The Innovator Bringing Artificial Intelligence into the Real World

Embodied cognition, the notion that our physical bodies influence our cognitive processes, is a concept with profound implications for the advancement of truly intelligent machines. According to Josh Bongard, a prominent roboticist at the University of Vermont, the realization of artificial intelligence’s full potential hinges upon its capacity to directly engage with and experience the physical world.

Bongard stands as a leading authority in the realm of embodied AI, which seeks to integrate artificial intelligence and robotics. He contends that simply amalgamating an AI chatbot with a robotic arm, as seen in Google’s PaLM-E system, falls short of the mark. Instead, he advocates for the application of “evolutionary robotics,” a methodology that leverages the principles of natural selection to rapidly iterate through various robot designs, many of which are comprised of pliable materials. Moreover, Bongard is actively engaged in a project involving the use of living cells to fabricate rudimentary biological robots known as xenobots. These xenobots not only execute rudimentary tasks, but they also demonstrate an ability to interact with and respond to their surroundings.

In an interview with New Scientist, Bongard delves into how these efforts are reshaping perspectives on embodied cognition and the definition of a robot. He believes that these insights stand to transform the approach to developing intelligent machines.