Sandhya Das Thuraisingham
In the early months of 2009, as the United States reported its first few cases of H1N1 swine flu, China looked to its pig population with a little more enthusiasm. Internet gaming company NetEase was setting up an AI-powered pig farm hoped to radically transform pork farming thereon.
Despite being the world’s largest producer of pork, China regularly struggles to meet its demand for the meat, which makes up more than 70% of the population’s animal calorie intake. As local consumption climbs steadily alongside incomes and farmers struggle with the costs of rearing notoriously sensitive creatures, the advent of AI farming seems a natural and timely solution to China’s bovine woes.
On facilities like NetEase’s (there are now several such farms in existence), facial recognition technology allows pigs to be watched closely for signs of illness. The animals lead perfectly optimised lives, all aspects of which, from their type of swill to amount of exercise and specific stress-reducing soundtracks, are carefully calibrated to maximise their wellbeing and, ultimately, end-use. While still in its infancy, machine-learning will facilitate the continual improvement of this technology as more data is amassed over time and the AI better understands how to optimise conditions. It is already, however, far more efficient than its human counterparts at doing so, whose information and time constraints are little match for the AI’s access to endless data streams and computation time.
Such efficiency may constitute an economist’s dreams, but the promise of this new technology mounts unsurprisingly alongside substantial risks. Advances in biotechnology have done away with the variety of pig breeds and attributes that once habited China’s fields in favour of an optimised, physically standardised type. Now genetically homogenised, the pigs are less suited to their local climates and thus more vulnerable to disease, many of which affect humans–badly. The threat of zoonotic disease requires little explanation at present, and it is worth remembering that these animal-to-human viruses make up 75% of all emerging pathogens. 80% of these come from the world’s top pork producers.
While animal surveillance and product tracing capabilities provide consumers with momentary reassurance in the face of an ongoing (but subdued) African Swine Flu (ASF) outbreak, optimisation technology undoubtedly exacerbates the problem as much as it provides us with the solutions to mitigate against it. Indeed, it does not seem too great a stretch of the imagination to envision an extinction of humankind led by scientific hubris and an unyielding penchant for eco-engineering.
There is the secondary issue of big data taking agriculture out of farmers’ hands and placing it in the hands of government and big corporation. Already, widespread fear of ASF has led the Chinese government to mandate the mass-slaughtering of pigs belonging to small farms under the concern that these facilities are unable to uphold the same biosecurity standards as industrial farms. Such sentiment gives way to tech firms and large farms to lobby for tightened biosecurity standards such that farms cannot meet these without use of AI. Indeed, some already suspect that ASF has allowed the Chinese government to begin the shift away from small firms to favour more centralised production.
China will likely continue to make dramatic strides towards fully automated agricultural practise, but it is far from the only player in this space; globally, more than 90% of livestock are housed in industrial farms. Australia should take more notice of this shift in agriculture.
As an agricultural economy, optimising farming practice has the potential to allow us to better manage droughts and tackle the meteorological effects of climate change if AI is employed to predict harvest times, understand soil conditions and inform crop distribution. While the likelihood of China seeking to overtake us in beef and wool exports remains slim, we should wean ourselves of complacency–induced by enviable resource endowments–lest it allows others without natural advantages to outcompete us using AI. We are dangerously reliant on primary produce, after all.
With surveillance technology ever looming on the horizon, there seems little we can do to escape its eventual grasp. As it pertains to agriculture, we should seek to develop a robust international regulation while biotechnology is young to avoid catastrophic ecological outcomes that may eventuate if development progresses unchecked. In what time we have left, we might then muse on what big data could mean for our democratic institutions. Will we entrust AI to find the invisible hand of the market? Could it help us divine socially optimal operations and usher in a resurgence of planned economies, no longer hallmarks of failed socialist states but instead representations of the omniscience of big data?
In time, perhaps; our needs are yet a little more complicated than those of our bovine friends.
Sandhya Das Thuraisingham is undertaking a Bachelor of Philosophy, Politics and Economics at the University of Oxford. She is currently the Victorian Director of Communications at UN Youth Australia, and a member of Plan International's 2021 Youth Activist Series. Outside her work in the not-for-profit sector, Sandhya's interests span political theory, environmental economics and corporate cyber-ethics.