The Roboticist Who Wants Machines to Rot Like Leaves

Robotics has spent the past two years chasing a scaling problem: more units, more factories, more training data, more compute. Barbara Mazzolai has spent her career chasing the opposite question, what happens to a robot after it has done its job, and her answer is reshaping how a growing corner of the field designs machines in the first place. Mazzolai is associate director for robotics at the Istituto Italiano di Tecnologia, Italy's national research institute headquartered in Genoa, where she leads the Bioinspired Soft Robotics Lab. Her pitch to the field is direct: a robot built to solve an environmental or agricultural problem should not create a second one by leaving behind plastic, batteries and electronics that outlive their usefulness by decades.

Mazzolai calls the framework sustainability robotics, and it rests on three pillars she has spent recent work articulating alongside collaborators: minimal environmental impact across a robot's full life cycle, accessibility for users across different income levels rather than only well-funded labs and militaries, and symbiotic designs that create measurable benefit for both the humans who deploy a robot and the natural systems it operates in or near. The framework is not an abstract manifesto. It follows directly from two decades of bioinspired engineering in which Mazzolai has repeatedly gone to biological systems not for aesthetic inspiration but for functional answers to problems robotics has struggled to solve with conventional rigid-body design.

From Biology to Engineering, and Back Again

Mazzolai's path into robotics runs through biology rather than mechanical engineering, an unusual entry point that shapes how she frames a design problem. She trained first as a biologist before moving into engineering, and the switch shows up in her lab's output: robots that borrow structure and behavior directly from organisms that solve a version of the same physical problem a robot is being built to address, rather than robots that use biology as a surface metaphor for an otherwise conventional actuator-and-sensor design. An octopus can move through cluttered, unpredictable environments and manipulate irregular objects using a boneless, distributed nervous system rather than a small number of rigid, centrally controlled joints, a capability that has informed soft-robotic manipulator designs across the field. A plant root grows by adding material at its tip and can sense and respond to obstacles, moisture gradients and mechanical resistance as it extends through soil, a mechanism Mazzolai's lab has translated into root-inspired robots that grow through soil from the tip rather than being pushed through it, reducing the energy and disturbance involved in subsurface exploration. A seed like that of the Erodium plant uses a self-coiling, humidity-responsive structure to drill itself into soil without any onboard power source at all, a passive mechanism Mazzolai has drawn on for low-energy, self-burying robotic designs.

Those three reference points, octopus, root and seed, are not separate research threads so much as three answers to the same underlying question: how does a biological system achieve capability, adaptability or efficiency that current robot design has not matched, and what happens to engineering practice once that mechanism is actually understood rather than loosely imitated. That question is what gives sustainability robotics its teeth. A soft, biodegradable actuator that performs a task and then breaks down is not a compromise on capability for the sake of environmental virtue in Mazzolai's framing. In several of her lab's designs, the same biological mechanism that inspired a robot's function, a root's growth-from-the-tip movement, a seed's passive drilling action, also points toward materials and structures that can be built to decompose, precisely because the biological original does.

What Soil Symbiosis Teaches a Robot About Where It Is

One of Mazzolai's more unusual current research lines studies the symbiotic relationship between rice roots and mycorrhizal fungi, the underground fungal networks that trade nutrients with plant roots in exchange for sugars, effectively extending a plant's sensing and resource-gathering range far beyond what its own root system could reach alone. Mazzolai's group is working to extract behavioral rules from that relationship, how the fungal network responds to soil conditions, how it routes around obstacles and toward resources, with the goal of encoding comparable rules into robots built to explore soil autonomously. The application she has pointed to is not limited to agriculture: a robot that can sense its way through soil using rules borrowed from a plant-fungus symbiosis is a candidate for subsurface exploration on Earth and, longer term, for soil analysis on other planets, where a machine cannot be resupplied, repaired or retrieved if it becomes stuck or damaged.

That planetary framing sharpens the stakes of the sustainability argument in a way that is easy to miss when the same argument gets made about a warehouse robot or a lawn-care device. A robot sent to explore soil on Mars cannot be recovered at the end of its mission, which means every material choice made at design time is permanent in a way that a factory robot's material choices are not. Mazzolai's insistence that engineers think through a robot's entire life cycle, not just its operating life but what happens to its body once that operating life ends, reads less like an environmental preference and more like an engineering requirement once the deployment environment is one where cleanup is not an option. The same logic applies, at smaller scale, to a soft robot deployed in a river, a field or a forest: recovery is often impractical even when it is theoretically possible, and a robot's material choices determine whether an unretrieved unit degrades harmlessly or becomes another piece of persistent waste in the environment it was built to study or protect.

The Argument Against Treating Sustainability as an Add-On

The most direct version of Mazzolai's argument is that engineers have to reduce the footprint of their technology by designing for reuse, recycling or biodegradation from the outset, not as a retrofit applied once a design already works. That ordering matters because the alternative, treating end-of-life material choices as a problem to solve after a robot's mechanical and software design is finalized, tends to produce exactly the kind of compromise sustainability robotics is meant to avoid: a robot that works well and then gets a biodegradable coating or a recycling plan bolted onto an otherwise conventional design, without the coating or the plan meaningfully changing how much of the robot ends up as landfill waste. Building life-cycle thinking into the initial specification, alongside cost, weight and performance targets, is a harder constraint to satisfy than adding it later, but Mazzolai's own robots are evidence that the harder path produces designs where the biological inspiration, the sustainability outcome and the functional capability reinforce each other instead of trading off against one another.

Accessibility, the second pillar of the framework, follows a similar logic. A soft, passively actuated seed-inspired drilling robot that requires no battery and little specialized manufacturing infrastructure is inherently cheaper and easier to deploy at scale than a robot built around precision electromechanical actuators and onboard power management, which means the same design choices that reduce a robot's environmental footprint also tend to lower the cost of building and operating it. That alignment is not guaranteed in every case, and Mazzolai's framework does not claim it is, but it is common enough across her lab's output that sustainability and accessibility function less as competing priorities and more as correlated outcomes of the same bioinspired design discipline.

A Discipline Still Finding Its Footing in a Fast-Moving Field

Sustainability robotics is not yet a standard line item in how most commercial robotics programs plan a product, and Mazzolai's framework arrives at a moment when the industry's dominant conversation is about scaling humanoid production, cutting per-unit cost and winning deployment contracts rather than designing for decomposition. That gap is itself informative. The field's current center of gravity rewards a robot that ships fast and works reliably far more than one that biodegrades cleanly, and a procurement team evaluating a robotics vendor today is unlikely to see end-of-life material strategy on a spec sheet next to payload capacity and battery life. Mazzolai's bet is that this changes as robots move from controlled factory floors into open environments, farms, waterways, disaster sites, planetary surfaces, where recovery is harder and the cost of leaving behind non-degrading hardware becomes a liability rather than an externality someone else absorbs.

Whether sustainability robotics becomes a mainstream design constraint or remains a research specialty concentrated in labs like Mazzolai's will depend less on the strength of the biological inspiration behind it and more on whether commercial deployment starts moving into the kinds of unrecoverable environments where a biodegradable root-inspired probe has an obvious advantage over a machine built to be discarded and forgotten.

This account synthesizes public reporting and published research on Barbara Mazzolai's work at the Istituto Italiano di Tecnologia. It is for general information purposes only and does not constitute investment, financial, or legal advice.

Hero image credit: IEEE Spectrum.