Swiss artist Matthieu Cherubini was kind enough to share some his thoughts and process behind the social bot rep.licants.
rep.licants.org is a service allowing users to install an artificial intelligence (bot) on their Facebook and/or Twitter account. From keywords, content analysis and activity analysis, the bot attempts to simulate the activity of the user, to improve it by feeding his account and to create new contacts with other users.
The experience of an enhanced virtual self as users are invited to install a bot on ther favorite social network account and become a replicant. Provided with “virtual prothesis for the social media introvert”, people who use the service have started to uncover what it means to automate social interactions.
GH: You are an artist! Why did you start working on rep.licants? Is it similar to your other work?
MC: rep.licants is the work that I did for my master thesis. During my studies, I developed an interest about the way most of people are using social networks but also the differences in between someone real identity and his digital one. I do not have a big experience about creating personal projects, I began seriously to do it during the past 2 years, so previous to rep.licants I did two other “serious” personal projects and they were related to this thematic aswell. One The Pursuit of Happiness is about hacking into Facebook account of random users in order to steal their private messages for seeing what they were looking for on this social-network. The second one Afghan War Diary is about linking data coming from Counter Strikes servers and Wikileaks and displaying the result on Google Earth.
Back to rep.licants – when I began to think about a project for my master thesis, I really wanted to work on those two thematics (mix in between digital and real identity and a kind of study about how users are using social networks). With the aim to raise discussions about those two thematics.
GH: What was the process like for you?
MC: At first I just had the idea about creating a webservice where people could subscribe on it and mix themselves with a robot. But I really didn’t know where I was going exactly. So I began the project as an experiment.
The first step was to study what people could do on Facebook (I began with Facebook only) and how a bot could reproduce those same actions by linking himself to other services that its user is using or by getting new informations on other sources.
[img_assist|nid=3969|title=|desc=This schema roughly demonstrates how the bot is working.|link=node|align=left|width=400|height=142]
After I programmed a version of the bot, according to this previous study, with few functions and I asked to some of my teachers, tutors and classmates if they wanted to be volunteers for this experiment…I had 3 volunteers who did the experiment for like 4 months. I was asking them weekly what the bot was doing, if they were feeling the bot was lacking of something important, … During this 4 months I redesigned a bot for Facebook by taking into account the feedbacks of my three volunteers.
In same time I also decided to do a version of the bot on Twitter and the process was almost the same as Facebook.
GH: What have people’s responses been? Have Facebook or Twitter responded at all?
MC: The responses has been very mixed some are over negatives and some are very positive ! I’m happy about that because it’s very interesting: the negative responses are mainly from people who were thinking rep.licants is a real and serious webservice which is giving for free performant bots who are able to almost perfectly replicate the user. And if they are expecting that I understand their disappointment because my bot is far from being performant ! Some were negatives because people were thinking it is kind of scary asking a bot to manage your own digital identity so they rejected the idea.
For the positive responses it’s mainly people who understood that rep.licants is not about giving performant bots but is more like an experiment (and also a kind of critics about how most of the users are using social networks) where users can mix themselves with a bot and see what is happening. Because even if my bots are crap they can be, sometimes, surprising.
But I was kind of surprised that so many people would really expect to have a real bot to manage their social networks account.
Twitter never responded and Facebook responded by banning, three times already, my Facebook applications which is managing and running all the Facebook’s bots.
GH: How do people use it? Have there been any interesting stories of how people have used rep.licants?
MC: For what I know and after some questions/feedbacks received by users, I would say that some people use the bot:
a. Just as an experiment, they want to see what the bot can do and if the bot can really improve their virtual social influences. Or users experimenting how long they could keep a bot on their account without their friends noticing it’s runt by a bot.
b. I saw few time inside my database which stores informations about the users that some of them have a twitter name like “renthouseUSA”, so I guess they are using rep.licants for getting a presence on social networks without managing anything and as a commercial goal.
c. This is a feedback that I had a lot of time and it is the reason why I am using rep.licants on my own twitter account: If you are precise with the keywords that you give to the bot, it will sometimes find very interesting content related to your interest. My bot made me discover a lot of interesting things, by posting them on Twitter, that I wouldn’t never find without him. New informations are coming so fast and in so big quantities that it becomes really difficult to deal with that. For example just on Twitter I follow 80 persons (which is not a lot) all of those persons that I follow is because I know that they might tweet interesting stuffs related to my interests. But I have maybe 10 of those 80 followers who are tweeting quiet a lot (maybe 1-2 tweet per hour) and as I check my twitter feed only one time per day I sometimes loose more than one hour to find interesting tweets in the amount of tweets that my 80 persons posted. And this is only for Twitter ! I really think that we need more and more personal robots for filtering information for us. And this is a very positive point I found about having a bot that I could never imagine when I was beginning my project.
GH: Have there been any interesting disasters or failures in the interactions as a result? Or any surprising bugs?
MC: One surprising bugs was when the Twitter’s bots began to speak to themselves. It’s maybe boring for some users to see their own account speak to itself one time per day but when I discovered the bug I found it very funny. So I decided to keep that bug !
GH: My own experience with rep.licants revealed to me an aesthetic of antagonism. What does an aesthetics of antagonism mean to you?
MC: I’m not sure but maybe something which is hostile in my project ? Or maybe the way the bots are running ? As they are very buggy, they do a lot of things which could be opposite. One time they could find a content which is absolutely match to the user but the next one is absolutely opposite of what the user is or like.
One of the things I’ve discovered during my own use of a rep.licants Twitter bot is that it likes to retweet messages I’ve exchanged with an acquaintance – sometimes even the same mmessage more than once. This has a somewhat awkward effect of bringing attention to that interaction when it wasn’t really warranted.
Around the same time Matthieu and I conducted the interview, this video of a chatbot having a conversation with itself went viral – perhaps in part because the conversation immediately turned towards more existentialist questions and responses. The conversation was recorded at the Cornell Creative Machines Lab, where the faculty are researching how to make helper bots.
The best part of the video happens when one chatbot implores that he is not a robot – but rather a unicorn. How the bot determined that is not widely known, but it does invoke an important visual element and narrative theme from Ridley Scott’s Blade Runner – where the main character Deckard (played by Harrison Ford) has dreams of unicorns. The character of Gaff (played by Edward james Olmos) is also seen making origami unicorns – an apparent reference to his knowledge that Deckard is replicant.
The questions that rep.licants poses are deep human and social ones – laced with uncertainties about the kinds of interactions we count as normal and the responsibilities we owe to ourselves and each other. Seeing these bots carry out conversations with themselves and with human counterparts (much less other non-human counterparts) allows us to take tradition social and technological research into a different territory – asking not only what it means to be human – but also what it means to be non-human.
Design is a sticky practice. It is looped with contradictions, uncertainties, and material constraints. Bringing something new into the world, be it an artifact or service raises challenges that few individuals can surmount – if at all. Despite the dominant view that geniuses, visionaries, and otherwise crafty individuals are solely responsible for designed creations, organizations play a far greater and often unattributed role. Perhaps it is because of the aesthetic flair worked into the surface of the object or experience, or maybe it’s the personality of the driving individual that points us in the direction of these myths. And they are myths, because even the most brilliant designer owes their success at the end of the day to at least one group – their participants, their users. More likely is “rock-star” designers owe the production of a product or service to many more who inhabit a long chain in the process of design, implementation, and distribution.
Diego Rivera's "Detroit Industry, South Wall"
Somewhere along the chain of causation between creative individuals and their users there exists a group of people, places, ideas, and things that operate synchronistically and synergistically to develop ideas into concepts, concepts into prototypes, prototypes into experiences, experiences into practices, and practices into lessons. These sets of translations encompass different skill sets and relationships, few of which are possible without deep and varied interactions across different environments.
Taking stock of an emerging design practice is something we do often these days. I think it springs from places that have recognized and internalized failures for what they are – opportunities – and from people who embrace reflection as positive forces for learning and adaptive change.
Our environments are changing. And they will continue to do so. Even if we find pathways to design static landscapes that include fixed social interactions, the resources and habitat available to us and other species will remain in flux. Consider that in 2008, we reached the threshold where 50% of the world’s human population resides in urban dwellings (and possibly also 50% of the world’s population of cockroaches, starlings, street dogs, and sewer rats).
It’s also true that the biosphere can no longer be considered ‘natural’ in the same terms that 18th century Romantics did, as something pure, something to be conserved, something separate. The landscapes of our contemporary experience are human enmeshed – neither dominated nor resistant to our desires to interact, to use, and to understand. They show our preferences for stable communities supported by agriculture that reinforce a growing feedback loop between population growth and energy consumption. The Anthropocene, as this epoch is now commonly referred to, places a point on some linear timeline where people demonstrated their best applications of the idea of progress. Perhaps it is only our external concept of the sublime that are disappearing from the human range of experience.
There is much greater landscape diversity than has ever existed, but certainly it is less inhabitable by the majority of the world biological diversity. Landscape diversity is created not only by people and their continued interpretations of “safe” and “prosperous”, but also by animals and plants that push and get pushed into their own new and divergent niches. Patches of materials are being collected and redistributed to form wild hybrids and pure spaces– bacteria-resistant surfaces, show rose gardens, crude oil-slicked sandy beaches, tourist-friendly rainforest, wildlife mobility solutions, skyscraper concrete pillars, semiconductors, and extra-terrestrial orbiting robots – to name just a few.
Each time new patches are created, they exemplify the desires and possibilities available for their inhabitants. They provide food, space for living, courses for exercise, obstacles for navigation, challenges and threats between groups that aim to occupy more patches, places to hide, and places to trade. Evolutionary history has demonstrated that cooperation confers a significant strategic advantage to those who choose to communicate, share, and build together. In human terms, one need only look at the migratory patterns of individuals from rural to urban settlements to understand that there is a direct and perceived economic advantage from sharing land, resources, infrastructure, and culture on people’s livelihoods – not to mention social mobility.
Detail from wall illustration at the Golden Temple, Namdroling Monastery, India
Design practices are widening. They are gaining breadth proportional to their influence on economic productivity, their ability to expand social engagement and political empowerment, and perhaps because of the impact that social studies of science and technology has provided to our appreciation of artifacts as catalysts for knowledge. Scientists and technologists are viewed as inventors, individual carries of the modern ideal of progress. We now recognize that images, laboratory spaces, institutions, public media, and mechanical parts play as significant a role in chance events, innovation, and the acquisition of scientific and technological dogma by civil society.
One of the implications of an expanded design practice is the gradual inclusion of organizations as ‘objects’ for design. Organizations were once the purview of managers, business executives, policy makers, and human resources consultants, but they can now be confidently lumped together with paint, plaster, and photo emulsion.
I’m sure this is raising red flags for some who read this, and it should. It’s a scary proposition for some to think that individual behaviors can and should be designed and organized. But it is a fact that individual and group behaviors are already structured by the designed and so-called natural environment along with normalized social interactions and perceptions of social agency. The only thing we gain by ignoring the structures that are already in place (albeit unconsciously) is the freedom from self-awareness, individual and collective agency to solve more challenging and complex problems. The more we ignore these unconscious behaviors (eating habits for example) that already exist, the more they leads us into deep patterns and habits that can be difficult to get out of for reasons of fear, inexperience, ability, or just a lack of awareness.
This is not to say it is all negative. If we had to pay attention to everything we did, we would fall apart from exhaustion while trying to make complicated decisions. Many of our biases may have developed because they habituate us into safe spaces for interaction. Unfortunately, as our societies and environments change, those safe spaces may be retreating, and it’s worth reflecting on our biases and how our individual and group dynamics promote infrastructures for flourishing.
Organizational management has become a major discipline of the 20th century with the adoption of increasingly complicated tasks and industrial processes. It stands to become more integrated into our systems and psyche, but will management theories dominate – or will design envelop management in favor more distributed processes of self-organization consistent with cybernetics and decision theory?
Groups change, and so do their goals. It is a part of life and society, and it always will be. The questions that we ought to be asking is how, where, through whom, and when do they change?
There is ample evidence that organizational behavior is at the root of innovation and robustness across enterprises. The shape and tenor of a group of people, each with different tasks, and working towards a common goal varies widely – not to mention the tasks, people and goals – and that’s assuming those goals are shared among the group members! Without going into the theory and practice of organizational behavior for which there is a massive literature, I simply want to raise the point that organizational design may be a more recent practice and one that plays a role in or strategies for adaptation, sustainability, and inclusive growth.
In part II, I’ll look at some examples where designers are tackling organizational design as project and process.
This post from frog design reminded me of a short presentation I attended a couple of year’s back on Maslow’s Hierarchy. I have to admit that I forgot what the main point was that was being made, but I think it had something to do with design and how attention to this classification could help designers with the “important” things. There is something about Maslow’s pyramid that fits quite nicely with Indian social philosophy. Perhaps it is the hierarchy part – or maybe the path to enlightenment.
All I really remember is that I was a little frustrated by the talk, and I made some sketches to explain my unease (recreated here).
In India (where I live), I often see some obvious trade-offs between one person’s self-actualization and another’s basic needs. This is true most everywhere, but luckily in India, many more of these trade-offs are visible and not isolated or placed elsewhere – although that is changing.
The basic maneuver of my sketch inverts Maslow’s pyramid and adds another. I think it’s somewhat useless to consider an individual in isolation – which is why Maslow’s hierarchy has a certain irony in India – very little exists in isolation. In a social setting, this inversion starts to get you somewhere near the interactions one experiences if everyday life. One person’s need for security means that another person has to sit and guard the front of a building all day – with little in the way of engaging, goal-oriented work. Likewise, the more “enlightened” one becomes, the less they feel pressured to advocate for material wealth for themselves or others.
So my comment would be that I now see Maslow’s hierarchy more as an interaction diagram. Drawing it out this way makes me also think that perhaps that brown middle band – being part of the group – is where the pyramids pivot and heterarchies begin. But that’s just me reasoning from sketches….
UPDATE: I’ve been looking at food pyramids lately, and found this inverted food pyramid to show a comparison between diet choices and environmental cost. thx GOOD
The image above was the first draft. This is the second. Thanks to Aliya for good, perceptive comments.
attachmentModel_v2
Premises:
Culture as the processes that allow the uptake of processes, procedures, information, beliefs, values and social norms.
Cultural affiliations are attachments.
Attachments and reattachments are limited (quantity) and constrained (quality) by pressures.
Aspiration is a cultural step in creating capability.
Based in part on: Appadurai, A., 2004, ‘The Capacity to Aspire: Culture and the Terms of Recognition’, in Rao, V. and Walton, M., (eds.) Culture and Public Action, Stanford University Press, Palo Alto, California, pp 59-84.
Daniela Plewe’s discussion brings me back to some thoughts and notes I made about Marcel Duchamp’s Coefficient d’Art. Duchamp described it as:
“An arithmetical relation between the unexpressed but intended and the unintentionally expressed.”
It is intended to describe the difference between what artists intend and what the spectator perceives. For Duchamp, this difference is in the act of communication or transaction, where certain differences and attributions of value are made out of the interaction among individuals. It this coefficient that structures the viewers engagement with artifacts and allows them opportunities to appropriate objects to their own needs and ends.
For Duchamp, the coefficient of art could be good (+), bad (-) or indifferent (=), but the sign of the coefficient had no bearing on the effectiveness of the work itself–only the difference between the agency of the artists to produce a desired effect in the minds of the spectators. The effect itself is up for further negotiation between them.
Mutual information is a similar concept to the coefficient of art, but it comes from information theory and describes the amount of information one thing tells about another thing. In other words, it is the reduction in uncertainty of one thing due to knowledge of another. If we ask how information (and consequently, meaning) is shared between different sources of uncertainty (like an object and a spectator or an object and its artist), we may be able to get a sense of how they are connected and how they might respond to each other.
Mutual information is helpful as a concept because we want to understand how interactions vary with one another–i.e. how interaction values may/may not change as a result of signals, actions, and assumptions.
A component of mutual information is information entropy. Entropy is a measure of uncertainty associated with a variable and quantifies the information contained in a message. It is similar to the coefficient of art; it may describe the uncertainty associated with an artwork as judged by the spectator. Conversely, it could describe the absence of meaning when one does not know the value of the work. Likewise the spectator may themselves exhibit high entropy (high uncertainty) relative to the artist if the artist knows little about the spectator and how they will perceive the artwork….at least that’s how I think it would go.
The coefficient of art is a compelling concept. It suggests that that art has an effect, and if an effect–value in context. Describing that value is very close to the describing what difference the work of art makes, either to the spectator or some chain extending through them.
Borrowing from evolutionary and network theory, one could pull in a set of relationships between interacting agents that describe how networks evolve and persist. Relationships endure over time from the benefits of interaction. In network reciprocity, entities pay a cost, c, while their number of neighbors, k, receive a benefit, b. If b/c > k, where the ratio of benefits to costs is greater than the sum of neighbors, the network persists because its members are gaining as a result of their interactions.
Duchamp’s coefficient of art (hereafter described using the greek letter psi, ϕ; see also: epistasis), approximates the number of neighbors, but as indicated by it separation from the actual effect of the work itself, says nothing about costs and benefits. ϕ approximates k, or rather the reciprocal of k, because as the number of neighbors (or spectators of the work) increases, the likely ability of the artwork to communicate intent, decreases. This is because of variation among the spectators who may either not be well-understood by the artist or who are perceiving differently or because the artist. Interestingly, ϕ always assumes artistic intent. If ϕ is low, it may be the ‘fault’ of the spectator, the inability of the artist to realize that intent, or of some other intervening factor.
But what about art that is created beyond intent such as generative, algorithmic, or emergent artworks?
ϕ may also be a bound on the ability of artifacts to bridge social groups, as in the case of boundary objects that have multiple uses. The intent of the maker of that object is only partially achieved, but may clearly be appropriated to serve other purposes. Here we might similarly invoke a coefficient of use–or a measure of intent in use that transforms the intent of the artist.
Far from achieving certainty, at least the idea of ϕ, of a coefficient of art, starts to unlock more questions about translation and meaning between objects and people–and of the directionality of interactions between people.
The CEMA homepage is showing an image of scanner that has opportunistically been colonized by ants (anyone know which species?). I was present at the offending attack, and I have this to say. I didn’t see it so much as an attack as it was (more perversely) an underanticipated observation that ants had quietly moved into an (apparently) unused and undisturbed piece of late 20th century technology- that of the document scanner.
While this may have been felt by some as an attack on our morals of human-hood and right-living (ants and scanners shouldn’t mix, right…er…right?), to me this was much more the most delicate and profound expression not of nature but of the social world in which we live. The most amazing thing to me is that a colony of ants could have arrived and decided that a scanner would make a good home. Perhaps there were some legacy muffins adding allure to the crystal glass and step-motor, but maybe the ants were looking for something held up in the ambient waves of electrical heat left over from un-nourished scans of students’ faces, buttocks, book chapters, and collages.
No..I think this is exactly where we want to be…where mixes and happenstances converge out of nothing more than the desire to find place, continence in the “other”, and the cheap thrill of being where you aren’t supposed to.
On checking up on their status, they are gone from the scanner…pupae and all. I’m not sure if they left on their own accord or if they were kicked out. Where did they go? The water cooler perhaps? As for next time, I’m keeping my fingers crossed that discovery doesn’t correlate with disentanglement. I’d like to keep my scanner ants…who knows…they may have figured out something that we haven’t.
Ok, the pic isn’t great but you get the idea. This is “The Cube” at the MIT Media Lab. I visited the Lifelong Kindergarten group there and saw how their close proximity to tools, shared workspaces, and each other facilitated their work in progress. I really liked how the space was large with high ceilings, that it was a mess of projects, and that there was a table where lab members would work individually with a tacit sociality.
Most of what I liked about the GROCS lab at the University of Michigan Media Union was how the activities of various groups, classes and projects were in an open, shared space. Mix that with some movable furniture and a close proximity to other resources in the Media Union and it made for a good space to come up with ideas, share them and work them out with each other.
I was up this morning thinking about the kinds of spaces, communities and interactions I would like to see. Somewhere between physical computing, synthetic biology, evolutionary ecology, and design is a space where species can speak and be recognized by each other, where urban infrastructure becomes adaptive in the space of days and not decades, where the threshold of difference is lowered to such a degree that new networks between otherwise unrelated groups and individuals can find common ground.
Perhaps for the first time, I am beginning to see how things can be connected for the purpose of builing empathy. Whereas previously, I think the difficult work of etting to know a species was largely out of many peoples’ desires and time banks, perhaps there are now ways of making the opportunities both immediate and resource-efficient.
Rather than always seeking to decouple tightly-linked host-parasite relationships, can we find ways to make new ones…perhaps ones that can grow into mutualisms and symbioses? Is hardwiring a step in the process? What are the costs, benefits, sources and sinks? Can we create or link networks of co-dependence? What models of covariation should we adopt: linear, dominance, epistatic, topological?
This semester I have the pleasure of being able to lead and help two teams of students create engaging, socially-embedded, interactive design projects. The experience was a success both for me and the students. I learned a lot about my students and what they needed to do excellent work. I think we also found some new ways of working here at Srishti that may prove valuable in the long-term.
The teams also took part in a competition in which the winning team is invited to present their work at the Microsoft Research Design Expo, part of the Faculty Summit held in Redmond, WA in late July. We’re all looking forward to attending because we are very proud of the students’ accomplishments.
The ‘Moon Vehicle’ project consisted of a system to create interactive storytelling experiences around themes of the moon, space exploration and colonization, and India’s forthcoming launch of the Chandrayaan-I moon satellite.
Screen captures from the \'Moon Vehicle\' project design.
The Moon Vehicle team’s design developed in part from the Bangalore Space and Culture Initiative, an interdisciplinary endeavor of artists, scientists, designers, and technologists that began in late September, 2007 and coordinated by Srishti, NIAS, and ISRO.
The Play Revolution project changed many times, but it was always focused on the idea of building a socio-econo-technical system for improving the knowledge-networking opportunities of children living in slums in and around Bangalore.
The lab itself and the social interactions were influenced in part by the GROCS lab at the University of Michigan. Thanks go to Linda Kendall-Knox for her willingness to share aspects of their process.
The course started as a relatively straightforward user interface design series of topics, but this plan was quickly abandoned for a more socially-embedded model that would adapt to the different concerns and questions we were going to encounter. The primary article guiding this process was entitled “Products and Practices: Selected Concepts from Science and Technology Studies and from Social Theories of Consumption and Practice” (Ingram et al. 2007). The article stressed six stages of technological adoption: acquisition, scripting, appropriation, assembly, normalization, and practice.
We used these stages to guide our design process.
The students were given a design brief that consisted of two challenges: one consisting of Srishti’s existing commitments to cultural, educational, artistic,and design-based engagements with society, and another consisting of a more general challenge to design a user interface and/or interactive experience around the theme of learning and education. They were asked to develop a project that synthesized these challenges into one unique approach that incorporated the concerns, commitments, and constraints that were implicitly and explicitly embedded in the issues raised.
The theme of this year’s competition was “Learning and Education”, and students were challenged to design a user interface and/or interactive experience around the theme of learning and education that improves the daily life of a wide variety of users through learning and education, promotes creativity and curiosity in new topics, demonstrates novel ways of providing instruction, and rethinks education systems and tools.
