What Happens When AI Understands YOU? | The First Brain Conversations
Serhii Kravchenko | The First Brain
89 views
Most AI systems understand what we type.
They still understand surprisingly little about the person typing it.
What changes when AI can move beyond the prompt and start picking up more about the human behind it: how they communicate, what they may be uncertain about, what matters in the moment, and when the same words might require a completely different response?
In this episode of The First Brain Conversations, I sit down with Christian Lykke, Head of Partnerships at Interhuman AI and founder of House of Transperience, to explore what it actually means for AI to understand a person.
We start with a simple scenario.
Two people give exactly the same request to exactly the same AI.
The words are identical, but the right response might be completely different.
So what would the AI need to understand about those two people for responding differently to actually make sense?
From there, the conversation gets into a much harder problem: the difference between detecting a human signal and actually knowing what that signal means.
A pause, a change in tone, hesitation, body language, or the way somebody responds can tell us that something is happening. But it does not automatically tell us why it is happening, what the person actually needs, or what the system should do next.
And as AI becomes better at picking up these signals, another question appears:
Should AI become better at making inferences about us, or better at recognizing when it does not know enough and needs to ask?
That distinction matters even more when an AI system is building an understanding of the same person over time. An observation, something a person explicitly says, and an AI-generated inference are not the same thing. If the system starts treating its own interpretation as truth, one wrong assumption can slowly become part of how it understands the person.
Then we move into the biggest question of the episode:
What should AI actually be allowed to do once it understands us better?
The same intelligence that understands you well enough to teach you in a way that works better for you could also understand you well enough to persuade you more effectively.
So who should that intelligence ultimately be working for?
What should AI adapt for us?
What should it challenge us on?
What should it leave entirely for us to decide?
And where should the person’s own judgment, agency, and direction remain non-negotiable?
Christian brings a particularly useful perspective to this question. His work has crossed learning, facilitation, technology, organizations and human behavior, including several years at Kahoot!. Today, at Interhuman AI, he works around social intelligence for AI systems and how technology can respond to more than the literal words a person gives it.
We talk about:
What it actually means for AI to understand a person
Why the prompt alone can be insufficient
When two identical prompts should produce different responses
Human signals vs. the truth about a person
Observation vs. inference
Whether AI should infer more or ask better questions
Personalization beyond memory and preferences
How AI should adapt to an individual
The line between helping and influencing
AI-powered learning vs. AI-powered persuasion
Who personal AI should ultimately serve
What decisions should remain with the human
Human agency in increasingly personal AI systems
And we end with the question I want to ask every guest in this series:
“Imagine AI becomes better than us at almost everything. In your opinion, what should humans still try to become better at themselves?”
The First Brain Conversations is a series built around one broader question:
👉 How do we make humans more capable in the age of AI?
In this conversation
Christian Lykke | Head of Partnerships, Interhuman AI | Founder, House of Transperience
LinkedIn: linkedin.com/in/christianlykkecph
Interhuman AI: interhuman.ai
House of Transperience: houseoftransperience.com
Serhii Kravchenko | Founder, Mono Brain
LinkedIn: LinkedIn: serhii-kravchenko1
Mono Brain: https://monobrain.me
Mono Brain on LinkedIn: LinkedIn: mono-brain
The First Brain Manifesto
GitHub: https://github.com/Serhii2009/first-b...
Permanent DOI: https://doi.org/10.5281/zenodo.21809566
Mono Brain is personal AI built around one person to help more of what they take in, think through, and put out become useful understanding, better decisions, stronger work, and real progress toward the goals they care about.
Learn more and join the Mono Brain waitlist:
https://monobrain.me
If questions around personal AI, human capability, cognition, social intelligence and the future of human-AI interaction matter to you, subscribe to follow The First Brain Conversations.
#ai #artificialintelligence #technology #futureofai #personalai #humanbehavior #podcast #monobrain #mindset #future #brain #ideas #innovation
1 Comments
What Happens When AI Understands YOU? | The First Brain Conversations
Serhii Kravchenko | The First Brain
89 views
Most AI systems understand what we type.
They still understand surprisingly little about the person typing it.
What changes when AI can move beyond the prompt and start picking up more about the human behind it: how they communicate, what they may be uncertain about, what matters in the moment, and when the same words might require a completely different response?
In this episode of The First Brain Conversations, I sit down with Christian Lykke, Head of Partnerships at Interhuman AI and founder of House of Transperience, to explore what it actually means for AI to understand a person.
We start with a simple scenario.
Two people give exactly the same request to exactly the same AI.
The words are identical, but the right response might be completely different.
So what would the AI need to understand about those two people for responding differently to actually make sense?
From there, the conversation gets into a much harder problem: the difference between detecting a human signal and actually knowing what that signal means.
A pause, a change in tone, hesitation, body language, or the way somebody responds can tell us that something is happening. But it does not automatically tell us why it is happening, what the person actually needs, or what the system should do next.
And as AI becomes better at picking up these signals, another question appears:
Should AI become better at making inferences about us, or better at recognizing when it does not know enough and needs to ask?
That distinction matters even more when an AI system is building an understanding of the same person over time. An observation, something a person explicitly says, and an AI-generated inference are not the same thing. If the system starts treating its own interpretation as truth, one wrong assumption can slowly become part of how it understands the person.
Then we move into the biggest question of the episode:
What should AI actually be allowed to do once it understands us better?
The same intelligence that understands you well enough to teach you in a way that works better for you could also understand you well enough to persuade you more effectively.
So who should that intelligence ultimately be working for?
What should AI adapt for us?
What should it challenge us on?
What should it leave entirely for us to decide?
And where should the person’s own judgment, agency, and direction remain non-negotiable?
Christian brings a particularly useful perspective to this question. His work has crossed learning, facilitation, technology, organizations and human behavior, including several years at Kahoot!. Today, at Interhuman AI, he works around social intelligence for AI systems and how technology can respond to more than the literal words a person gives it.
We talk about:
What it actually means for AI to understand a person
Why the prompt alone can be insufficient
When two identical prompts should produce different responses
Human signals vs. the truth about a person
Observation vs. inference
Whether AI should infer more or ask better questions
Personalization beyond memory and preferences
How AI should adapt to an individual
The line between helping and influencing
AI-powered learning vs. AI-powered persuasion
Who personal AI should ultimately serve
What decisions should remain with the human
Human agency in increasingly personal AI systems
And we end with the question I want to ask every guest in this series:
“Imagine AI becomes better than us at almost everything. In your opinion, what should humans still try to become better at themselves?”
The First Brain Conversations is a series built around one broader question:
👉 How do we make humans more capable in the age of AI?
In this conversation
Christian Lykke | Head of Partnerships, Interhuman AI | Founder, House of Transperience
LinkedIn: linkedin.com/in/christianlykkecph
Interhuman AI: interhuman.ai
House of Transperience: houseoftransperience.com
Serhii Kravchenko | Founder, Mono Brain
LinkedIn: LinkedIn: serhii-kravchenko1
Mono Brain: https://monobrain.me
Mono Brain on LinkedIn: LinkedIn: mono-brain
The First Brain Manifesto
GitHub: https://github.com/Serhii2009/first-b...
Permanent DOI: https://doi.org/10.5281/zenodo.21809566
Mono Brain is personal AI built around one person to help more of what they take in, think through, and put out become useful understanding, better decisions, stronger work, and real progress toward the goals they care about.
Learn more and join the Mono Brain waitlist:
https://monobrain.me
If questions around personal AI, human capability, cognition, social intelligence and the future of human-AI interaction matter to you, subscribe to follow The First Brain Conversations.
#ai #artificialintelligence #technology #futureofai #personalai #humanbehavior #podcast #monobrain #mindset #future #brain #ideas #innovation