From Coral Reefs to Robot Swarms: Rethinking Technology Through Nature

Hi Didem, on Tuesday 7 October, you’ll be joining UppTalk from 12:00 to 12:30. What will you be talking about?

- I’ll be talking about how we can make multi-robot systems more resilient and sustainable and how nature can actually teach us a lot about that. Right now, we’re at an unique crossroads: the design choices we make today will influence whether intelligent machines help or harm life on Earth in the future.

I want to show how ideas from natural ecosystems – from ant colonies to coral reefs – can inspire us to build robot swarms that are not only smart and efficient, but also ethically-aligned in how they behave towards each other and their surroundings.

What are multi-robot systems, and why is it important that they can work together without human control?

- Multi-robot systems are groups of robots that work together to carry out tasks that would be too difficult or inefficient for a single robot to handle. Examples include robots used for environmental monitoring, search and rescue, construction, or precision agriculture.

- It’s important that these robots can cooperate and make decisions on their own, especially in dynamic, unpredictable environments where things change quickly. In those cases, humans might not be able to respond fast enough or might not have all the information needed to make the right decisions.

- But full autonomy, where robots operate entirely on their own, comes with challenges. For instance, if we design them to compete instead of collaborate, they might end up working against each other. And if we removehuman oversight completely, we risk losing the ability to ensure ethical behaviour.

- That’s why it’s so important to get the coordination mechanisms right and nature offers great examples of how to solve these kinds of problems.

You’ve mentioned being inspired by nature’s ecosystems – how does that influence your work with robots?

- Nature has been running successful experiments in distributed coordination for billions of years. When I look at how mycorrhizal networks share resources between trees, how birds fly in flocks without a leader, or how bacteria make collective decisions, I see solutions to many of the challenges we face in robotics.

- These natural systems are incredibly resilient – they have built-in redundancy, adapt to change through feedback loops, and optimize for collective well-being rather than individual maximization. This has changed how I think about designing robot intelligence. Instead of building centralised control systems that can fail completely if something goes wrong, I focus on distributed systems that can learn and adapt.

- I develop reinforcement learning algorithms that encourage robots to form symbiotic relationships – helping each other out. And rather than just aiming for maximum performance, I prioritise sustainability and cooperation.

- So nature isn’t just a source of inspiration, it’s a blueprint for building the next generation of intelligent systems that can truly work in harmony with the world around them.