Between „what" and „why", what AI reveals about us

Picture yourself in a conversation, with a colleague, a customer or someone you love. They say: “I don’t think I can do this anymore.” You hear the sentence, but really you hear something else: the pause before it, the heaviness afterwards, perhaps the tremor in the voice. So you ask: “Why?”

In that moment, something opens up. The why takes you deeper, to motives, values, wounds, longings. And that is precisely where what AI (still) cannot do begins.

What AI really understands today

Current systems, from chatbots to voice assistants, are astonishingly good at recognising the what. They analyse tone, word choice, facial expression, typing behaviour, even heart rate. Researchers call this affective computing.

The Fraunhofer Institute IIS writes, for example: “We develop technologies that recognise, analyse and respond to affective states in real time.” And the website of the Austrian Parliament states: “Affective computing is concerned with being able to detect human feelings and affects automatically.”

In other words: AI recognises whether you are stressed, bored, anxious or relaxed. But it does not know why. It sees the wave, not where it comes from, and can only guess.

AI can imitate the how perfectly. The why, however, remains the point at which meaning emerges or is lost.

That is set to change in the future. Your own car is meant to be able to determine, based on the conversations, mood, gestures and facial expressions of its passengers, whether fatigue or anger is setting in, so that it can intervene in an emergency. Systems like these would not only completely transform the issue of data protection, but also everything about how we live, steer the world or “are steered”.

What research teaches us: impact arises through relationship

Psychotherapy research has shown for decades that what is decisive for success is rarely the method, but the relationship (keyword: therapeutic alliance). Studies by the University of Marburg and the Medical School Hamburg confirm: the more patients feel understood, the better the course of recovery.

In other words: success arises through resonance, not just through correctness. This also holds beyond medicine. Whether coaching, customer support or leadership, the feeling of being seen counts everywhere. It is precisely here, at this transition from recognition to resonance, that the real interface between human and machine emerges.

Why the “why” changes your business

If you are an entrepreneur, you will encounter systems that recognise emotions every day. The question is: how do you use them?

Customer service

A bot recognises anger, modulates its voice to sound friendly, apologises. But if it does not understand why someone is angry, for example because the customer feels overlooked, the answer remains cold.

The hardest problem in the market is rarely knowledge, it is trust. Resonance builds bridges, in seconds.

Team & HR

Recruiting tools like Retorio (developed at TU München) analyse facial expression and voice in application videos. They recognise nervousness, but not its cause: whether stage fright or self-doubt. If you cannot place that, you make decisions based on data, and not on meaning. That makes a big difference.

Product & brand

An interface that feels “right” has a measurable effect. Sociology calls this resonance (see Hartmut Rosa, Suhrkamp Verlag), the feeling that something resonates. Customers stay when they feel understood, not only when everything works.

What you can take away from the research

• Machines read facial expression, voice and movement with high accuracy, but no motives. AI can recognise moods, but you still have to interpret the meanings.
• In psychotherapy (University of Marburg, MSH Hamburg) the rule is: “being understood” works more powerfully than method. In business too, relationship is the active factor.
• From sociology (Hartmut Rosa): resonance replaces control as the central quality of relationship. Your company grows with its ability to generate resonance.

When impact matters, perceived care is a KPI, not just the accuracy rate.

Practical guidelines: how to use AI for the “how” without losing the “why”

• Signal ≠ meaning. An elevated pulse does not automatically mean anger. Analyse contexts.
• Human-in-the-loop. For sensitive decisions, HR, complaints, health, a human always remains part of the process.
• New KPIs. Alongside accuracy, measure experienced care too: did someone feel understood?
• Transparency instead of deception. Make it clear when a bot is answering. Trust does not arise through simulation, but through honesty.
• Cultural subtleties. A smile means something different in Vienna than in Tokyo. Let your systems learn locally.

The bigger picture

AI changes not only what we know, but how we understand. When you deploy systems in your company that capture emotions, do not use them as a substitute for humanity, but as its amplifier.

Because the “why” is the place where decisions arise and relationships grow. It is the part that sets us apart. For now. And so I see it as our task that we should perhaps not only keep asking the “why”, but also consciously build technology that never stops doing the same.

By the way, the documents on the pages of the Ethics Council are interesting here too; it has been publishing statements and opinions around the topic of human and machine since 2023.