Embodied agency and emotion: When AI learns to read the room

For decades, we’ve interacted with technology as if it were emotionally tone-deaf. We clicked buttons, filled in forms and spoke to chatbots that responded politely but mechanically.

Technology processed data, but it didn’t understand us. That’s now changing.

The NTT DATA Technology Foresight 2026 report explores six current macrotrends, including what happens when intelligence becomes embodied and emotionally aware — in other words, when systems perceive our tone, gestures, stress levels and context, so they can respond not only to what we say but also to how we say it.

This shift moves AI from purely cognitive interaction into relational experience.

• ALSO READ → Technology Foresight 2026: What happens when intelligence is everywhere?

A new world of emotionally responsive technology

Digital humans are being equipped with emotionally expressive interfaces. These systems combine voice, gesture, facial expression and movement to create presence across physical and digital spaces, using both emotional detection and emotional response.

Imagine you’re in an online consultation with a virtual doctor. In addition to recording your symptoms, the digital assistant on screen also notices a slight tremor in your voice. Having deduced that your breathing pattern suggests stress, it gently slows the pace of conversation and asks a clarifying question in a softer tone.

In a factory setting, a collaborative robot might sense that a human operator is tired by analyzing multimodal cues: posture, movement patterns and tiny pauses. It then adjusts its speed and prompts a break before a mistake happens.

Similarly, in customer service, a digital human avatar might detect a customer’s rising frustration before it escalates out of control. The avatar adjusts its tone, proactively escalates the call to a human when appropriate or suggests a targeted solution, not through scripted programming but by interpreting emotional signals in real time.

These scenarios are extensions of the technologies described in our report: Multimodal sensing, affective computing, digital humans, embodied AI and emotionally adaptive systems.

The promise — and the tradeoffs

Emotionally intelligent systems offer powerful upsides across industries, including more inclusive services, more responsive healthcare, better collaboration and safer mobility.

But they also raise complex questions. If a system can detect stress, can it manipulate it? If digital humans become highly persuasive, where is the boundary between empathy and influence? If workplaces begin tracking the emotional climate, who owns that data?

There are also deeper technical and ethical risks to address. Emotional AI systems are trained on human data — voice recordings, facial expressions and behavioral signals. Those datasets can reflect cultural bias, demographic imbalance or contextual blind spots. A system that interprets tone or facial cues inaccurately across different ethnicities, languages or neurodiverse behaviors could misclassify emotion.

Our report makes it clear that bias testing, representative training data, transparency around emotional inference models and clear human override mechanisms must be built into these solutions from the start, with strong norms for privacy and human agency. Emotional data carries psychological weight and cannot be treated like ordinary telemetry.

A new layer of infrastructure

What’s striking is how quickly emotional intelligence is moving from experimentation to becoming foundational as multimodal sensing becomes standard and digital humans scale across industries.

Emotion-aware interfaces are already being embedded into service models — especially in healthcare, automotive safety, education and customer experience.

In the longer term, our report also considers autonomous affective environments, referring to workplaces or care settings that adapt lighting, sound and interaction flows based on humans’ collective emotional states.

In other words, emotional intelligence is set to become part of our digital infrastructure, as fundamental as connectivity or cloud. And systems that can understand emotional context and respond appropriately will outperform those that merely execute tasks.

There’s more to the architecture of the future

Embodied agency and emotion is just one of six interconnected macrotrends in the era of mass intelligence.

The full NTT DATA Technology Foresight 2026 report also explores human-orchestrated autonomy, trustworthy intelligence, informed infrastructure, sovereign silicon ecosystems and the move from illusory efficiency to sufficiency. These trends describe a future where intelligence becomes adaptive, embodied and deeply human-centered.

For insight into how these forces converge and what they mean for your organization, download the full report.