Director-General  QU Dongyu
A statement by FAO Director-General Qu Dongyu

Speaking Points

FAO Director-General

Ethics of Artificial Intelligence event

As prepared

 

 

Global food systems require innovative solutions to address global challenges such as ensuring food security and nutrition for all, contributing to inclusive growth, and managing natural resources sustainably.

Digitalization can play an increasingly important role in addressing these challenges and help meet the Sustainable Development Goals, especially SDGs 1 and 2 – eradicating poverty and zero hunger.

And by 2050, global agriculture has to provide 50 percent more food to feed a growing and increasingly more affluent population, and we have to meet this target sustainably.

The new technological frontier, Artificial Intelligence, (AI), can have a tremendous positive impact, making agriculture more productive and sustainable.

On the other hand, the opportunities AI offers can also give rise to economic, social and ethical challenges and risks.

FAO is working and committed to innovation-powered solutions through digital technologies, Big Data, and Artificial Intelligence.

And we aim to help harness their transformative power in making food systems more efficient, sustainable and inclusive.

Artificial Intelligence in Agriculture: Benefits

Agriculture can benefit significantly from Artificial Intelligence across all regions and farm structures.

Farming, transporting, processing and marketing food gives rise to complex value chains and Artificial Intelligence can help make these more efficient and sustainable.

An example of an FAO application of Artificial Intelligence that benefits smallholders is Nuru. Nuru, for instance, used machine learning and artificial intelligence, and helped tackle Fall Armyworm in sub-Saharan Africa, which threatened to devastate crops and farmers’ livelihoods.

The FAO system for earth observations (SEPAL) helps countries measure, monitor and report on forests and land use. It uses advanced Cloud computing, AI and machine learning to enable detection of small-scale changes in forests, such as those associated with illegal or unsustainable timber harvesting.

Precision Agriculture methods that are applied in farms across developed countries and emerging economies are based on Artificial Intelligence. This involves intelligent machines such as drones, sensors, robots and self-driven tractors. These machines are complemented by data and software that can ‘think’ like the farmer, ‘learn’ like the farmer and ‘solve’ problems like a farmer.

This is very different from what we know: Farmers produce food using the knowledge they acquired through experience being supported by scientific evidence provided by extension. Farmers are also key in providing, exchanging and using agricultural data and information.

Artificial Intelligence can also help farmers manage crops and livestock, detect pests and diseases, and optimize the use of labour, fertilizers, pesticides, feed and water.

All of this can have a huge impact on productivity, better management of natural resources, such as soil and water, reduction of pollution and environmental impacts and increase of food safety.

At the same time, Artificial Intelligence will have a profound impact downstream the food value chain: Transportation, food processing and the marketing and trade of food will also be affected by automated systems.

In agriculture and across all sectors, traders increasingly rely on computer algorithms to learn from economic data and identify opportunities.

We have to understand these impacts, and try to maximize the benefits, while minimizing the potential risks.

Artificial Intelligence in Agriculture: Challenges

Despite all the remarkable potential to benefit farmers and consumers, AI in agriculture comes with challenges and risks.

First of all, we have to make sure that Artificial Intelligence tools are designed, developed and used to be consistent with the universal human rights principles. This relates to linking technology with data and information, privacy issues, and the way information is disseminated. It is therefore necessary to translate human rights responsibilities into guidance for technical design by both the public and private sectors.1

In agriculture, the development of Artificial Intelligence technologies should also be guided by animal welfare principles, environmental considerations and food safety. All of this has to be taken into consideration when designing ‘intelligent’ machines and software that can ‘think’ like the human brain. To achieve this, a code of ethics to guide the development of the technology is crucial. It is also important that this code is integrated in the design and development of artificial intelligence tools.

There are other challenges which, although not ethical in nature, are related to economic and social impact in the long term. One such issue is the impact of Artificial Intelligence on agricultural labour. There is no doubt that the technology will result in less demand for labour, as many tasks on the farm can be carried out by ‘intelligent’ machines. This can mean lower wages and also what is called ‘technological unemployment’, raising the concern of inequality and negative social change.

The tractor is an example of such a technology that revolutionized agriculture. In the United States, the tractor was diffused between 1910 and 1960 replacing horses and workers. In 1910 there were just under 12 million farm workers. By 1960, this number fell to 6 million. This decline also mirrors the rural-urban migration in the country during the same period. Artificial Intelligence can also give rise to important social challenges, especially because, unlike in the case of the tractor, workers might face limited opportunities to remain employed by upgrading their skills.2

There are other issues also related to inequality and the way the gains from technology are distributed. By replacing labour, Artificial Intelligence can reduce agricultural wages and increase the gains of those who own land. As the supply of land is limited, such gains can be significant, giving rise to inequality. A large part of the gains will accrue to innovators – the developers of Artificial Intelligence tools. Depending on the magnitude of the technology’s impact on inequality, we should be ready to discuss regulations such as antitrust policies and intellectual property rights frameworks that can promote sustainable and inclusive growth in agriculture.

FAO’s efforts to tackle economic, social and ethical issues related to digital technologies

At FAO, together with the development of Artificial Intelligence tools, we work towards establishing the International Platform for Digital Food and Agriculture – an inclusive multi-stakeholder forum for identifying and discussing the potential benefits and risks of digitalization of the food and agricultural sectors.

This Platform will address an important gap between multilateral fora for the digital economy and those for food and agriculture. It will bring agriculture to global discussions on digitalization. For Artificial Intelligence, an important forum, is the AI for Good Global Summit that is organized by ITU. 

The Platform will create linkages with all stakeholders and support policy-makers with recommendations such as voluntary guidelines and best practices that can address the technical, economic, social, and ethical challenges that agriculture is facing in the context of digital technologies.

-------------------

1 Report of the Special Rapporteur on the promotion and protection of the right to freedom of opinion and expression. UN General Assembly Seventy-third session, 2018, Item 74 (b) of the provisional agenda Promotion and protection of human rights: human rights questions, including alternative approaches for improving the effective enjoyment of human rights and fundamental freedoms (available at https://www.ohchr.org/EN/Issues/FreedomOpinion/Pages/ReportGA73.aspx)

2 Data from Manuelli, R.E. & Seshadri, A. 2014. Frictionless Technology Diffusion: The Case of Tractors. American Economic Review, 104(4): 1368-1391.

Send
Print