AI satellite checks trigger 79,700 farm subsidy queries

Satellite monitoring of agricultural subsidies generated 79,700 requests for clarification from Finnish farmers last year, Yle reports. The system compares satellite images with subsidy applications, but still makes many mistakes.

In Satakunta, 2,500 farms received a total of 5,500 requests. The Finnish Food Authority sends a request when a satellite image conflicts with information in a farmer’s subsidy application.

Satakunta farmer Anne-Mari “Amski” Tarkkio, who is also known on social media, is growing wheat, barley, rye, peas and hay this year. She is also applying for subsidies for soil-improvement fields and wildlife fields.

Tarkkio has received four requests for clarification this summer, compared with eight last year.

Technology struggles with mixed crops

The satellite system is particularly prone to errors when several plant species are growing in the same field. Wildlife fields must contain a varied selection of food suitable for forest animals, but the Vipu application used to monitor subsidy conditions told Tarkkio that the satellite image could not establish whether her field met the requirements.

After receiving a request, the farmer must photograph the field using the Vipu application and use the image to show that the satellite data is incorrect. Tarkkio was able to demonstrate that her wildlife field met the requirements.

Her soil-improvement field caused more concern. Such fields are intended to improve the soil and make it more nutrient-rich in the future. Tarkkio planted oilseed radish, chicory and mustard there.

The middle of the field was too wet in early summer, leaving the vegetation sparse. The satellite image detected the deviation, and in this case the request for clarification was justified. Tarkkio expects that she will not receive the subsidy for that field.

System has improved since launch

Satellite monitoring was introduced in 2023 and initially faced strong criticism from farmers because of the flood of clarification requests.

Hanna Mäntylä, a group leader at the Southwest Finland Centre for Economic Development, says artificial intelligence and algorithms still make mistakes, although the system has improved over the past three years. Natural phenomena, such as cloudy weather, can also lead to requests being sent to farmers.

Farmers’ photographs correct mistakes made by the artificial intelligence, and almost all requests are approved after the images are submitted.

Despite receiving several requests, Tarkkio considers satellite monitoring a good and fair system.

“I would rather take photographs of my fields than have inspectors come here in person,” Tarkkio said, comparing the system to choosing between brushing your teeth every morning and evening or going to the dentist.

She also believes satellite monitoring treats all agricultural subsidy applicants equally. In her view, subsidies should go to people who genuinely work their land and produce food rather than to someone using a wildlife field as a pension fund.