Robust AI, a company co-founded by MIT Professor Emeritus Rodney Brooks and Anthony Jules '92, SM '92, builds collaborative robots intended to slot into existing workplaces rather than replace them. Jules argues that results in these settings depend on the whole system: robots handle part of the work, people handle part, and backend systems handle the rest, while workers keep an appropriate level of agency and input so the operation keeps moving forward.
Brooks served on the MIT faculty from 1984 to 2010 and directed the Computer Science and Artificial Intelligence Laboratory (CSAIL) from 1997 to 2007. During his time at MIT he co-founded iRobot, maker of the Roomba vacuum, which went public in 2005, and later co-founded Rethink Robotics, another collaborative-robot venture. Jules met Brooks in the 1990s as a CSAIL graduate student; their labs sat on the same floor, though they never worked together. After graduating, Jules co-founded Sapient, which went public before being acquired by Publicis in 2015.
The company grew out of a 2019 meeting at a robotics conference, where Brooks told Jules he was considering a robotics software startup he compared to a video game engine for training robots. Robust AI was founded later that year. The Covid-19 pandemic first pushed the team toward robots that could autonomously disinfect surfaces; once it became clear the virus was not transmitted via surfaces, they had to pivot again. Jules recalls warning Brooks that the new idea would both please and frustrate him: it solved problems customers had been describing, but it required building hardware.
After years of interviewing potential customers through the company's early pivots, the founders settled on logistics and manufacturing as initial markets. Jules, now Robust AI's chief innovation officer, notes that the popular picture of factory automation is the roughly 5 percent of facilities that are fully automated with $50 million to $200 million systems, while the vast majority of factories have no automation at all, not even a conveyor belt, and will not tear out their operations for a $200 million installation.
The company's Carter robots can be phased into operations as needed and support several workflows: guiding human workers to the spots where items are retrieved, following workers to loading points, or moving goods from one place to another. That last workflow is the primary one, aimed at large warehouses and factories where the company says employees can walk 30,000 steps a day on average. Brooks, the chief technology officer, describes a worker loading a robot, saying "done," and letting the machine cover a distance that might be 400 feet; each customer decides how the robots are integrated into the workspace.
The design philosophy also draws on failures elsewhere. Leila Takayama, Robust AI's vice president of human robot interaction, came across a peer-reviewed study of hospital robots used to gather dishes, move sheets, and transport equipment; the researchers found that in many hospitals large numbers of the machines were powered off and pushed against walls, with nurses reporting they had broken. According to Brooks, what actually happened was that the robots stopped dead in hallways when nurses rushed past with a patient on a gurney, blocking the gurney's path — the machines were getting in the way of the staff's primary tasks. Robust AI's development approach is therefore to understand workers' primary tasks deeply and build features that accelerate them, whether a person or a robot performs them. Humans, Brooks says, have had millions of years of evolution to get good at the physical world, and the digital system should respond to that and supply what it does best.