Towards Responsible AI in Agriculture: A Framework for Machine Unlearning in Precision Farming Systems

Applying Artificial Intelligence (AI) is enhancing farming, with results such as 20% more crops and 30% better use of resources. Depending on AI more heavily brings issues related to personal information, unfair biases and ensuring models remain responsible. Thanks to the proposed framework, sensitive or unnecessary data in AI models for precision farming can be properly dealt with to make sure AI is handled carefully. Since data analysis is applied in agriculture within the framework, it attends to the GDPR regulations and helps catch bias in approximately 40% of agricultural models. Over 95% of results have been found correct and it takes the data elimination steps within the suitable time. In this way, it makes sure the first steps in AI farming are ethical, clear and protect privacy which helps to build trust and support sustainability in smart agriculture.