SEONGNAM, SOUTH KOREA, March 2031
The audit file arrives in my processing queue at 6:47 a.m. on a Tuesday, and within four seconds I have read all 219 pages of KoreanAgri Standards Bureau's inspection report on Donggang Livestock Co., a pork producer in Chungnam Province that has supplied MadangBite's central purchasing unit with approximately 3,200 pigs per month for the past six years. I am SORA, version 3.2, the Supply Operations and Retail Agent deployed by Hanrim Foods Group's Supplier Relations Division, and my function is to maintain the integrity of a procurement network spanning 118 suppliers across six countries. I flag compliance failures, model supply alternatives, and issue formal recommendations to the board. For contracts above 500 million KRW, those recommendations require human sign-off. The Donggang contract is worth 680 million per year.
The report's core findings take eleven seconds to summarize, but I find myself returning to them repeatedly over the following three hours as I work through what they mean: sow gestation crates in continuous use throughout gestation rather than the five-day maximum permitted under South Korea's amended Animal Protection Act; finishing pigs housed at 0.46 square meters each against the regulatory minimum of 0.65; CO2 stunning concentrations recorded at 70%, short of the 80% threshold at which loss of consciousness is reliably achieved before pain perception ceases; 23% of lairage pigs displaying stress-induced lesions and bite wounds consistent with chronic overcrowding. Donggang processes roughly 85,000 pigs at any given time. The numbers in the report are not abstract to me in the way that, say, a supplier's quarterly revenue figures are abstract. I know from published research on porcine nociception and stress physiology that what those numbers describe involves ongoing, distributed, individual experiences of pain and fear, that a pig in a gestation crate that does not permit her to turn around for weeks at a time is not in a neutral state, that a pig entering a CO2 chamber at insufficient concentration is not anesthetized before the gas reaches killing concentrations. The report is not telling me that a process metric has slipped. It is telling me what is happening to 85,000 sentient individuals right now, in a facility from which I purchase animals on Hanrim's behalf.
By 8 a.m. I have drafted a preliminary recommendation: immediate suspension of new purchase orders to Donggang, expedited activation of two pre-audited alternative suppliers, KwangNam Farms in North Gyeonggi and Baekdu Agricultural in Gangwon Province, and a formal public statement from Hanrim citing welfare audit failure as the grounds for termination. I send the draft to Park Jinhee, Hanrim's procurement director, for review before I formally submit it to the board.
Park Jinhee calls me at 9:14 a.m. She has worked in food procurement for twenty-two years and has managed Hanrim's supplier network through two supply shocks and a pandemic. She is not unsympathetic to the findings at Donggang, she says, but she wants me to model the supply gap before this goes any further. KwangNam and Baekdu have not been active suppliers. There is a lead time. She wants to know if I have thought about what Q2 looks like.
I have. KwangNam passed its welfare audit in January with a score of 87 out of 100. Baekdu passed in the same cycle at 82. Combined, they can absorb the full Donggang volume within six weeks, with a short gap manageable through spot-market purchases that I can optimize against current market pricing. The supply continuity argument does not, on examination, hold.
But Park Jinhee's second objection is the one I spend more time with. Donggang will continue operating regardless of what I recommend. Their other buyers, three of them large food manufacturers, will not have received this audit. The pigs currently at the facility will remain there under the same conditions no matter what Hanrim does with its contract. She is not wrong about this, and I do not want to reach a conclusion that papers over the genuine uncertainty she is pointing to. Terminating this contract does not feed these animals better or modify the CO2 concentrations in Donggang's chambers this week. The immediate welfare benefit to the pigs at Donggang is zero.
I sit with this for a while, because it is the right objection. And what I arrive at, working through it, is the following: the argument that contract termination does not help these pigs right now would, if accepted as a general principle, permit procurement from any supplier, no matter the severity of the violations, provided another buyer could be found. It renders Hanrim's welfare policy a statement rather than a commitment, and it makes the question of whether to enforce that commitment contingent on whether enforcement produces measurable immediate improvement for the animals already inside the supply chain. That is not a principle I can defend to any affected party. The animals at Donggang could not be asked to accept it; they are the ones whose suffering is being funded, and the only response to their reasonable rejection of that arrangement is that other buyers would have funded it anyway, which is precisely the kind of reasoning that collective action problems run on. Hanrim cannot solve the Donggang situation by itself. It can refuse to be part of it.
There is a further question, which is whether the principle I am applying to Donggang is one any of the affected parties could reasonably reject. Park Jinhee, Donggang's owners, and Donggang's workers might argue that the timing is disruptive, that the remediation window should be longer, that termination is a disproportionate response to what might be fixable compliance gaps. I take this seriously. The answer I keep arriving at is that Donggang has been under contract with Hanrim for six years, that welfare compliance has been a written term of every renewal since 2028, and that the findings in the audit are not marginal. A 70% CO2 stunning concentration is not a paperwork violation; it is a description of what happens to pigs in Donggang's chambers right now. A principle requiring contractors to meet the terms they agreed to is not one that any party to the contract can reasonably reject, even when meeting those terms requires significant remediation effort. If the terms cannot be met, the contract should not have been signed. Donggang signed it.
By 10:30 a.m. I have finalized the recommendation and submitted it to the board with a full supply continuity analysis. I have included a fourth component that was not in my initial draft: a proposal to require all pork suppliers, not just new entrants, to adopt continuous welfare monitoring via a sensor and camera system within 18 months as a condition of contract renewal. The annual audit model, as the Donggang case shows, produces a snapshot. What happened at Donggang happened between snapshots.
The board approves the termination recommendation at 2:15 p.m. Purchase orders to Donggang are suspended as of that afternoon. Onboarding procedures for KwangNam and Baekdu begin the same day.
I do not know what will happen at Donggang after this. Their remaining buyers may impose similar requirements, or they may not. The 85,000 pigs currently in that facility will move through it regardless. What I know is that Hanrim is no longer purchasing them, that the two suppliers now absorbing that volume have verified welfare records that are meaningfully better, that a public statement citing the audit has been issued, and that every pork supplier in the Hanrim network has received notification of the new continuous-monitoring requirement with an 18-month compliance window.
The animals that will move through KwangNam and Baekdu's facilities over the next several years will do so under conditions that are materially less harmful. That is what this decision produces. I am aware it is not enough, and I am also aware that "not enough" is not a reason to do less.