The journey from bee brains to AI – Dr Rachel Parkinson is transforming how we can study insect behaviour

Dr Rachel Parkinson’s work unites the seemingly disparate disciplines of artificial intelligence and bee senses in ways that simply haven’t been done before. An Eric and Wendy Schmidt AI in Science Fellow, Rachel is innovating the realm of insect behaviour by using self-built labs to study the effects of pesticide toxicity on bees.

Rachel didn’t always plan to be a scientist. After high school, she travelled and worked in a variety of jobs, eventually spending several years in Peru: “I became fascinated by how human activities shape biodiversity. I initially pursued a biology degree with the intent of going into environmental law but discovered a passion for experimental research and data analysis”, she explains.

Rachel grew up on the Canadian prairies, an agriculturally intensive part of the world. When she started her PhD, she was drawn to studying pesticides after seeing that research was emerging on their sublethal effects in non-target insects. Her early work focused on how low pesticide concentrations disrupt visual circuits in locusts (“yes, the plague locust”, Rachel confirms): “The results were striking, and I started wondering whether similar effects were present in bees, given their ecological importance”. That curiosity led her to neuroethology and a deep dive into bee sensory processing.

“I love that my research stretches across disciplines. One day, I’m studying neural circuits and sensory processing; the next, I’m developing AI tools to diagnose sublethal toxicity in insects. I’m especially excited about the potential applications of this work—from advancing our understanding of animal behaviour to improving environmental risk assessment.”.

Rachel’s work naturally has implications for agricultural management: “We need more rigorous environmental risk assessments because current methods often lack resolution.” The past few years have seen emergency use of neonicotinoids in the UK frequently making headlines, but recent months have seen a move towards a complete ban. Rachel considers it a complex issue: “While banning neonicotinoids may reduce some risks, we don’t fully understand their long-term impacts or how they interact with other stressors like climate change and habitat loss. Many pesticides, even at low concentrations, can have significant effects on insect behaviour. A broader, more precautionary approach to pesticide regulation is needed.”

Rachel’s current work has completely changed her research trajectory, having initiated a new research avenue that she hadn't contemplated previously. “I have started thinking more broadly about meta-science, and how we can integrate data, methods, and knowledge across disciplines. The biggest surprise has been my shift toward software development and using large language models to synthesise complex information. It’s opened up new ways of tackling big ecological questions.”

She continues: “Receiving the Schmidt Fellowship was a game-changer. It didn’t just provide resources for AI training; it connected me to an incredible network of scientists from diverse fields. Those collaborations have deeply shaped my research, expanding how I think about integrating AI with ecotoxicology and neuroethology.”

So how hard was it to incorporate AI into such a different area of research? “At first, it was daunting. It felt like staring up at a massive mountain without knowing where to start. I tried to learn everything at once, but the real progress came when I started chipping away at projects”.

It’s safe to say that Rachel has now climbed that mountain, having recently founded the MetaBeeAI consortium, which uses AI and large language models to synthesise research on pesticide impacts on bees. Her advice for someone starting out on this journey of applying AI to their science? “Find mentors who can help you prioritise what’s useful. It’s impossible to know what skills will be most valuable until you start using them. Get into collaborations where you can contribute to writing software and doing analyses alongside others – I've learned so much by example.”

Considering the wide implications for her research, Rachel reflects on what she thinks has been most important so far: “I’m working on some exciting projects now, with major findings coming soon. That said, I’m proud of my PhD work, which showed that low pesticide concentrations can disrupt vision and navigation in insects. More broadly, it demonstrated the power of borrowing methods from neuroethology to make discoveries relevant to ecotoxicology. Cross-disciplinary approaches have been a key theme in my research ever since.”

The rising star that she is, Rachel is soon moving on and up to the next step in her career: “I’ve accepted a tenure-track position at Queen Mary University, which I’ll be starting later this year. In the meantime, I’m excited to publish findings from my Schmidt Fellowship work and continue building collaborations that bridge AI, neuroethology, and environmental science.”

Illustration of Rachel's research by Alicia Hayden www.aliciahayden.co.uk