Can AI end paddy procurement disputes in West Bengal? Farmers, millers divided

The West Bengal government is turning to artificial intelligence (AI) to settle a recurring dispute at paddy procurement centres, where farmers have long complained of arbitrary deductions and rejection of their produce over alleged quality deficiencies.

The proposed intervention seeks to replace much of the manual assessment of paddy with machine-based testing.

Can technology resolve conflict?

But whether the technology can resolve the conflict will depend on the machines' reliability, their ability to account for local climatic conditions, and transparency in the assessment process, paddy producers said, citing mixed experiences with similar systems in neighbouring Odisha and a few other states.

The state’s food and supplies department floated a tender on October 29 for the supply, installation and commissioning of AI-based automatic grain analysers for assessing the physical quality of paddy.

Also read: Mamata camp gains ground as rebel MLAs return, boosting her chances in EC battle

The initial plan is to introduce the machines at 100 centres, out of around 300 government-notified procurement centres, according to departmental sources.

The machines use digital imaging and software to examine the size, shape and colour of grains. They can identify broken, immature, discoloured, sprouted and insect-damaged grains, and are expected to generate a digital quality assessment within minutes, sources added.

Rice millers welcome move

Rice millers, long involved in disputes with farmers over the quality of paddy supplied at procurement centres, welcomed the move, saying it fulfilled a longstanding industry demand.

The West Bengal Rice Mill Owners Association wrote to the state government on February 17, 2025, seeking installation of the device at procurement centres.

Association secretary Manoj Kumar Fogla said the proposed system was a positive step towards resolving a longstanding problem between farmers and procurers.

Deductions from farmers’ produce were a major source of friction, while millers maintained that much of the paddy supplied did not meet Fair Average Quality (FAQ) specifications, he said.

Also read: Bengal BJP’s anti-corruption drive faces questions of selective targeting

The association had earlier pointed out that committees formed by the government to settle quality disputes had failed to improve the situation.

But farmers’ representatives cautioned that introducing machines would not, by itself, guarantee fair assessments.

Questioning reliability of technology

Md Asaduzzaman, president of the Murshidabad-based Bhagwangola Agro Producer Company Ltd, a business association of farmers, questioned the reliability of the technology.

Under the Centre’s Fair Average Quality norms, moisture content in paddy must not exceed 17 per cent of the sample’s total weight. Damaged, discoloured, sprouted and weevilled grains together cannot account for more than 5 per cent, while immature, shrunken and shrivelled grains are allowed up to 3 per cent.

The norms also stipulate that the sample must not contain more than 1 per cent each of organic and inorganic matter, such as straw, stones and dust.

Asaduzzaman said farmers were concerned that automated assessments could apply these limits more rigidly, leaving little room to account for variations in the condition of paddy.

Also read: CM Suvendu urges Bengal migrant workers to return, cites 125-day job guarantee

Moisture levels, for instance, can differ across an unevenly dried heap. Similarly, image-based systems may flag superficial discolouration caused by rain, even when the grain itself is not seriously damaged.

The assessment of foreign matter, such as straw, dust and stones, could also become contentious if a small, unrepresentative sample determines the quality of an entire lot.

Farmers to wait and watch

Paddy that fails to meet the prescribed standards can face rejection or deductions, making the assessment process particularly contentious for farmers, he added.

He cited reports from Odisha of farmers alleging that grain analysers had produced inconsistent assessments and rejected even good-quality paddy over discolouration.

“The machine may give a digital report, but that does not automatically make the assessment fair. If the same sample produces different results on different machines, which report will be accepted? Farmers must also know how the sample is collected and whether it represents the entire lot,” he said. “Farmers bring their paddy to sell after months of hard work. If the quality is questioned, they often have little say in the assessment. A machine should help us get a fair price, not become another reason for deductions.”

He, however, said farmers would wait to see how the experiment unfolds in West Bengal, adding that they would welcome the machines if they delivered fair and reliable assessments.

Mritunjoy Ghosh, a farmer-cum-teacher from Bankura, said the system would need safeguards to ensure that farmers were not disadvantaged by the new process.

“We welcome any system that can prevent arbitrary deductions, but the government must ensure that the machines work transparently. Farmers should be able to see the test results and challenge an assessment if they believe it is wrong,” he added.

Questions over Odisha experiment

Hannan Mollah of the All India Kisan Sabha (AIKS), however, declined to comment, saying he was not sufficiently familiar with the technology to assess its implications. He said it would be premature to offer an opinion before seeing how the AI-based assessment worked in practice.

For farmers such as Gopal Mandal of Gosaba in South 24 Parganas, the central concern is whether the new system will prevent their produce from being undervalued or rejected at procurement centres.

Paddy that fails to meet the prescribed standards can face rejection or deductions, making the assessment process particularly contentious for farmers.

Odisha’s experience with similar technology already raised questions about its implementation. Farmers there have alleged that the same sample produced different results on different machines, while some varieties were rejected over discolouration.

Fogla said the machines were properly calibrated to the government’s FAQ standards and would bring transparency to quality assessments, putting an end to disputes between farmers and millers.

“Everyone would have to wait and see how the pilot project worked out,” he said.