WATCH / RESEARCH NOTES · EMPATHETIC TECHNOLOGY

I watched Poppy Crum’s TED talk Technology that knows what you’re feeling because I wanted a clearer, more visual understanding of how technology can interpret signals from the body. What was useful about this talk was that it moved away from thinking only about a wearable collecting heart-rate data and showed a much wider idea: our bodies are constantly producing signals that technology can potentially sense, interpret and respond to.

The “end of the poker face”

Crum’s central idea is that we may have less control over what our bodies reveal than we think. Even when a person is not consciously expressing how they feel, physiological changes can still be detected through things such as pupil dilation, body temperature, voice, language and breath chemistry.

One example in the talk shows the pupil changing size as cognitive effort changes. When the listening task becomes harder, the autonomic nervous system affects pupil dilation. This was useful for me because it made physiological sensing feel much more immediate: the body can change in response to effort before a person has consciously explained what they are experiencing.

For my project, this connects with the idea that an autistic traveller may experience increasing physiological demand before they actively ask for help. However, I do not want to translate this into “the system knows the person is overloaded”. The signal only shows that something has changed.

Important for my project: physiological data can reveal change, but it cannot automatically explain the cause of that change. The system should support interpretation rather than pretend to know exactly what the person is feeling.

More than one kind of signal

The talk was useful because it showed how broad the idea of biosensing can be. Crum discusses technologies that can detect changes through thermal imaging, patterns in speech and language, and even the chemical composition of breath.

This helped me move away from thinking that my project has to depend on one specific signal such as HRV. A future system could potentially combine several types of information, although for my Studio Practice outcome I still need to decide which signals are realistic, non-invasive and appropriate within an airport.

The broader point I am taking from this is that the body produces multiple forms of data. The design challenge is deciding which ones are useful enough to include without making the system invasive, overly complex or falsely authoritative.

Technology as an “empathetic” partner

Crum describes the possibility of technology becoming more responsive to the person using it rather than remaining static. If technology can recognise meaningful changes in a person’s physiological state, an environment or interface could potentially adapt in response.

This is particularly relevant to the direction of my project because I have already identified a problem with fixed accessibility provision. Airport support often exists as a set of predetermined services, while an individual’s needs may change throughout the journey.

The idea of “empathetic technology” gives me another way to think about a bioresponsive system: not as a device that diagnoses the user, but as a system that becomes more sensitive to changes and offers support accordingly.

For example, a change in physiological arousal could trigger an optional prompt, a route adjustment, a quieter information mode, haptic support or another intervention. The important part is that the person should still decide whether that response is useful.

Agency becomes even more important

One part of the talk that feels especially relevant is the tension between what technology can detect and what a person chooses to share. If technology becomes capable of sensing internal states that someone has not deliberately communicated, that immediately raises questions around privacy, consent and control.

This is important for my project because I do not want accessibility technology to become surveillance technology. An autistic traveller should not lose agency simply because a wearable can collect physiological information.

That means the final system would need to make it very clear:

• what is being sensed;
• when sensing is active;
• what the system thinks it has detected;
• what data are stored or shared;
• and whether the user wants any response to happen.

This reinforces the argument I have already been developing that independence is linked to control rather than the absence of support.

What this changes for my project

This talk did not give me a specific wearable form, but it helped me understand the bigger design space around biosensing. I am no longer thinking only about a wristband that “detects stress”. I am thinking about how different physiological signals could become inputs into a responsive support system.

The most useful things I am taking forward are:

• the body produces signals that may change before someone verbally communicates a need;
• physiological sensing can come from much more than heart rate alone;
• combining signals may be more useful than relying on one measurement;
• the technology should identify change, not claim certainty about emotion or sensory overload;
• any adaptive response should remain optional and user-controlled;
• privacy and transparency need to be designed into the system from the beginning;
• the most interesting design opportunity may be how sensing changes the interface, environment or support response, rather than the sensor itself.

Questions to carry forward

This leaves me with a more specific set of questions for the next stage of Studio Practice. Which signals are realistic to use in an airport? Which are discreet enough for everyday use? What kind of intervention would actually be helpful? How should the system communicate uncertainty? And how can the user remain in control if the technology is sensing something they have not consciously chosen to communicate?

The talk also makes me want to explore adaptive and immersive technologies more closely, because it suggests that sensing becomes most interesting when it changes what the user experiences next.

Source / Further Reading