WUR project to explore opportunities and dangers of using AI in agri-food

09 Feb 2022 --- Researchers from Wageningen University and Research (WUR) are to investigate what artificial intelligence can bring to the agri-food space, following US$2.5 million in funding from the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO).

They will design a methodology for the use of AI in the agri-food sector which will include tools, guidelines; ethical and legal; as well as engagement strategies for the application of AI. 

“Right now people have no clue what AI can bring to agri, nor what dangers it may imply. If a farmer contributes to an AI project by sharing data on their daily milk output, they want the data to be used anonymously so as not to inform competing farmers on their company’s output,” says Simone van der Burg, senior researcher ethics and responsible research and innovation at Wageningen Economic Research and co-lead of the project.

“But how do you protect their data?”

The research will be conducted in an AI for Sustainable Food Systems (AI4SFS) laboratory following ethical, legal and societal (ELSA) impacts protocols. 

The researchers will conduct six case studies where ELSA factors will be integrated into the AI design.

Robots and humans togetherWUR researchers hope to realize an AI hub with NWO grant funds. (Credit: WUR)
The objective of the research is to expand the number of case studies and projects dedicated to the ELSA lab and integrate the network of intelligence expertise of Wageningen Research Centers, closer to the lab. 

“AI robots will take over jobs of humans and are expected to profoundly change the labor market in the agri-food sectors,” explains Van der Burg.

“The questions which the researchers will tackle are “what are the threats for human laborers” or “what opportunities do robots offer to create better and more meaningful work”. 

“A weeding robot can accidentally take out the crops instead of the weeds, thus destroying the harvest. Who can be held responsible for that? And who must pay for the damage? An algorithm is trained to seek the best investments for a large agri company, which leads to enhanced profits,” Van der Burg continues. 

“However, the algorithm therewith eventually commits fraud. The definition of ‘fraud’ presupposes a human criminal. Can we blame the algorithm?,” she asks.

AI hub of data protection
When the project reaches completion, the researchers hope an ELSA hub for responsible agri-food AI applications would be realized for advanced research and practical expertise on how to develop and implement ELSA and AI protocols. 

“This allows us to show the relevance of the Social Sciences Group (SSG) work for digitalization. This is sometimes ignored, yet often it is lack of trust in AI that hinders its success. We think the provision of grants from this type of work by the AI Coalition shows recognition of this,” says Vincent Blok, project lead.

AI is being used more frequently in the food sector. For example, fragrance and flavor expert Firmenich’s Formulae Generator provides customized perfumery and taste solutions to be delivered to market quickly.

Symrise is harnessing AI to discover flavor modulating ingredients in plant varieties. With this technology, the German supplier has looked into decoding 40,000 ancient Chinese recipes to identify new potential materials with taste balancing capabilities.

The use of AI can help predict how consumers will respond to particular sensorial properties of NPD.

By Inga de Jong