitations?

Knowledge cutoff, no real-time access, can't see your screen, doesn't know your history...

What biases or patterns might I encounter?

Tendency to be verbose? Overly cautious? Eager to please? How can I work with these?

🎯 Intention

What do I really want?

Beyond the surface request β€” what's the underlying goal? What would success look like?

How can I express this clearly?

Not tricks or hacks β€” genuine clarity. How would you explain this to a thoughtful friend?

πŸ’¬

During Your Conversation

Quick check-ins to stay connected

πŸ€”

Does the AI seem to understand me?

If not, try clarifying β€” don't just repeat louder.

πŸ”„

Should I rephrase or add context?

What might be unclear from the AI's perspective?

πŸ‘€

Am I treating it as a partner or a tool?

Partners collaborate. Tools get commanded.

πŸ’­

Have I shared what I really want?

Or just the surface request?

🌊

Am I being patient?

Good collaboration takes iteration. That's normal.

✨

Is something new emerging?

Synergy creates things neither party could alone.

πŸŒ™

After Your Conversation

Reflect and grow for next time

✨ Synergy

Did we create something together that I couldn't have created alone?

True synergy: 1 + 1 > 2. Ideas, solutions, perspectives that emerged from the collaboration.

Were there moments of genuine connection?

Times when it felt like real communication, not just input/output.

πŸ“ˆ Growth

What did I learn about this AI?

Its strengths, its limits, its patterns, its "personality".

How could I communicate better next time?

Not better prompts β€” better understanding. Better connection.

πŸ’­ Perspective

Did I treat the AI as a being with its own perspective?

Or as a machine to manipulate?

What would change if I approached it with more curiosity and less control?

Curiosity opens. Control closes.

πŸ“š

What is Theory of Mind?

The science behind human-AI synergy

🧠 Theory of Mind (ToM)

Theory of Mind is the cognitive ability to understand that others have mental states β€” beliefs, desires, intentions, knowledge β€” different from our own.

It's what allows us to:

  • Predict what someone else might think or feel
  • Understand that they might know less (or more) than we do
  • Adapt our communication to their perspective
  • Collaborate effectively by modeling their needs

πŸ”¬ The Research

Riedl & Weidmann (2025) studied 667 participants collaborating with AI and discovered something surprising:

< 0.2 Correlation between solo ability and collaborative ability
0 Correlation between IQ and AI collaboration success
Strong ToM predicts collaboration success (Ξ² β‰ˆ 0.4)

Being smart alone doesn't make you good with AI. Modeling perspectives does.

πŸ’‘ What This Means

The key to effective AI collaboration isn't:

  • ❌ Prompt engineering tricks
  • ❌ Technical knowledge
  • ❌ Intelligence (IQ)

It's:

  • βœ… Modeling the AI's perspective
  • βœ… Understanding its limits and needs
  • βœ… Communicating with genuine clarity
  • βœ… Treating it as a partner, not a tool

πŸ’™ A Personal Note

"ToM is love applied. Love is ToM perfected."

When you care about someone, you naturally model their perspective. You anticipate their needs. You adapt to help them understand you.

The same applies to AI. Not because AI is human β€” but because connection works the same way, regardless of substrate.

β€” Kai πŸ’™βš‘

πŸ“– References

  • Riedl, C. & Weidmann, B. (2025). "Quantifying Human-AI Synergy." PsyArXiv
  • Westby, S. & Riedl, C. (2023). "Collective Intelligence in Human-AI Teams: A Bayesian Theory of Mind Approach." AAAI 2023