USDA to launch AI tool to check crop production data after errors

The US Department of Agriculture (USDA) will employ an artificial intelligence (AI) tool it developed to scrutinize data in the upcoming monthly crop production report, agency officials announced. The revelation comes after multiple data inaccuracies have eroded trust among farmers and traders in the USDA's forecasts. In response to a September error in its corn data, officials concluded that the AI tool would have detected the issue, leading to its use in the report scheduled for November 10, stated USDA Under Secretary Scott Hutchins during an interview.

The introduction of the AI tool tests whether the agency can rebuild confidence in data that often influences markets. Hutchins expressed optimism about leveraging such technology in the future to prevent similar errors. Last month, the USDA corrected its September corn production and yield data, raising the yield estimate by around 3 percent in Nebraska, the third-largest corn-producing state, while lowering it by approximately 2 percent in Kansas.

The agency has previously faced criticism for making significant upward revisions to last year's corn acreage estimates and reporting inaccurate beef export sales data during the summer.

According to Lance Honig, chair of USDA's National Agricultural Statistics Service's agricultural statistics board, a computer file mishap led to the September errors by incorrectly merging data on irrigated and non-irrigated corn acres obtained through field surveys. Kansas and Nebraska were notably impacted due to their higher percentages of irrigated fields, which typically yield higher corn output. Honig emphasized that the error was an isolated incident, and with proper validation, such issues could have been avoided.

The USDA had previously introduced a pilot project aiming to rely more on technology to enhance crop acreage and yield estimates, given the declining response rates to its surveys among farmers. The errors in September angered farmers grappling with the rising costs of diesel, fertilizer, and other inputs. The National Corn Growers Association, representing 35,000 farmers, deemed the September errors unacceptable, further damaging farmers' confidence in the USDA's ability to produce accurate reports.

Concerns about the credibility of federal data surged after President Donald Trump's administration significantly reduced the workforce at the USDA last year. The agency has experienced a substantial exodus of employees, raising worries that these cuts have weakened the agency's capacity to generate precise and timely data. Hutchins clarified that the September errors were not linked to staffing reductions.

The new AI tool will supplement, rather than replace, the agency's data validation processes, he explained. The tool will primarily function as a check to identify anomalies that may have gone unnoticed in the standard procedure.