Scroll ⬇
😊 😢 😠 😲 😏

NATIONAL GEOGRAPHIC LEARNING

🤖

Unit 10

Emotion in Technology

Can Machines Learn to Feel?

Lead-in 01

Can a robot tell if you're happy or sad? Should it be able to? 🤔

Scientists have mapped over 5,000 facial muscle movements. Now, technology can read what you feel — even before you say a word.

😶

5,000 Expressions

Paul Ekman mapped them all

📱

Blind-Assist App

Dr. Asakawa's life-changing work

🤖

Pepper Robot

Reads emotions live

🔒

Privacy Risk

Who owns your feelings?

The boundary between human emotion and machine intelligence is getting blurrier — and it raises profound questions.

Reading 02

Skimming Task ⏱️

Read the article quickly (90 seconds). Answer three questions:

🧑‍🔬

WHO IS EKMAN?

What was Paul Ekman's contribution to emotion technology?

🤖

WHERE IS PEPPER?

Where is the Pepper robot already deployed, and what does it do?

⚙️

HOW IT WORKS

How does facial recognition software learn to recognize emotions?

✅ EKMAN: Analyzed 5,000+ facial muscle movements and discovered micro-expressions — companies use his work to build emotion-recognition software  |  PEPPER: Retail stores in Japan — comforts sad customers, jokes with playful ones  |  HOW: Software analyzes thousands of faces and learns to recognize emotions with increasing accuracy
Introduction
Science Fiction Becoming Real
The gap between imagined AI and real AI is closing fast.
Reading 03
Sci-Fi Scenarios & Reality
People have long imagined a world where we interact with computers and robots as if they were normal human beings. Science fiction movies such as Her and Chappie show computers and robots that think and feel just like humans. While scenarios like these exist only in the movies for now, we may be getting close to making technology emotionally intelligent.
The adverb 'long' with present perfect signals that this imagining has been sustained over a significant period — not a recent fantasy but a deep-rooted human dream. It establishes that the desire for emotionally intelligent machines is not new (science fiction has explored it for decades), but it is only now potentially achievable. The combination of 'have long imagined' with 'may be getting close' in the next sentence creates a sense of historical patience finally meeting technological possibility.
Reading 04
Science Fiction Examples
People have long imagined a world where we interact with computers and robots as if they were normal human beings. Science fiction movies such as Her and Chappie show computers and robots that think and feel just like humans. While scenarios like these exist only in the movies for now, we may be getting close to making technology emotionally intelligent.
Naming specific films (Her and Chappie) rather than vague 'science fiction movies' gives the claim concrete, verifiable grounding. Readers familiar with these films immediately activate their knowledge of the stories — Her features an AI operating system capable of love; Chappie features a robot that develops consciousness and emotion. This specificity also signals the author's credibility and awareness of the cultural conversation. The phrase 'think and feel just like humans' pairs cognition (think) with emotion (feel), previewing the article's core theme: both intelligence and emotion must be replicated for true AI humanity.
Reading 05
Reality vs. Science Fiction
People have long imagined a world where we interact with computers and robots as if they were normal human beings. Science fiction movies such as Her and Chappie show computers and robots that think and feel just like humans. While scenarios like these exist only in the movies for now, we may be getting close to making technology emotionally intelligent.
'While' creates a concessive acknowledgment: yes, sci-fi is still fiction. But immediately, 'may be getting close' introduces cautious optimism. 'May' is a modal of possibility; 'getting close' (progressive) suggests movement, approach — an ongoing process. Together, the sentence creates a sense of tantalizing proximity: we're not there yet, but we're heading there. This is an effective hook for a technology article: acknowledging the gap while promising that it's narrowing. The hedging ('may') keeps it intellectually honest.
Para 2
Understanding Emotion
Even scientists struggle to define what emotions are.
Reading 06
Defining Emotions Is Hard
The first step toward this is understanding what emotions are. It's a complicated area of study. Scientists are often unable to define emotions in exact terms, even though we generally understand what people mean when they say they're sad or happy.
'The first step' signals a sequential, problem-solving structure: the author is framing the path to emotion AI as a series of logical steps, the first of which is definition. This sets up expectations: if step one is defining emotion, subsequent steps will be measuring, replicating, and applying it. The phrase also implies that this step is necessary but not sufficient — more steps follow. It gives the paragraph forward momentum and signals to the reader that a structured argument is unfolding.
Reading 07
A Complicated Area of Study
The first step toward this is understanding what emotions are. It's a complicated area of study. Scientists are often unable to define emotions in exact terms, even though we generally understand what people mean when they say they're sad or happy.
The short sentence creates a rhetorical pause — a moment that acknowledges complexity before the paragraph continues. The brevity is itself a signal: this cannot be explained in one breath. "It's a complicated area" also shifts register to conversational directness, like the author briefly breaking the academic frame to speak informally. This transparency builds trust: the author doesn't hide the difficulty, which makes the subsequent technical details feel more credible. Short sentences after long ones create emphasis through contrast.
Reading 08
Universal Feeling, Elusive Definition
The first step toward this is understanding what emotions are. It's a complicated area of study. Scientists are often unable to define emotions in exact terms, even though we generally understand what people mean when they say they're sad or happy.
The paradox is: humans experience and recognize emotions intuitively but cannot define them precisely. For technology, this is a profound challenge: machines need formal definitions to be trained, but the very thing they need to recognize resists formal definition. The 'even though' concession acknowledges shared human understanding while revealing its limits as a foundation for programming. This is why emotion AI is hard: teaching a machine to recognize something that even experts can't define requires a different approach entirely — pattern recognition from data, not rule-based logic.
Para 3
Paul Ekman's Discovery
5,000 muscle movements — and one breakthrough insight.
Reading 09
Ekman's Pioneering Research
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
'Back in' is an informal temporal marker that signals distance — both in time and, implicitly, in scientific perspective. It creates a slightly nostalgic or retrospective tone, like looking back at a less enlightened era. The word 'few' (not 'no') acknowledges some researchers existed, but positions the field as underdeveloped. Together, 'back in the 1950s, few' sets up the contrast with today's flourishing emotion tech industry — making Ekman's choice to study emotion then even more remarkable (ahead of his time).
Reading 10
Ekman Sees Potential
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
Starting with 'But' creates a sharp rhetorical pivot — it positions Ekman as an exception against the grain of his era. Grammatically, beginning a sentence with 'But' is acceptable in modern English and is used for dramatic contrast. Here it signals: most scientists ignored emotion, but this one man did not. The word 'potential' is also key: Ekman did not just study emotion — he 'saw potential' in it, implying he had the foresight to recognize future value where others did not. This frames him as a visionary, not merely a researcher.
Reading 11
5,000 Muscle Movements
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
The large number (5,000+) demonstrates both the complexity of human expression and the scope of Ekman's work. For the tech argument, it is crucial: emotion recognition is not simple because faces are extraordinarily complex. The scale also makes the achievement of software that can read this complexity more impressive. The number contrasts powerfully with the simplicity of our everyday emotional vocabulary (happy, sad, angry) — showing that what seems simple at the surface level is deeply complex underneath. This complexity justifies why the technology took decades to develop.
Reading 12
Muscle Movements Form Expressions
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
'Combine' reveals that facial expressions are not single, atomic units but rather composite systems — built from multiple smaller components working together. This is crucial for technology: it means a machine must track the interaction of many muscles simultaneously, not just detect one signal. The word 'different' before 'expressions' implies a finite but rich taxonomy of combinations. This sentence essentially provides the engineering logic for Ekman's entire system: catalogue the components (5,000 movements), understand how they combine, and you can map the full range of human expression.
Reading 13
Micro Expressions — Emotions We Try to Hide
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
This phrase introduces a fundamental ethical tension: humans have always had the ability to hide certain emotions (for dignity, strategy, or protection). Technology that can read micro-expressions essentially removes the choice to conceal. The phrasing 'try to hide' implies effort — humans work to maintain emotional privacy. If machines can bypass that effort, it represents a profound shift in human agency. This quietly foreshadows the privacy discussion in Para 6, planting the ethical seed here in what seems like a scientific description.
Reading 14
Technology Companies Adopt Ekman's Work
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
'Have now started' (present perfect) positions the commercial adoption as a recent development — it implies that Ekman's 1950s research has only recently become commercially viable. The word 'now' emphasizes the contemporary moment, suggesting that something has changed to make this adoption possible (likely the rise of computing power and big data). The phrase 'a number of' (rather than naming specific companies) generalizes the trend, suggesting this is an industry-wide shift rather than one company's decision. The gap between 1950s research and today's application also underlines the long incubation period of fundamental research.
Reading 15
Software Learns from Faces
Back in the 1950s, few scientists studied emotion. But American psychologist Paul Ekman saw a lot of potential in this field. He began analyzing facial expressions, and compiled a list of over 5,000 muscle movements. These muscle movements combine to form our different expressions. His discovery of micro expressions — facial expressions that last only a fraction of a second — allows us to read the emotions that people try to hide. A number of technology companies have now started to use Dr. Ekman's work to create software that recognizes human facial expressions. By analyzing thousands of different faces, the software learns to recognize different emotions with greater and greater accuracy.
The repeated comparative 'greater and greater' creates a sense of continuous, accelerating improvement — each iteration is not just a little better, but the improvement itself keeps growing. 'With increasing accuracy' is correct but static — it describes a direction. 'Greater and greater' conveys momentum and compounding progress. The repetition also has a rhythmic, almost cinematic quality — it mirrors the machine learning process of iterative improvement. This is an example of how word choice can encode the nature of a process, not just describe it.
Para 4 & 5
Real-World Applications
From blind-assist apps to retail robots — emotion tech is already here.
Reading 16
Many Possible Uses
There are many possible uses of emotion-sensing technology. Dr. Chieko Asakawa, a researcher at Carnegie Mellon University, has been blind since the age of 14. She has been developing a smartphone app that might be able to help people with disabilities. Using the smartphone's camera and audio, the app helps the user navigate their environment. It also recognizes people's faces and facial expressions as they approach. Dr. Asakawa is working to refine the app to enable it to read people's moods.
'Possible' is a deliberate hedge — it acknowledges these uses are not all fully realized yet. It also signals breadth: the paragraph will not exhaust all uses, just illustrate a few from many. This topic sentence acts as a signpost: readers understand that what follows will be examples, not an exhaustive list. The use of 'there are' (existential construction) rather than 'scientists use' keeps the focus general before narrowing to a specific case (Dr. Asakawa). This is a classic general-to-specific paragraph structure.
Reading 17
Dr. Asakawa — Researcher and Blind Since 14
There are many possible uses of emotion-sensing technology. Dr. Chieko Asakawa, a researcher at Carnegie Mellon University, has been blind since the age of 14. She has been developing a smartphone app that might be able to help people with disabilities. Using the smartphone's camera and audio, the app helps the user navigate their environment. It also recognizes people's faces and facial expressions as they approach. Dr. Asakawa is working to refine the app to enable it to read people's moods.
Including Dr. Asakawa's blindness as part of her professional introduction creates a profound narrative connection: the very condition she has lived with since 14 is what motivates and informs her research. The appositive (set off by commas) signals that this is background information — but its placement immediately after her title makes it feel essential, not incidental. This technique is called biographical motivation framing: her personal experience is presented as both context and credibility. She doesn't just study accessibility; she lives it.
Reading 18
Developing a Smartphone App
There are many possible uses of emotion-sensing technology. Dr. Chieko Asakawa, a researcher at Carnegie Mellon University, has been blind since the age of 14. She has been developing a smartphone app that might be able to help people with disabilities. Using the smartphone's camera and audio, the app helps the user navigate their environment. It also recognizes people's faces and facial expressions as they approach. Dr. Asakawa is working to refine the app to enable it to read people's moods.
'Has been developing' (present perfect continuous) signals an ongoing, unfinished process — the app is not yet complete. 'Might be able to' is a double hedge: 'might' (possibility) + 'be able to' (conditional capability). Together these two hedges signal that this is work in progress with uncertain outcomes. The author is being honest: the app is not finished, and even when finished, it may not work perfectly. This careful hedging reflects good scientific communication: promising but not overpromising, which actually increases the author's credibility.
Reading 19
Navigating the Environment
There are many possible uses of emotion-sensing technology. Dr. Chieko Asakawa, a researcher at Carnegie Mellon University, has been blind since the age of 14. She has been developing a smartphone app that might be able to help people with disabilities. Using the smartphone's camera and audio, the app helps the user navigate their environment. It also recognizes people's faces and facial expressions as they approach. Dr. Asakawa is working to refine the app to enable it to read people's moods.
Fronting the participial phrase 'Using the smartphone's camera and audio' prioritizes the mechanism of action before the outcome. It answers 'how?' before 'what?', which is natural in technology writing where readers want to understand the method. This structure also emphasizes the innovation: the phone's existing hardware (camera + audio) is being repurposed for a new function — navigation assistance. The phrase makes the technology feel accessible and practical: no special equipment, just the smartphone already in everyone's pocket.
Reading 20
Recognizing Faces as They Approach
There are many possible uses of emotion-sensing technology. Dr. Chieko Asakawa, a researcher at Carnegie Mellon University, has been blind since the age of 14. She has been developing a smartphone app that might be able to help people with disabilities. Using the smartphone's camera and audio, the app helps the user navigate their environment. It also recognizes people's faces and facial expressions as they approach. Dr. Asakawa is working to refine the app to enable it to read people's moods.
'Also' signals that the app has multiple layers of capability: navigation is not its only function. The addition of face and expression recognition moves the app from spatial awareness to social awareness — knowing not just where obstacles are, but who is approaching and what mood they are in. This is a significant step: it means the app does not just help a blind user move through physical space, but also helps them read the social environment — something sighted people take for granted but which is inaccessible without vision.
Reading 21
Refining the App to Read Moods
There are many possible uses of emotion-sensing technology. Dr. Chieko Asakawa, a researcher at Carnegie Mellon University, has been blind since the age of 14. She has been developing a smartphone app that might be able to help people with disabilities. Using the smartphone's camera and audio, the app helps the user navigate their environment. It also recognizes people's faces and facial expressions as they approach. Dr. Asakawa is working to refine the app to enable it to read people's moods.
The chain of infinitives creates a goal-within-a-goal-within-a-goal structure: she is working (immediate action) to refine (short-term goal) to enable (medium-term capability) to read moods (ultimate objective). This layered structure linguistically mirrors the iterative nature of technology development: each step is a prerequisite for the next. The final goal — 'read people's moods' — sits at the end of the chain, emphasizing that it is the ultimate aim, not yet achieved. The present continuous 'is working' also keeps this firmly in ongoing, present-tense effort.
Reading 22
Pepper Robot — Emotional Companion
Another use of emotion-sensing technology can be illustrated through human-shaped robots like Pepper. Launched in Japan in 2015, Pepper is an interactive companion robot. It's capable of recognizing basic human emotions and responding appropriately. For example, it comforts someone when it senses the person is sad, or cracks a joke when the person is feeling playful. In Japan, Pepper is already serving customers in retail stores.
'Another' explicitly signals continuity from the previous example (Dr. Asakawa's app), reinforcing the paragraph's topic sentence that 'there are many possible uses.' It tells the reader: we are still exploring the same theme, but from a new angle. 'Human-shaped' is a significant qualifier: it signals that Pepper's physical form mirrors ours, which creates the expectation of human-like interaction. The design choice of human shape is itself a statement — makers believe we respond better to machines that look like us. This connects to the article's deeper theme about the boundary between human and machine.
Reading 23
Pepper — Interactive Companion Robot
Another use of emotion-sensing technology can be illustrated through human-shaped robots like Pepper. Launched in Japan in 2015, Pepper is an interactive companion robot. It's capable of recognizing basic human emotions and responding appropriately. For example, it comforts someone when it senses the person is sad, or cracks a joke when the person is feeling playful. In Japan, Pepper is already serving customers in retail stores.
Fronting 'Launched in Japan in 2015' gives the reader immediate context: country of origin and date. This is efficient journalism — it grounds Pepper in reality before describing its function. 'Japan' is also significant: Japan has a strong cultural tradition of humanoid robotics and is more accepting of social robots than many Western countries, which adds credibility to Pepper's deployment. 'Companion' is the key word in the description: not 'worker,' not 'tool,' but companion — implying relationship, emotion, and social interaction. This is precisely what emotion-sensing technology enables.
Reading 24
Recognizing and Responding to Emotions
Another use of emotion-sensing technology can be illustrated through human-shaped robots like Pepper. Launched in Japan in 2015, Pepper is an interactive companion robot. It's capable of recognizing basic human emotions and responding appropriately. For example, it comforts someone when it senses the person is sad, or cracks a joke when the person is feeling playful. In Japan, Pepper is already serving customers in retail stores.
'Basic' is a careful qualifier that limits the claim to fundamental, primary emotions (happy, sad, angry, surprised) — not complex blends like 'wistfully nostalgic' or 'guardedly optimistic.' This honestly acknowledges that emotion AI is still at an early stage. The word also prevents overstating Pepper's capabilities. 'Capable of' (rather than 'can') is a formal construction that frames the ability as a designed feature rather than inherent intelligence — maintaining a distinction between simulation and genuine understanding.
Reading 25
Comforting the Sad, Joking with the Playful
Another use of emotion-sensing technology can be illustrated through human-shaped robots like Pepper. Launched in Japan in 2015, Pepper is an interactive companion robot. It's capable of recognizing basic human emotions and responding appropriately. For example, it comforts someone when it senses the person is sad, or cracks a joke when the person is feeling playful. In Japan, Pepper is already serving customers in retail stores.
The contrast between 'comforts' (gentle, empathetic response to sadness) and 'cracks a joke' (playful, light response to playfulness) demonstrates that Pepper's responses are emotionally calibrated to the detected mood — it does not apply one response to all situations. 'Cracks a joke' is particularly interesting: the colloquial phrase (rather than 'tells a joke') implies spontaneity, humor, and timing — qualities associated with social intelligence. This pairing also shows the range from supportive to entertaining, suggesting Pepper can function as both emotional support and social companion.
Reading 26
Pepper Already in Retail Stores
Another use of emotion-sensing technology can be illustrated through human-shaped robots like Pepper. Launched in Japan in 2015, Pepper is an interactive companion robot. It's capable of recognizing basic human emotions and responding appropriately. For example, it comforts someone when it senses the person is sad, or cracks a joke when the person is feeling playful. In Japan, Pepper is already serving customers in retail stores.
'Already' implies that this deployment has arrived sooner than expected — it carries a tone of mild surprise or emphasis on speed. It pushes back against any reader who might assume emotion robots are purely science fiction: they are not future; they are present. 'Serving customers' uses service vocabulary, framing Pepper as a functional employee rather than a novelty or experiment. 'Retail stores' grounds the technology in an everyday, commercial setting — not a research lab. Together, these choices signal that emotion AI has crossed from prototype to deployment.
Para 6 — Challenges
The Privacy Problem
Emotionally intelligent devices come with major ethical risks.
Reading 27
Fascinating — but Risky
Although the idea of emotionally intelligent devices may sound fascinating, this technology can create some major challenges. The issue of privacy is something that many people, including Paul Ekman, are concerned about. For example, as we walk on the streets, devices and scanners could record our facial expressions without our knowledge. This could allow many people to monitor or view our feelings without permission. It may leave us no control over who we share our feelings with. However, if we can negotiate these challenges successfully, there could be many benefits for all of us if our devices become a little more human.
The two modals: 'may sound' (possibility of a perception) vs. 'can create' (established capability). 'May sound' is more tentative — it's about how things appear. 'Can create' is stronger — it's about what the technology is objectively capable of doing. The contrast signals a gap between surface appeal (fascinating) and underlying danger (challenges). This is a classic persuasive move: acknowledge the allure, then deliver the warning. The 'although…' structure creates a warning letter structure: I know you're excited, but consider this.
Reading 28
Privacy — A Shared Concern
Although the idea of emotionally intelligent devices may sound fascinating, this technology can create some major challenges. The issue of privacy is something that many people, including Paul Ekman, are concerned about. For example, as we walk on the streets, devices and scanners could record our facial expressions without our knowledge. This could allow many people to monitor or view our feelings without permission. It may leave us no control over who we share our feelings with. However, if we can negotiate these challenges successfully, there could be many benefits for all of us if our devices become a little more human.
Including Ekman's name here creates a powerful irony and credibility move: the very person whose research made emotion-recognition software possible is himself worried about its implications. This is not a critic from outside the field — it is the founder of the field raising the alarm. This rhetorical move — using an insider's concern — is far more persuasive than citing an external critic. It signals that the risks are not hypothetical fears but are recognized even by those who created the technology. 'Many people, including' also broadens the concern: it is not just Ekman, but a wide community of thinkers.
Reading 29
Recorded Without Your Knowledge
Although the idea of emotionally intelligent devices may sound fascinating, this technology can create some major challenges. The issue of privacy is something that many people, including Paul Ekman, are concerned about. For example, as we walk on the streets, devices and scanners could record our facial expressions without our knowledge. This could allow many people to monitor or view our feelings without permission. It may leave us no control over who we share our feelings with. However, if we can negotiate these challenges successfully, there could be many benefits for all of us if our devices become a little more human.
The passive 'could record' (with no named agent) makes the threat feel diffuse and omnipresent — it could be governments, corporations, or anonymous individuals. The unnamed agent is more unsettling than a specific one because it's everywhere and nowhere. 'Without our knowledge' completes the violation: not only is the agent unnamed, but we wouldn't even know it's happening. This is a textbook example of how passive voice in surveillance discourse amplifies anxiety by removing the possibility of identifying or confronting the threat.
Reading 30
Monitoring Feelings Without Permission
Although the idea of emotionally intelligent devices may sound fascinating, this technology can create some major challenges. The issue of privacy is something that many people, including Paul Ekman, are concerned about. For example, as we walk on the streets, devices and scanners could record our facial expressions without our knowledge. This could allow many people to monitor or view our feelings without permission. It may leave us no control over who we share our feelings with. However, if we can negotiate these challenges successfully, there could be many benefits for all of us if our devices become a little more human.
The shift from 'facial expressions' (physical, observable) to 'feelings' (internal, private) is a deliberate escalation. It reveals that the technology does not just record surface behavior — it decodes interior emotional states. This is a profound distinction: we accept that cameras record our faces in public, but we do not accept that they can also read our minds. The word 'feelings' makes the invasion feel more intimate and more violating. 'Without permission' reinforces the consent issue: the problem is not the technology itself but the removal of individual agency over one's most private experience.
Reading 31
No Control Over Who Knows How We Feel
Although the idea of emotionally intelligent devices may sound fascinating, this technology can create some major challenges. The issue of privacy is something that many people, including Paul Ekman, are concerned about. For example, as we walk on the streets, devices and scanners could record our facial expressions without our knowledge. This could allow many people to monitor or view our feelings without permission. It may leave us no control over who we share our feelings with. However, if we can negotiate these challenges successfully, there could be many benefits for all of us if our devices become a little more human.
'Share' is a deliberately ironic word choice: sharing normally implies consent, willingness, and a social act. By using 'share' in the context of involuntary emotional surveillance, the sentence creates a dark inversion: our feelings are 'shared' not by our choice but by the technology's action. This is a form of semantic subversion — taking a positive word (share) and exposing its hollow meaning in a surveillance context. 'No control over who' emphasizes that the violation is about autonomy and consent, not just exposure. We can share feelings willingly; what we cannot do is control who accesses them without our knowledge.
Reading 32
Balanced Conclusion — Benefits if Negotiated
Although the idea of emotionally intelligent devices may sound fascinating, this technology can create some major challenges. The issue of privacy is something that many people, including Paul Ekman, are concerned about. For example, as we walk on the streets, devices and scanners could record our facial expressions without our knowledge. This could allow many people to monitor or view our feelings without permission. It may leave us no control over who we share our feelings with. However, if we can negotiate these challenges successfully, there could be many benefits for all of us if our devices become a little more human.
The conditional 'if' (not 'when') maintains genuine uncertainty about whether the challenges will be solved. This is intellectually honest: the author does not claim the problems are resolved or inevitable. 'Negotiate' is a careful word choice — it implies active engagement, compromise, and effort (not elimination or suppression). 'A little more human' is deliberately modest in scale: not asking for fully sentient AI, just technology with more emotional intelligence. The hedged, conditional ending mirrors the nature of the technology itself: promising but uncertain, powerful but fragile.
Language 33

Non-Restrictive Relative Clause

Adding background info without changing the main meaning

who + non-restrictive comma + who extra information
A) "Paul Ekman, who saw a lot of potential in this field, began analyzing facial expressions." — text example
B) "Dr. Asakawa, who has been blind since 14, is developing a smartphone app." — same structure
C) Restrictive: "Scientists who study emotions are called affective scientists." — no commas; changes meaning
D) RULE: Commas + who/which = non-restrictive (adds info, removable); No commas = restrictive (defines which one, not removable)
Quick test: Remove the relative clause. Does the sentence still make clear sense?
— "Paul Ekman began analyzing facial expressions." ✅ (Non-restrictive — info was extra)
— "Scientists began analyzing facial expressions." ❌ (Restrictive — which scientists? We need the clause)
Writing tip: Non-restrictive clauses are excellent for adding biographical or contextual information about a named individual without interrupting the narrative flow.
Language 34

"Capable of + -ing" vs "Able to + Infinitive"

Two ways to express ability — register and context matter

capable of able to can designed to
A) "It's capable of recognizing basic human emotions." — formal; suggests designed/built-in ability
B) "The app is able to navigate environments." — neutral; general acquired ability
C) Informal: "The robot can recognize emotions." — simplest; doesn't indicate how ability was acquired
D) RULE: capable of = often for designed/inherent capacity (machines, skilled people); able to = broader, context-free ability; can = simplest, most informal
Key distinction: 'Capable of + gerund' is preferred when describing a designed or trained ability — it implies the ability was engineered or developed, fitting perfectly for technology articles. 'Able to + infinitive' is more general and often refers to current situational ability. 'Can' is the most conversational. In formal academic writing about AI and technology, prefer 'capable of'.
Language 35

Passive Voice in Privacy/Surveillance Discourse

When the agent is hidden — and why that's terrifying

could record without our knowledge agentless passive surveillance
A) "Devices could record our facial expressions without our knowledge." — text (active but unnamed subject)
B) "Our feelings could be monitored by unknown parties." — passive; agent "unknown parties" added for clarity
C) "We would not know who was recording us." — active but emphasizes ignorance
D) RULE: Removing the agent from surveillance language creates diffuse, unidentifiable threat — more psychologically unsettling than naming a specific actor
Rhetorical effect: When you can name your enemy, you can fight them. When the agent is absent or unnamed, resistance feels impossible. This is why privacy advocates (and this article) often use agentless constructions — 'data is collected,' 'faces are scanned,' 'emotions are recorded' — to amplify the sense of exposure and helplessness. It's not just grammar; it's a rhetorical choice with real emotional impact on the reader.
🤖

LESSON COMPLETE

😶

5,000 Expressions

Ekman's lifetime of research

📱

Accessibility Tech

Dr. Asakawa's blind-assist app

🤖

Pepper in Retail

Emotion recognition deployed

🔒

Privacy at Stake

Who controls your emotions?

"If we can negotiate these challenges, there could be many benefits if our devices become a little more human."