“The Misbehavior of Animated Objects”: how to design an object with an attitude? How to give them a personnality, or to simulate their free will? Such was the goal of the workshop we contributed to organise for TEI 2014.
“The Misbehavior of Animated Objects” workshop and the MisB Kit was developed for the TEI 2014 conference in relationship to the MIT Medialab–Tangible Media Group; initiated by EnsadLab/ Reflective Interaction (program under the direction of Samuel Bianchini), with Emanuele Quinz, Cécile Bucher, Benoît Verjat and Alexandre Saunier; and EnsadLab/ Sociable Media (Program under the direction of Rémy Bourganel), with Émeline Brulé and Max Mollon; in collaboration with Didier Bouchon, Martin Gautron; with the participation of Jean-Baptiste Labrune and Nicolas Nova; and with the support of the Bettencourt Schueller Foundation.
“TEI 2014 is the eighth international conference dedicated to presenting the latest results in tangible, embedded, and embodied interaction. It will be held 16th to 19th February 2014 at the Ludwig-Maximilians-Universität München (LMU), Munich, Germany.”
The work presented at TEI addresses HCI issues, design, interactive art, user experience, tools and technologies, with a strong focus on how computing can bridge atoms and bits into cohesive interactive systems. The intimate size of this single-track conference provides a unique forum for exchanging ideas and presenting innovative work through talks, interactive exhibits, demos, hands-on studios, posters, art installations and performances.
TEI is sponsored by the ACM SIGCHI.”
(Quick Access: http://misbehaviour.ensadlab.fr/)
The 8th TEI conference held 10 different workshops on Sunday the 16th, the day before the conference. The one we (EnsadLab) ran in collaboration with the MIT Media Lab Tangible Media Group was named “The misbehavior of animated objects”:
“How to create & animate an object of simple, abstract form which movement would confer a behavior? How to give the impression that such an object have a personality allowing to be proactive, with self-motivated behavior, not directly responding to our expectations, or even challenging them through the demonstration of mis-behavior? To address these questions, we propose a studio allowing to design and rapid-prototype such object from an easily accessible toolbox (hardware and software) for the studio’s audience to experiment with. This toolbox will have been developed beforehand through the collaboration of MIT Tangible Media Group and the Ensadlab composed of researchers in art, design, technology (HCI and tangible media) for this specific studio experiment.”
The workshop was an important step that allowed to make public the progress of our research program those last months, to physically meet our collaborators from MIT MediaLab TMG and to gather feedbacks from the participants.
It was organised as follow:
- Introduction to the context, state of art and brief
- Introduction to Hiroshi Ishii’s “vision” of a world constituted of programmable matter
- Presentation of the toolkit, both software and hardware
Three main activities:
- Discovery of the hardware toolkit and its animation by midi-controller. Each team was asked to fast-prototype three behaviours (agressiveness, shyness, lazyness)
- Scoring of a behaviour, and constitution of a library of movements, using recordings of the motors
- Prototyping of a behavioral object using the entire toolkit and sensors
All projects were presented during a 1min demo and a 1min talk.
Participants were also asked to write down their definition of misbehaviour, at the beginning and the end of the workshop.
To be noted: the workshop explores the first interaction with such objects, however our project continues with a second step, considering the relationship over time with these entities.
The toolkit is presented in a suitcase containing:
- a set of structural elements (velcro, … aluminium, K’nex, a variety of connectors)
- a collection of dynamixel motors, in different settings
- a midi-controller
- various tools, such as scissors, velcro stickers…
- identifying a typology of object behaviours
- a collective definition of a misbehaviour
MisB KIT has been initiated by the whole team for the project The Behaviour of Things, coordinated by Emanuele Quinz for the Labex Arts-H2H, under the direction of Samuel Bianchini. It is now developed by a team composed of Didier Bouchon, Cécile Bucher, Martin Gautron, Benoît Verjat and Alexandre Saunier from the EnsadLab program Reflective Interaction.
All MisB KIT components, hardware as software, are under LGPL License, excepted for the Bioloid proprietary modules (Dynamixels) and the K’nex elements used as structural elements.
Looking back to the workshop
Before drawing any conclusion, we would like to report participants’ reactions and our first thoughts on the workshop.
Overall, the workshop ran well and was dynamic. Participants quickly assimilated the technical and physical constraints of the toolkit and began to create animated objects.
The participants’ productions were diverse, from raw animated structures to fine tuned movements and sophisticated objects, and correspond to categories we identified in our preparation phase:
- Object fixed to a location VS Object moving on trajectories
- Inner movements VS External movements
- Physically disabled (doing wrong) VS Socially mis-behaving (being wrong)
Though we encouraged participants to avoid them, the following categories showed-up anyway:
- Existing functional objects (we asked to start with abstract forms/objects)
- Anthropomorphic and zoomorphic objects
- Object—to—object communication (we asked to focus on Human—object communication or Human—object—human communication)
Relying on participants feedbacks, improvements could be brought through:
- a better introduction of the “writing a score” activity;
- a score more adapted to the broad range of behaviours (the first draft being a proposal for linear movements)
- running the workshop over two days;
- running a brainstorming session on scenarios and applications of misbehaving objects.
A work-in-progress: the definition of misbehaviours
Among many topics, the question of the “relation to a good-behaviour” was discussed. The first model that arose from participants productions was the contrast between behaviors: first the good one and then the surprising misbehavioral one.
As expected, the repetition issue was discussed: A misbehaviour might become a simple behaviour if repeated changelessly over time. For others, such a repetition was not a considered a problem as an object can be compared to other similar objects providing us with a reference to assert whether or not it shows a misbehavior.
Furthermore, the notion of misbehavior is also context-dependent. For exemple, what should we think of a dust bin which doesn’t want you to throw away a paper as it should go to the recycle bin?
Or what about a “free-candy” machine that slaps your finger if designed by someone who seeks and likes that ?
Therefore we believe that different relationships can emerge from a same interaction. As said before, the workshop only explored the first interaction with such objects, considering the relationship over time will come in a second time.
ON BEHAVIOUR AND FREE WILL
Reflecting on the discussions that took place during the workshop, the following observation came up: defining the concept of behavior of an object is problematic.
What comes naturally out of the workshop proposal (expressing behaviours) are objects designed in conformity to the given behaviour: bags trying to run away from you or spheres mimicking the movement of the mouth. Objects embodying human behaviors.
The half sphere was designed otherwise. Participants equipped it with motors, tried to make it move and then optimized the movements. As a result it could switch from its spherical side to its plane one and move in different fashions. Though they described such movements as the expression of a human-like misbehavior, some kind of suicidal half-sphere begging for help, they later admitted they made it up for the demo and didn’t have it in mind when designing the object. Nevertheless spectators accepted that story and had empathy for this object.
What comes out of this example? First of all the behavior is dependent on the context. Here the storyteller made up the suicidal behavior by describing it before showing the object. Were there no beforehand explanation the half-sphere could have easily been perceived as trying to save itself (just imagine a turtle on its back trying to get back on its feet). Nevertheless no one brought it up.
Among the six objects demonstrated, the half-sphere is the only one whose behavior purely emanated from its internal elements, constraints… In short: an object-like or object-o-centric behavior. This lead to central questions about behavioral objects: can objects express a behavior which would not be programmed as such by their designers? If so, what kind of a behavior is it? And further than that, how can we initiate a relationship with an object expressing an object-o-centric behavior, freed of anthropomorphisms like aggressiveness or fear, which means a behavior expressed in the paradigm of the object and not in our human paradigm? In short: if there was to emerge a self-awarenesss in non-biological entities, what would it be?
Those issues are pretty central as they force us to question the interaction we seek with objects. Do we want them to be human derivates or spin-offs, objects that could act – up to a certain point – as humans? Or can we consider them as beings, with their own behaviors and mechanism, with whom we could have a relation based on the elaboration of a common “grammar” taking into account both the human’s and the object’s paradigm?
Objects are defined by Simondon as “crystallised human” (in Du mode d’existence des objets techniques, 1958). He argues for a society built around both human and objects, without domination from one on the other.
Exploring object-o-centric behavior and their relation to humans might be the most promising path to follow as it forces to consider objects for what they are and not for what we expect them to be. It doesn’t consider behavioral objects as the sole result of algorithms or artificial intelligences, it also takes into account their physical characteristic and allows to turn what might have been considered flaws as a potential for unexpected and self-affirming personality.
So – if objects had a voice, what would they be telling us?
• running various workshops and identifying object proposals that could be evaluated on a long term basis
• developing speculative work on objects behaviours and scenarios for a possible future including them.
THANK YOU TO ALL THE PARTICIPANTS
Bruns, Miguel TU Eindhoven
Fetter, Mirko Human-Computer Interaction Group, University of Bamberg
Schoessler Philipp, MIT MediaLab TMG
Alex, technical support
Max or Samuel, stimulate/facilitate/document
Rizzo, Tim IMG Institut für Mediengestaltung,
Epp, Felix University of Applied Sciences Darmstadt
Rasmussen, Majken Aarhus University
Daniel Tauber, MIT MediaLab TMG & RCA London
Alex, technical support
Grah, Thomas Hochschule Darmstadt
Hanke, Christoph Muthesius Kunsthochschule
Ou, Jiffei, MIT MediaLab TMG
Didier, technical support, 2 tables
Schmidt, Deborah Technische Universität Dresden, ?
Vandevelde, Cesar Ghent University, HCI/arts/sciences ?
Walmink, Wouter Exertion Games Lab, RMIT University, game lab
Niiyama, Ryuma, MIT MediaLab TMG
Cécile, technical support, 2 tables
Hamidi, Foad York University
LIU, Xin Rhode Island School of Design
Yao, Lining, MIT MediaLab TMG
Cécile, technical support, 2 tables
RELATED PROJECTS, INSPIRATIONS AND REFERENCES
• One of the first article presented at the conference was about Tangible Autonomous Interfaces and their behaviours, by Siana Nowacka. Here’s the short introduction to the talk:http://www.tei-conf.org/14/program/136.php
Max, Alex & Emeline on behalf of the misbehaviour team.