Smart agriculture needs urgent data quality overhaul to harness AI’s potential

The future of agriculture is increasingly being shaped by artificial intelligence, but a new study published in Frontiers in Artificial Intelligence warns that poor data quality could undermine its potential. Their analysis shows that as Artificial Intelligence Generated Content (AIGC) becomes central to farming operations, the challenges of data noise, data fog, and data islands threaten the efficiency, fairness, and sustainability of smart agriculture.

The paper, “Data Quality Challenges of AIGC Application in Smart Agriculture,” evaluates the risks of relying on flawed or fragmented agricultural data and outlines a structured roadmap to address these systemic weaknesses as China moves forward with its National Smart Agriculture Action Plan (2024–2028).

How does poor data quality affect AIGC in agriculture?

According to the authors, AIGC is transforming agricultural decision-making by enabling automated recommendations in irrigation, fertilization, pest control, and market planning. However, these benefits are only as strong as the quality of the data being used.

The study identifies three interlinked threats:

• Data noise emerges in the quality design stage. It comes from sensor malfunctions, environmental fluctuations, transmission errors, and even human mistakes. If not corrected, noise produces biased instructions that lead to over-irrigation, pesticide waste, or inaccurate yield forecasts. Farmers may suffer economic losses while ecosystems absorb the cost of unnecessary chemical inputs.

• Data fog appears during the quality control phase. It stems from weak correlations between diverse datasets, inconsistent standards, and integration challenges. This fog slows down real-time decision-making and complicates analysis. In practice, it reduces the precision of agricultural operations, creating delays that can harm both crops and resource use efficiency.

• Data islands dominate the quality improvement phase. These silos occur when data cannot be shared across platforms or regions due to incompatibility, infrastructure gaps, or institutional barriers. The result is a fragmented landscape where knowledge remains locked within individual farms or organizations. This not only prevents effective large-scale planning but also widens inequalities between regions with advanced digital infrastructure and those left behind.

The authors emphasize that unless these threats are addressed, AIGC systems may provide misleading outputs, reinforce inefficiencies, and erode trust among farmers and stakeholders.

What solutions are needed to safeguard AIGC-driven farming?

To tackle these data challenges, the study applies the Juran quality improvement model and frames its recommendations around a quality loop approach, which emphasizes continuous improvement across three phases.

In the design phase, systematic noise detection and cleaning must be prioritized. Algorithms and sensor calibration protocols can help identify outliers and prevent corrupted data from feeding into AIGC models.

In the control phase, standardized data formats and unified protocols are essential. Establishing common standards would reduce inconsistencies between weather data, soil metrics, crop health indicators, and market statistics, allowing integrated analysis. Without such alignment, the scalability of AIGC tools remains limited.

In the improvement phase, stronger digital infrastructure and the development of regional data platforms can help eliminate data silos. Shared access will allow farmers, cooperatives, and policymakers to collaborate on resource allocation, crop planning, and market forecasting. The authors argue that addressing data islands is particularly urgent, as siloed systems exacerbate inequalities by giving digitally advanced areas disproportionate advantages.

By linking data quality management with agricultural governance, the roadmap aims to ensure that AIGC delivers on its promise of precision agriculture, efficiency, and sustainability.

What are the broader risks if challenges remain unaddressed?

The researchers caution that ignoring these data quality issues will not only limit the effectiveness of AIGC but also risk worsening agricultural imbalances.

On the environmental side, poor data quality could result in wasted water, excessive fertilizer use, and pesticide over-application, all of which degrade soil health and biodiversity. Economically, noise and fog reduce the competitiveness of farmers who rely on AIGC outputs, while data islands hinder large-scale optimization of production and distribution.

Socially, uneven access to clean, shareable data could widen the gap between technologically advanced farming regions and those with limited digital resources. Without equitable data sharing frameworks, smallholder farmers and rural communities risk being excluded from the benefits of smart agriculture, deepening regional inequalities.

The study also brings to light the governance dimension. For AIGC to support national and global food security goals, policymakers must establish clear rules for data quality assurance, standardization, and platform interoperability. If left unresolved, fragmented data governance will make it harder to respond effectively to food crises or climate-related disruptions.