Robots are broken, but we can take care of them. In order to explore the role that art can play in innovating robot-human interaction and the agency of robots, Waag organised multiple VOJEXT S+T+ARTS collaborative exploration efforts. Which lessons can we extract from these efforts?
In the past year, Waag organized and hosted talks, interviews and co-creation workshops with artists and designers that are working with robotics and AI. Some transversal questions and concerns arose from these events, and the first and foremost one is: in order to co-learn, co-live, co-exist and co-work with robots, we need to recognise that the robots are broken, but we can take care of them. This draws from Marleen Stikker’s argument of “the internet is broken, but we can fix it”, following the observation that “shareholder value is preferred above useful technology”.
Robots are broken
This brokenness can point to the physical state of robots—as artist Sanela Jahić mentioned in her interview, robots have their boundaries too—she gave an example of a programmer saying that the robot was forced to work for long hours that was beyond its limit. Artist Justine Emard also remarked about the fragility of technology in her talk, as she has worked with an AI that turned out to be too strong for itself.
The brokenness also refers to the human’s conception and perception of robotics in the human-robot relations, and that of workers who work with robots in factories. Justine Emard emphasised that questions relating to machines being able to have mental breakdowns and needing repair can tie into the topic of care and repair, just as you would have in relation to workers and conditions of labour.
Madeline Gannon suggested that we need to come up with ways that are not prescriptive but emergent for this co-habitation to be beneficial to everyone, and we should think about how to address the needs of those who suffer from unbalanced impact caused by technology.
The brokenness of humans and workers in terms of agency is also implied in Patrick Tresset’s robotic installation Human Study #4, La Classe (2017) in which a robot teaches the other robots to draw: the uncanny parallel between the disciplined robots and disciplined human in a class setting invites us to reflect on our relationships to rules and norms. The robots’ agency of revolting also mirrors human’s agency of refusing to be disciplined. Madeline Gannon similarly mentioned that the alienation in industrialised labour happens to both human and robots, in which one’s existence and meaning become side-lined.
As for the imagination and applications of robotics, the brokenness for instance shows in the limitedness of robots’ functions. Anna Dumitriu expressed that she and Alex May explored questions such as what is naughty, what is malfunction in their works that involved robots. In her eyes, most machines and robots tend to be fairly limited in their particular roles.
The brokenness also appears in the human-technology interactions and relationships—robots are often regarded either as mere tools or threats. Madeline Gannon, who brought industrial robots outside of the factory and let them exist as curious beings, considered giving attention is an act of generosity, and it also applied when it came from a robot.
For Gannon, the interactions between the visitors and Mimus, the giant industrial robot, were exchanges of curiosity and generosity, rather than pure information. From the co-creation workshop, some participants asked from the lens of a robot: “why do the other non-robotic workers consider me as a threat?” As for a collaboration in the setting of human-robot, Sanela Jahić thought that collaboration with humans meant to discuss about different perspectives and to move one’s own boundaries in order to combine these differences.
As for robot, Jahić remarked that humanising a robot is a proposal to replace what is missing, which would not work in the long run since it evades the problem itself. Jahić compared the robot to an organism with all these interesting, intricate parts, and one could see that in the interaction with it. “It is not a dead entity, but a materialisation of energy”.
Jahić said about her works: “It's an orchestra of motion”. In order to achieve co-learning and co-creation with humans and robots, there should be no subordination. Otherwise, “the possibility of learning or possibility of influencing is completely diminished”.
Therefore, we can say that technology is not neutral and that robots can have agency deriving from the way they are made. But whose agency are we talking about?
Robots and agency? Whose agency?
In the Q&A session of the artists talks, an audience member asked whether robots have agency since they are programmed. It seems that the general assumption is that they have no agency at all. However, artists’ practices and visions challenge this assumption.
In Justine Emard’s work Co(AI)xistence (2017), the human-robot improvisational dance is beautiful and it makes the audience to feel that there is life inside this partly humanoid robot. In response to the question asked by this audience, Emard thought that her practices were collaborations with different forms of life including robots. In her works, the robots were programmed to not to be programmed, which meant that they could not be predicted. “Madeline Gannon pointed out the complexity of infrastructure stacks behind the robotics. We are creating software that creates itself, which is beyond the control and prediction of human.
Jan Hein Hoogstad tried to define AI as agency intelligence. For Hoogstad, agency is always contextual. Agency is defining values both in the more economic sense of the word, and the more moral sense of the word. “If you manage to create value within the values of a context, then you have defined an agent”.
Following this reasoning, if both humans and technology manage to create values within a context, then the collaboration is not an unequal relationship. In terms of AI learning loop and human’s self-learning loop, Hoogstad considered that if AI challenges both human and itself in order to learn, then these two learning loops come together and the collective agency increases.
Jesse Howard: Hacking Households
Agency also tightly connects to the ability to act. In terms of agency or artistic agency, as a teacher, Frank Kolkman aimed at nurturing curiosity, nurturing skills to engage with some of these technologies that they are presented with, and enabling spaces for critical reflection.
Through their work and vision, these artists seem to argue for a view in which both humans and robots can have agency if they have a critical understanding of not only the technological interface but also of an underlying system. Designer Jesse Howard for example considered the act of design as thinking about a complex system, and then finding places in the system where it could act. It allows someone to take the agency a bit, or to take a position and be able to act from that position.
So how to solve the problems regarding brokenness and agency? Collaborative and art-driven innovation can be a solution!
Collaborative and art-driven innovation
For both designers and end users, it can help them to make informed decisions if they understand the logic is of the systems and can retrace the decision-making processes. This will decrease the feeling of becoming subjected to the machine. Regarding his practices in relation to robots, Frank Kolkman thought that trust is important in order to facilitate collaborative robotics.
For Kolkman, technologies such as artificial intelligence are currently very much black boxes. Even the people that develop them often don't really have a very clear understanding of the technologies, especially in deep learning situations. “If you're presented with just an outcome, but not a roadmap. How do you trust it?”.
Having previously worked in a textile factory, Sanela Jahić shared a similar view. She thinks that textile workers would have a higher feeling of fulfilment and accomplishment in the setting of a factory, if they are beforehand informed about changes in investment plans or in the production process. The idea behind this is that it allows workers to think along and give input on new, different solutions. The idea of involving workers is also applicable in the factory setting where workers need to work with robots: by involving workers into the design process of robots, their sense of ownership over their work and workspace will be enhanced.
In the co-creation workshops, participants also advocated that workers should be co-designers of robots: in the process of robotization, decision making should not only focus on efficiency and productivity, but involve workers and ask them: what kind of technology or robot do you think would be beneficial?
The trust issue is bilateral since robots also have agency. Imagining themselves as robots, the workshop participants discussed the issue of trust: “I'm not as complex as a human, nor as simple as a tool. Do I trust my programming or advice from my colleague? Who do I trust more, the worker that I work with, the director, or the technologist that developed me?”.
Robotisation and technological innovation should also take into account inclusivity and bottom-up collectivity. The second group of workshop participants asked: “are robots making the workplace more inclusive for people with disabilities? Is the worker safe? Can we diversify movements/actions in the workplace? Is the robot designed and tested to collaborate with workers with different conditions?” They also raised the question of care for malfunctional robots, and whether human disability can become a point of departure for the development of robots and their behaviour.
Moreover, inclusivity implies the appearance and operation of robots that are not based on certain biases or set ideas that restrict people’s feeling and perceptions about them. This requires openness and creativity.
Anna Dumitriu and Alex May said: “All these choices that we make about how robots look and how they operate, are designed to create a certain ambiguity that allows people to not just see what this robot is doing, but how they [the humans] feel about it. It allows them an opportunity to question that”. Their work creates a space where people's ideas and preconceptions are allowed to be manifest and reinforced. For Frank Kolkman, this open and creative experiment with robots can be achieved through art, since artists can claim their artistic freedom and to use exhibition spaces or musea as a place to conduct a public experiment.
As for concrete suggestions on what it means for robots to be inclusive and supportive, the co-creation workshop participants have made these points: firstly, technologies should be flexible and adaptive, which means that technologies should be non-intrusive but adjustable to the workers needs and replicable to different industrial applications. Secondly, technologies should be supportive, collaborative, and inclusive to the workers. Inclusive meaning: to all kinds of workers and bodies, affordable, presented as multi-coloured and gender neutral, where both workers and robots are autonomous.
Thirdly, humanised and embodied operations: the robot could have a “voice” like providing suggestions, information and jokes, and also communicate through other forms such as haptics, movements, emotions and gestures. Lastly, robots should have environmental responsiveness and connectedness, meaning that they are connected to internet and data platforms and can therefore learn to cope with unexpected things and be aware of the workers.
Robots are people, too
If there’s one thing to conclude from these three events, it is that robotic technology is not neutral. Robot technology needs care and reimagination from designers and artists. Let’s open up the vision that robots are merely tools, designed to assist people. They are people too in the sense that they are fragile, they have limits, and they have their own agency.
In order to have a more collaborative innovation about robotisation in factories, an artistic and collaborative approach can help to create a framework in which workers are more involved in the design of their robotic colleagues. They can help create robots that are more inclusive and supportive to all kinds of human workers. The development of automation could benefit from a more ambiguous and non-utilitarian approach, which can open up new possibilities for robot-human interaction.
Want to know more?
- Would you like to learn more about collaborative robotics? Tune in during Robots Are People, Too at Ars Electronica Festival 2021.
- Are you a fan of media art? Ars Electronica Festival organizes hundreds of online sessions of conferences, talks, concerts, workshops, exhibitions and guided tours. Interested? Register here.
- Next to the event Robots Are People Too, Waag participates in many other activities at Ars Electronica 2021. Find out more in this overview.
