AI System Reveals How Tiny Doses of Pesticides Disrupt Entire Honeybee Colonies

A groundbreaking AI and modeling approach reveals that even sublethal neonicotinoid exposure hinders honeybee foraging and jeopardizes colony survival, offering a new standard for pesticide risk assessments.

Research: Reduced Honeybee Pollen Foraging under Neonicotinoid Exposure: Exploring Reproducible Individual and Colony Level Effects in the Field Using AI and Simulation. Image Credit: Baronb / Shutterstock

Honeybees are essential pollinators for agriculture and natural ecosystems. Stressors such as climate change, habitat loss, and pesticide exposure threaten their ability to forage for pollen, a critical resource for colony survival. Researchers in the ACS' Environmental Science & Technology journal demonstrate that an artificial intelligence (AI)-based monitoring system, combined with a computer model, can link the exposure of neonicotinoid pesticides to individual honeybees with the health of the entire colony.

Neonicotinoid pesticides are widely used in agriculture. Plants absorb and distribute neonicotinoids throughout their tissues, making the pollen from these plants potentially harmful to honeybees. According to past field studies, bees exposed to neonicotinoids make fewer trips to collect pollen throughout the day. However, the connection between changes in individual honeybee pollen-foraging behavior and overall colony health hasn't been widely researched. Now, a multidisciplinary team led by Ming Wang hopes to change this by combining field study data with computer modeling and AI-based monitoring.

The researchers repeated their 2019 pollen-foraging field experiments, in which they exposed honeybees to sublethal doses of neonicotinoids and then tracked the insects' activities using AI-based camera technology and traditional ecotoxicology methods. In this experiment, the researchers analyzed recently collected data using a computer simulation called BEEHAVE, which was designed to explore the effects of stress on honeybee colony dynamics.

Using their new approach, the team members found that even low exposure to neonicotinoid pesticides led to less efficient pollen foraging at both the individual and colony levels, confirming their previous studies. "We were surprised that we could replicate the findings of our first field experiment in 2019," says Silvio Knaebe, one of the researchers. "Honeybee colony behavior varies so much that statistically significant effects are difficult to detect."

Given these initial results, the researchers say their new model, with pollen-foraging behavior as a key parameter, could be a unique candidate for pesticide risk assessment in the field at both the individual honeybee and colony levels.

The authors acknowledge funding from the Federal Ministry of Food and Agriculture (Bundesministerium für Ernährung und Landwirtschaft) via the Federal Office for Agriculture and Food (Bundesanstalt für Landwirtschaft und Ernährung), Germany.