The USDA will apply its own artificial intelligence tool to verify its Nov. 10 crop production report, following the September errors in the corn data. Nearly three-quarters of about 1,200 farmers surveyed distrust its reports.
The U.S. Department of Agriculture (USDA) will use an artificial intelligence tool it developed in-house to verify the data in its monthly crop production report, scheduled for publication on Nov. 10. The move was announced after a series of data errors eroded farmers' and market operators' trust in its forecasts, according to Reuters.
USDA Under Secretary Scott Hutchins said in an interview that, after the errors were detected in the September corn data, officials determined that the AI tool would have flagged the problem and decided to apply it to the crop production report. "I'm very bullish on the fact that we'll be able to utilize these tools in the future to help us avoid these kinds of really, truly unfortunate errors," Hutchins said on Thursday.
According to AgroLatam, the technology will work as an additional validation layer and will not replace existing statistical procedures. The September errors originated in a computer file that incorrectly combined irrigated and non-irrigated corn acreage data, according to Lance Honig of the USDA's National Agricultural Statistics Service (NASS); the original field survey information was accurate. Hutchins clarified that those mistakes are unrelated to federal staffing reductions, although restoring confidence will require consistent accuracy in the reports.
Confidence in USDA data is low among farmers following the major revisions to this year's reports, according to an Ag View Solutions survey cited by FarmWeek. Nearly three-quarters of the almost 1,200 farmers and farm managers surveyed said they have little to no confidence in USDA reports.
The online poll, conducted Oct. 25-28, drew responses from farmers in 38 states, including 175 in Illinois and 204 in Iowa. The results come after a year of significant revisions to USDA crop estimates that rattled commodity markets and fueled debate about the reliability of public data.
Only 38% of respondents said they had completed a USDA survey in the past 24 months, AgWeb reported. Among the 605 who skipped one and gave a main reason, 34% said they don't believe it matters, 29% said it takes too much time, 24% said the information is redundant and 13% said they don't want to share data.
Asked what would restore trust, 33% chose the USDA partnering with private companies to obtain better data and 30% chose a radical change in surveys and reports. In written responses, 23% suggested using data from the Farm Service Agency (FSA), the Risk Management Agency (RMA) and crop insurance; 12% mentioned satellite and artificial intelligence tools, and another 12% private and independent data.
Ag View Solutions consultant Shay Foulk told Brownfield Ag News that the recent listening sessions held by the USDA are a good step toward restoring trust in the agency, but that "the farmer is a little tired of optics and a little tired of platitudes and they're ready to see change." He added: "And then having some accountability of what is being done about it."
