Search the archive.

When considering AI in agriculture, you’ll face several challenges. High upfront costs and reluctance to embrace new technologies can be significant barriers.

Additionally, a lack of practical experience and the lengthy adoption process further complicate the implementation of AI solutions.

Data - Data plays a crucial role in developing AI solutions, and it is a significant bottleneck in getting labeled high-quality data to train AI models [213]. Recent surveys published by Ferrang et al. [214] and Alwis et al. [215] listed different datasets that can be used for different smart farming applications. However, they are not specially obtained from agriculture or smart farms. It is essential to represent the characteristics of agriculture in the data to develop a robust AI solution. This creates a need to obtain high-quality domain-specific datasets. One of the possible solutions to make diverse, high-quality data available is using agriculture testbeds. Recently, Batarseh et al. [216] and Oberascher et al. [217] reported soil testbeds that can help to produce agriculture-specific network and sensor data. This approach may help researchers to tackle data quality issues and provide robust AI solutions for agriculture.

Deployment – Deployment of AI solutions in real settings is another challenge observed in the agriculture domain. Most security solutions, such as malware, DoS/DDoS detection systems, and intrusion detection for UAVs, are not deployed. These security solutions are crucial for detecting and mitigating cybersecurity threats to farms. The deployment of AI solutions can also uncover the scalability and stability issues that may occur while using these solutions. Additionally, issues such as data and model drifts can occur with the deployment of models, which also needs attention to use AI as a decision-support tool.

The Agriculture Department is approaching generative artificial intelligence cautiously as it wades through questions about the technology’s efficacy, as well as potential impacts on privacy, equity and security, Chris Alvares, the department’s responsible AI officer and chief data officer, said during a Nextgov/FCW event on Wednesday.

“I do see a lot of opportunities, a lot of potential, for generative AI to help USDA be more efficient, deliver programs more effectively, understand the complexity of our organization and be able to describe that better than we can today without those tools,” he said. “But there's a lot of uncertainty around how to use those tools… in a federal agency and how to do that safely and securely, so we need to take a cautious approach.”

Several epidemics, such as the Black Death and the Spanish flu, have threatened human life throughout history; however, it is unclear if humans will remain safe from the sudden and fast spread of epidemic diseases. Moreover, the transmission characteristics of epidemics remain undiscovered. In this study, we present the results of an epidemic simulation experiment revealing the relationship between epidemic parameters and pandemic risk. To analyze the time-dependent risk and impact of epidemics, we considered two parameters for infectious diseases: the recovery time from infection and the transmission rate of the disease. Based on the epidemic simulation, we identified two important aspects of human safety with regard to the threat of a pandemic. First, humans should be safe if the fatality rate is below 100%. Second, even when the fatality rate is 100%, humans would be safe if the average degree of human social networks is below a threshold value. Nevertheless, certain diseases can potentially infect all nodes in the human social networks, and these diseases cause a pandemic when the average degree is larger than the threshold value. These results indicated that certain infectious diseases lead to human extinction and can be prevented by minimizing human contact.

1. Introduction

Although beneficial, the agriculture industry’s digital revolution comes with consequences. According to a 2018 Department of Homeland Security report, “The adoption of advanced precision agriculture technology and farm information management systems in the crop and livestock sectors is introducing new vulnerabilities into an industry which had previously been highly mechanical in nature.”17 The associated risks these technologies pose to the agriculture sector increase the exposure to cyber and other related threats and vulnerabilities.18 The report sheds light on some risks to confidentiality, integrity, and availability.

One of the confidentiality concerns is that major unmanned aircraft system “equipment manufacturers in the precision agricultural market are dominated by foreign built systems,” allowing foreign governments unfettered access to collected sensor data.19 Chinese manufacturer DJI is one example. Nathan Ord discusses the national security risks of foreign companies infusing IoT technologies into American companies and infrastructure.20 Ord explains Chinese IoT cellular modules allow for remote data exploitation and the possibility of remotely terminating connected devices. To illustrate this point, in 2022, Russian troops overran a John Deere dealership in the Russian-controlled city of Melitopol, Ukraine, stealing 27 pieces of farm machinery worth $5 million.21 The dealership was forced to access the stolen devices and disable them remotely, turning them into paperweights.

Artificial intelligence could improve agricultural productivity in myriad ways, from machines that zap weeds to crop-selection software. But its digital networks must be secure from attack or abuse, a panel of experts told senators on Tuesday. The panelists and members of the Senate Agriculture Committee agreed the technology must be within the financial grasp of farmers, possibly with the help of USDA cost-sharing funds.

 use so-called precision agriculture technology such as GPS steering and yield monitoring, drones to scout fields and monitor livestock, and robotic milking. By analyzing huge volumes of data, AI could make farmers more efficient or automate tasks. Data confidentiality has been an issue since the introduction of precision agriculture. Agricultural AI, reliant on transferring data from farm field to the cloud, raises the issue of cybersecurity.

It is vital to note the expected economic and social impact of a potential agroterror attack in order to develop thorough and efficient preparedness strategies. Historically, plant disease has precipitated ecological and social damage as evidenced by high mortality and migration rates (Velásquez et al., 2018). Perhaps the most infamous case was the Irish potato famine of the 1840s, caused by the fungus-like organism Phytophthora infestans, in which over a million individuals died of starvation or disease resulting from the infection of Ireland’s primary food source (DoChara, 2008). While modern societies are not reliant on monoculture, measures must still be taken to enhance the resistance of food crops essential for human nutrition. With an increasingly global society, national economies are intimately connected to the world market. Gill (2015) specifies that agroterrorism is not “flashy” (p. 10); while it may not produce the shock value of a bombing or other tactic, its use as economic sabotage will result in fear, instability, and loss of life and investments, including “heritage and environmental loss which may have psychological effects on populations” (Suffert et al., 2017, 225). Statistically, 1 in 6 jobs in the United States is linked to agriculture (Olson, 2012). Thus, it is vital to remember that an attack would not only inflict economic hardship on individuals working within the agriculture industry, but would inevitably spiral outward to impact the rest of the nation; the severity of the effects would depend on the type of pathogen.

In the recent past, analysts expressed concern that groups such as al-Qaeda would adopt tactics utilizing agroterrorism due to a “highly favorable cost-benefit ratio” (FAS, 2011, Section 6, para. 2); however, more recently the concern has shifted to include economic opportunists, domestic terrorists (including lone wolves and disgruntled employees), as well as militant animal rights groups (Gill, 2015). The Federal Bureau of Investigation (FBI) notes the extent of economic and social damage which an agroterrorism attack could inflict, particularly on the United States:

The agriculture industry has predominately been a mechanical and manual labor trade for centuries. With recent advancements in the technology sector, autonomous devices, sensors, and information systems are saturating the agriculture industry at an unprecedented rate. Economic investment futures forecast a significant growth in unmanned systems within the agriculture and logistics industries over the next five to 10 years.1 This growth is driven by drones and data applications, which are used to replace mechanical processes with more precise technological methods. The proliferation of unman[person]ed systems throughout the agriculture and logistics sectors increases the probability that threat actors, such as violent extremists, will exploit and weaponize commercial drones, which lack the more stringent security protocols found within the defense industry.

The threat from these substate terrorists to domestic and international food-supply depots and production facilities poses a direct threat to America’s homeland and national security. Substate terrorism denotes terrorism within five subcategories: leftist social revolutionary terrorism, right-wing terrorism, nationalist-separatist terrorism, single-issue terrorism, and religious extremist terrorism. Religious extremist terrorism is broken down further into new-religions terrorism and religious fundamentalist 35Parrott terrorism; al-Qaeda and the Islamic State in Iraq and Syria fall into the latter subcategory.2 Violent extremist organizations (VEOs) like al-Qaeda and the Islamic State in Iraq and Syria have successfully employed drone attacks in Africa, Asia, the Americas, Europe, and the Middle East increasing the probability of drone attacks in the future. In 2002, as US and coalition forces gained control of al-Qaeda territory in eastern Afghanistan, they uncovered a treasure trove of documents. “Among the thousands of documents they discovered were US agricultural documents and al Qaeda training manuals targeting agriculture,” according to Dean Olson.3 Al-Qaeda and the Islamic State in Iraq and Syria’s homegrown terrorist sympathizers have successfully carried out attacks in the United States, adding a layer of complexity to an already perilous threat. This chapter postulates within the next decade, religiously motivated terrorists (RMTs) like al-Qaeda, the Islamic State in Iraq and Syria, and their affiliates and adherents could leverage technological advancements in unmanned systems within the agriculture sector to conduct catastrophic attacks on the American agriculture industry and cause further global food insecurity and economic costs.

As compared with wider biodiversity “the special nature of agricultural biodiversity, its distinctive features, and problems needing distinctive solutions” was recognized by the Convention on Biological Diversity (CBD) in its early years (CBD, 1994), confirming the need for globally agreed management and governance measures that are distinct from those governing the preservation of wider biodiversity through, for example, Protected Areas governance measures. These measures recognize that, for example: people have a key role in sustaining and dynamically managing agricultural biodiversity; many components of agricultural biodiversity, such as the diversity of the seeds of food crops, would not exist without this dynamic management and associated indigenous and local knowledge and culture; intraspecific diversity and intravarietal diversity within productive ecosystems are at least as important as diversity between species; access to, and control over, components of agricultural biodiversity can be restricted by legal and technological means thereby potentially limiting its availability; and that many food production systems are based on species introduced from other communities, countries, and continents, creating great interdependence on agricultural biodiversity across the world.

Small-scale food providers’ perspectives on the importance of agricultural biodiversity

China has bought almost half the world’s food reserves in the last 12 months. This will cause worldwide inflation, we well as famine in some countries. This report provides us with clear information on what we should prepare for in the 21st century.

Prices will rise globally, including in the U.S. and Europe; a significant portion of lower-income states will be doomed to famine. The fact that the U.S. and Europe are gravely concerned about inflation for the first time is an indication that this process has already started.

Inevitably, the pursuit of these benefits will also involve risks. In this section, we highlight various possible unintended consequences of the applications of AI in agriculture and the impacts for agriculture of developments in AI elsewhere. We describe these as “risks” rather than “costs” because it is not certain that each will eventuate (or to what extent they will eventuate if they do eventuate) and because some of them, especially those involving changes in power relations or transformations in social practices, are not necessarily negative for all parties concerned.Footnote3 We have chosen to spend more time on the risks of AI than the benefits, not because we think that the risks are more likely than the benefits, or will outweigh the benefits, but because the incentives, particularly in the private sector, to realise the benefits are stronger, and more likely to be effective, than are the incentives to avoid the risks, which may require acting in the collective interest rather than in the pursuit of private gain. For this reason, we believe it is important to clearly identify the risks and draw them to the attention of those actors and institutions that might have the power to mitigate them. The risks should be of especial concern to governments, which are charged with serving the public good, and to the broader community who may be affected by them. Although we do not treat it separately below – not being a risk “of” AI – it is also clear that failure to secure the potential benefits from AI should itself be counted as a risk that should be taken seriously in deliberations about these matters.

3.1. Technologically mediated unemployment

For pool technologies, patent applications declined after the creation of a pool, both in absolute terms and relative to alternative definitions of the control. In pool subclasses, patent applications declined from 2.54 per subclass and year before the creation of a pool to 2.40 afterwards (Table 2), and from 2.80 to 2.48 within a 10-year window before and after the creation of a pool (Figure 1). In cross-reference subclasses, patent applications increased from 2.70 before the creation of a pool to 2.94 afterwards (Table 2), and from 2.94 to 3.02 within a 10-year window (Figure 1). In the broader control group that includes all subclasses in the same class, patent applications increased from 1.00 per year before the creation of a pool to 1.11 afterwards (Table 2).

Certain industries are said to be forced to rely on pools.121 As Steven Carlson argues, “in standard-dependent technologies, such as the MPEG protocol, there is essentially no choice but to deal with the patent pool.”122 This can also amount to a violation of abuse of dominance provisions.

Where horizontal competitors pool their patents, competition for licensees is reduced—and even where a patent holder retains the right to license outside of the pool there may not be much incentive to do so.123 Where the right to license out of the pool does not exist, this could discourage innovation.124 Moreover, where the pool requires the members to grant licences to each other for present and future technologies (“grant-back”) at minimal cost or without an adequate reward, this may further disincentivise members to engage in research and development or cost-reducing process and product innovations and lead them to just freeride on the works of others.125

On the flipside, patent pools have certain disadvantages associated with them as well. Patent pools may work ineffectively if there are regulation issues among the members of the pool. For instance, if members of the patent pool collectively disagree when it comes to governance issues or if prominent patent owners do not join the pool, the patent pool may be greatly unavailing. Patent pools also pose various issues related to antitrust law. Issues related to antitrust law arises when competitors are patent holders of overlapping patent rights. Consequently, neither of the competitors will end up developing the technology related to the patents they own without infringing each other’s patents. Primarily, there are two types of patent pools. An open patent pool may be defined as a patent pool that is largely controlled by management companies. These companies may control the patent pool by adding patent owners who meet certain criteria from time to time. The management company also coordinates among the patent owners and looks into licensing operations that occur within the patent pool. On the other hand, a closed patent pool may be defined as a patent pool wherein a patent administrator is selected. The patent administrator here regulates the pool, procures licensees, and ensures that the patent pool is conducted in a lawful manner.

In contrast, a pool comprised of pure substitute patents is very likely to offer limited technological choices to its own members and third parties. Such limitations are likewise capable of constraining innovation that generally relies on the supply of a variety of technologies. In a long run, it will likely lead to higher prices of, and constrain access to, standard products in downstream market as also noted by Gallini (2011). This is a negative consequence likely to arise from the scarcity of technologies alluded to early, which can have an opposite effect from that which might have otherwise occurred with a diverse supply of technologies. The latter generally help to lower the costs of production of standard products. From this standpoint, it would not be unseasonable to perceive pools loaded with substitute patents as a disguised stratagem that firms can use to control or fix the market prices, foreclose potential competitors and/or restrict downstream technology and knowledge spillover and licensing process.

The second type of exclusivity concerns the outside developer of technology who wants to include its patents in the pool but is denied. One clearly procompetitive explanation for the pool's refusal to accept an offeror's technology is that the excluded technology is inferior to or no better than technology that the pool already has. There are good reasons for a pool not to accept an inferior substitute technology; namely, it will not make the pool products better and it might raise fears of collusion. If the prospective licensor's technology is simply unwanted by pool members, that also indicates lack of competitive harm. The firms in the pool have a right to make their own technology choices, even when they do so by agreement. Indeed, even a dominant network should not have an obligation to accept a set of licenses to a technology that its members do not want, and no concern about competition arises.

Competitive harm can occur when some pool members already have technology in the pool that competes with the excluded firm's superior technology. In that case the pool can operate as an anticompetitive boycott.23 Assessing the merits of such claims is difficult, because it may require the fact finder to assess the merits of the excluded technology. As a result it is essential to make sure that structural prerequisites are present. For example, an essential element in such a claim would be the existence of an influential pool member or members who own competing technology and have sufficient power to foist an inferior technology on others, perhaps by manipulating the process.24

However, patent pools may also be restrictive of competition, which is perhaps not so surprising given they are basically agreements between competitors. The collective pricing of pooled patents, greater possibilities for collusion (e.g. through using the mechanism of the pool to exchange competitively sensitive information118) and generally larger number of market participants mean that patent pools need to be scrutinized for antitrust concerns. Specifically, they can lead to creation of price fixing cartels119, market allocation, and reduction of innovation in the form of standard setting or foreclosure of alternative technologies or barriers to the entry of new and improved technologies.120

While most patent pools provide overall social benefits, certain practices or structures within them can pose competitive challenges. The primary antitrust risk associated with pooling is collusion, which depends on two key factors. First is the market structure and the pool's influence within its market. Second is the nature of pricing and exclusivity arrangements within the pool (Merges, 1996).

Regarding market structure, cartels can only profit if they possess market power. Patent pools often cover entire industries, but they can also be limited to a small group of firms within a larger market that decide to share a unique technology. Whether this technology allows them to raise prices above costs is an empirical question. Naked price fixing is generally considered illegal per se. Courts have been cautious in applying this per se rule to agreements within pools, instead evaluating softer forms of price-affecting conduct under antitrust's rule of reason. This approach requires estimating the pool's market power.

Furthermore, such pools can distort the equilibrium of royalties payable to certain licensors (particularly those holding essential patents), if such distribution is a dependent variable of the aver-all population of patents present in a pool. Members holding large amounts of substitute patents are likely to shift such equilibrium to their side and internalise higher and undue benefits to the detriment of the other members holding essential patents. In the end, the owners of essential patents will likely exit such pools due to their respective imbalanced and unfair royalty-sharing scheme. The owners of essential patents, who elect to remain in such pools are likely to relax their expenditures in new technology development and dissemination initiatives because the conventionally pursued and expected payoffs will be prone to distortions, likely negligible and insufficient to sustain their innovation efforts.

The existence of a pool may distort innovation behavior. For example, a firm might decide to protect a particular innovation with a trade secret rather than a patent in order to avoid committing it to a pool. Alternatively, it may rely on simple first mover advantages in order to avoid patent disclosure requirements. Another distortion results from how royalties are divided up. For example, if the pool divides royalties by the number of patents, then pool members have an incentive to make patents thinner in order to increase their number. (Baron and Delcamp, 2015) Finally, one must distinguish anticipatory innovation which occurs when a patent pool is announced (Lerner & Tirole, 2004; Dequiedt & Versaevel, 2013), from equilibrium innovation which occurs on an ongoing basis after the pool is in operation (Baron & Pohlmann, 2015).

A well-functioning patent pool should reduce the costs of sharing technology, which it can do by making it easier for manufacturers to assemble complements, eliminate double marginalization, or reduce the costs of defining or defending patent boundaries. Since all of these costs operate as a deadweight loss on innovation one would think that the impact of a well designed pool is positive. On the other side, pooling may reduce the appropriation value of innovations, particularly if pool members must commit in advance to dedicate new patents to the pool. In that case, rather than obtaining the entire return the pool member must contemplate sharing with others, and this may create common pool management problems. Each firm will have a tendency to shirk, profiting instead from the inventions of others.