Architects and Automation

Robotic and software automation of human labour is a hot topic. Opinions vary from it causing mass unemployment and economic crisis, to it emancipating human energy from the harsh environments of production and the monotony of repetitive activity. The rapid innovation in machine learning and intelligence points towards automation extending beyond purely repetitive jobs to any role productive role, including creative practice. When the discussion around automation and ai emerged architects presumed their professional role would be safe as artificial creativity seemed like science fiction, not anymore.

Creative Agents

As computers and robots become more powerful and agile, and capable of design and documentation they will offer clients and developers a cheaper, more efficient and more manageable alternative to human labour. Research into the creative capacity of agents, or autonomous computation is already 15 years in the making meaning that potentially designers days are numbered. Fear not fellow architects, this is an opportunity to become unshackled from AutoCAD and Revit, think of this as a chance to outsource the crap you hate, bathroom details… robot, door schedules… robot, environmental impact assessment reports… actually even the robot might find this too much. Rather than thinking of the future take over of our robot overlords as them and us, think of it as a new opportunity to design and build, one where architects become robot owners, trainers, and mentors. A collaborative relationship with a software or hardware robot provides a new means of design, communication and production that could, as Jeremy Rifkin has described, enable architects to engage in the automated market based capitalist economy while freeing them up to participate in the emerging social economy .


A possible future was detectable at the Rob|Arch conference in March this year via weak(ish) signals. 2016’s conference involved collaborative robots (HAL Robotics & Bond University), interactive 3d printing robots (Institute of Advanced Architecture Catalunya, Harvard GSD & University of Sydney and UNSW), and real-time vision sensing generative design robots (RMIT Architectural Robotics Lab and Southern California Institute of Architecture). The first signal was that the human-robot relationship was not merely mutual, a third participant was evident, the material involved in fabrication. A three-way collaboration involving the agencies of human(s), computation/robot, and material in all the workshops showed that the previous treatment of robots as subservient machinery via deterministic algorithms is no longer the accepted approach. While the outcomes shown at Rob|Arch were pretty wild and unpredictable, treating material behaviour as a generative component within a design system provides architects, for now, a role in a dynamic partnership where the human is needed to train and tune both machine and material. Architects with knowledge of computation can take on the responsibility to continuously improve robots freeing them up to engage in the aspects of architecture that initially attracted them to the field.

Robot Work Cells

As soft and hard robots become more financially accessible and more manoeuvrable, it is possible to imagine robot work cells maintained and continuously improved through interaction with architects, liberating them from the banal aspects of everyday practice, and re-connecting with the social needs of the built environment. Architects will continue to provide the subjective spatial intelligence that is critical for wonderful architecture, but collaboration with artificial agents and materials will enable a more efficient production of sophisticated, complex and unexpected outcomes. The skills needed to train and collaborate with these computational tools will need to be incorporated into architecture schools, not just exclusive to design and construction but also history and theory. The ability to train and collaborate with intelligence robots could provide the means to explore the past and imagine the future.

The Singularity

In this scenario quality and ability becomes measured through the architectural intelligence of robots which will reflect its human collaborators’ knowledge and capacity to teach. Beyond this, if the singularity is to be believed, the skilled up robot architect can take over practice while the human architect goes and spends time with loved ones, or be killed, it’s hard to tell at this stage.