Agricultural AI: Issues & Values

by Jeffrey Guan, in the Brain Food Society

Introduction

In his article on Agricultural AI, Mark Ryan examines the social and ethical impacts of AI in agriculture. What? That’s right, AI software is used to monitor fields, soils, and crops, while AI robots work autonomously on farms as they replace farmworkers to harvest fruits and vegetables or to spray weeds and pesticides. So what is the current landscape of social and ethical impacts of Agricultural AI? To answer this question, the author analyzes eleven overarching principles of AI ethic guidelines, including transparency, justice and fairness, non-maleficence, responsibility, privacy, beneficence, freedom and autonomy, trust, dignity, sustainability, and solidarity. In other words: is AI safe? Is AI fair? Who’s liable when it comes to AI? How does AI use energy? Is AI sustainable?

When it comes to AI and its impact on the world, since we are all concerned about the food we consume, we should also be concerned about the impact of Agricultural AI. The AI being used in agriculture is narrow AI, as it is an artificial intelligence programmed to do specific tasks. When AI is designed and trained for certain tasks it’s very unlikely for AI to make mistakes or biased decisions, which would probably never happen in the agricultural sector. 

Method

The research to see whether or not AI is ethical or how AI impacts the agricultural sector used an outline of eleven ethical principles. To conduct his research, Ryan searched for articles on Agricultural AI through Scopus and Google Scholar searches. 269 articles were found and only a small subset was used because of the exclusion criteria. The research would be used to see how they connect to the eleven ethical principles and how it will impact the agricultural sector.

Results

The research result for each of the eleven ethical principles(in bold):

1. A concern for AI when it comes to non-maleficence is that AI can be hacked and it may unintentionally hurt something on the farm. 

2. A concern with being able to trust AI is that we don’t know how we can trust it with data from farms and we don’t know when AI might make a mistake.

3. AI benefits everyone as agriculture is the main way humans get food and faster, more efficient, and effective farming will only benefit people.

4. There are many legal rules that must be followed and this can make it harder for farmers when using their AI. (Legality)

5.  With AI having access to data we don’t know if AI may leak or give out private information only for farmers. (Privacy)

6. It may not be fair to smaller farming companies that may not be able to get access to AI. (fairness + justice)

7. Farming/AI companies may not responsibly use AI, an example may be that the company that creates the AI makes farmers sign an agreement that if anything goes wrong with the AI it’s on the farmers. (responsibility)

8. People worry that AI might be wrong and might be unreliable sometimes, or the AI may be interpreted incorrectly by farmers when reading data, etc. (reliability)

9. AI will make farming more sustainable.

10. With AI being so new, solidarity can’t really be answered. 

11. Dignity isn’t a huge concern in the agricultural sector.

Conclusion

The point of the research paper was to examine the societal and ethical impacts of using AI in agriculture while also identifying gaps in both agricultural literature and AI ethics guidelines. The main way to research this topic was by using eleven overarching ethical principles and researching how AI connects to the principles. The least discussed principles in agricultural AI literature were transparency, dignity, and solidarity. The most discussed principle was sustainability. To conclude, the paper identified the qualities of research being conducted on the societal and ethical impacts of AI in agriculture.

Reference

The social and ethical impacts of artificial intelligence in agriculture: mapping the agricultural AI literature - Mark Ryan (Published: January 3rd 2022)