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

The US Department of Agriculture (USDA) plans to utilize a new artificial intelligence (AI) tool to scrutinize data in its upcoming monthly crop production report, following a series of data discrepancies that have damaged farmers' and traders' trust in its forecasts. USDA Under Secretary Scott Hutchins revealed the AI tool's debut during an interview.

The agency previously launched a pilot project that relied more on technology to enhance crop acreage and yield estimations due to a decreasing number of farmers responding to its surveys. The tool's debut tests whether the USDA can restore trust in data that frequently impacts market movements. The USDA corrected its September corn production and yield data, which raised the yield estimate by around 3% in Nebraska, the third-largest corn-producing state, and decreased it by approximately 2% in Kansas.

The agency was previously criticized for making significant upward revisions to estimates for last year's corn acreage and issuing inaccurate beef export sales data. The September errors were caused by a computer file incorrectly merging data on irrigated and non-irrigated corn acres gathered through field surveys, disproportionately affecting Kansas and Nebraska, which have higher percentages of irrigated fields.

The USDA stated that the error was isolated and that no issues arose during data collection at the field level. However, the National Corn Growers Association criticized the September errors, stating they further undermined farmers' confidence in the USDA's ability to produce accurate reports. The new AI tool will complement, not replace, the USDA's data validation process.