We carried out a design workshop with 9 experts to gain an initial understanding on how we can leverage an agent's face as canvas for risk communication. From the workshop, we collected 23 design outcomes, three overarching design strategies as well as challenges and opportunities for employing an agent's face as interface for risk communication.
๐ฑ = Environment & Ecology
๐ = Health & Well-beingย
๐ฃ๏ธ = Social Issue
๐ฃ๏ธ
The design explores flowers as metaphorical representation for each harassment type. The falling flower petals symbolise the frequency of the respective harassment type, whereas the projected light intensity conveys the time of day. Besides, the design experiments with facial expressions, conveying sadness each time a petal falls.
๐ฃ๏ธ
The design assigns facial features and movements to represent harassment types. For instance, eye movement conveys stalking, whereas mouth muttering represents verbal assault. The time of day is represented through the tiredness of the eyes. The design explores head shakes as a way to represent the frequency of harassment.
๐ฃ๏ธ
Prototyping of an agent representing verbal assault as one potential type of harassment. The facial regions represent time, mapped as a clock. Combined with facial expressions and varying colour intensity to visualise the harassment frequency.
๐ฃ๏ธ
The design uses different types of agent faces for each type of harassment case. The number of tears falling down the robot's face represents the frequency of harassment. The design further explores facial expressions to elicit a stronger emotional response.
๐ฃ๏ธ
The design uses the idea of suppressing emotions, that gradually reveals themselves depending on the frequency of harassment cases. Each facial region represents one type of harassment.
๐ฑ
The design experiments with projected smoke, a spreading skin rash and pimples to convey each type of air pollutant, i.e. diesel conveyed by smoke. The design includes various types of face masks to convey the degree of air pollution.
๐ฑ
This design introduces "pollution make-up" and shows the agent's face as an artistic landscape, with dedicated facial regions to visualise each air pollutant. Depending on the degree of air pollution, the face is either covered by leaves, starts to crack or transitions into a Terminator-like appearance.
๐ฑ
The design represents the air quality, exemplified for low pollution levels. A fabric face mask covers the mouth. The design suggests to combine the visualisation with breathing sounds, that turn from peaceful breathing into coughing depending on the pollution levels.
๐ฑ
Gas mask as a visualisation for a high level of air pollution and thus a high risk. Threshold visualisation for high-level air pollution.
The design introduces pimples as a human response to air pollution. Each colour represents a specific pollutant type. The growth of the eczema is according to the level of air pollution. Further, the design plays with facial expressions such as disgust or sadness.
๐ฑ
The agent's face is segmented into three regions that each represents a threshold of air pollution. It is exemplified for exhaust gas. Low-level pollution is visualised at the bottom as green, healthy-looking plants. If the pollution rises, these plants slowly die and turn into cracks. Besides, the eyes turn red to further signalise high air pollution.
๐ฑ
The design experiments with dedicated facial regions, an environmental impact zone at the top and a human reaction zone at the bottom.
๐ย
The design explores using smiles and celebrating faces as a means to encourage physical exercise. The design further includes a pulsing face to mimic the user's heart rate and includes head shakes to signal dangerous heart rate levels.
๐ย
The design represents heart rate data as sweat, running down the agent's face depending on the level of exercise. Further, the skin is pulsating to signal the user's heart rate.
๐ย
The design uses the ''Tamagotchi'' concept to visualise user's physiological data as a means to incentivise physical exercise. An increased heart rate data functions as a way of taking care and feeding the Tamagotchi-like robot. It either grows tall, healthy and strong with sufficient physical exercise levels or stays weak and eventually dies.
๐ย
This design makes use of a "battery" metaphor to represent a user's heart rate data. It exemplifies a low heart rate, thus a low exercise level, as an empty and red battery.
๐ย
This design makes use of a "battery" metaphor to represent a user's heart rate data. It exemplifies a medium heart rate, thus a medium exercise level, as an orange battery level.
๐ย
This design makes use of a "battery" metaphor to represent a user's heart rate data. It exemplifies a high heart rate, thus a high exercise level, as a fully-charged green battery.
๐ย
The design explores a candy treat as an incentive to exercise. The treat gets bigger the more the user exercises, thus the higher the heart rate.
๐ฃ๏ธ
This design explores a clock projection onto the agent's face to convey the time of day. Depending on the time of day, the respective quadrant lights up.
๐ย
The design uses an anatomically-correct heart icon as representation of a user's heart rate. The aim is to enhance the accuracy and realism of the visual representation, providing a more precise depiction of physiological data.
๐ฑ
The design explores a face mask as a metaphor to convey air pollution as a human response. When the air pollution level is low, the face mask is still worn on the face but is pulled aside, allowing the agent to uncover the mouth and to breathe freely.
๐ย
The design visualises heart rate data by displaying a progressively tired and aged face, with wrinkles as a sign of ageing. A lack of physical exercise makes the agent appear more exhausted. Conversely, increased physical activity results in a younger, healthier appearance, accompanied by a happier and more vibrant facial expression.
Exemplified experimental setting for air pollution: participants were given a GUI to manipulate the different levels of particulate matter (PM2.5) by moving a slider. The respective information was conveyed through the embodied social agent placed next to them.
We conducted a second study with 28 participants, investigating the communication of risks related to air pollution and cholesterol. We refined and implemented our designs on a popular social agent and explored participants' risk perception, self-reflection, empathy and motivation to act in a mixed-method user study.ย