IEEE Technology and Society Magazine - December 2020 - 32

data quality is central in developing an efficient system
for dairy farming, as the data forms the basis to inform
any of the farmers' decision-making. Most development
thought and effort thus goes towards the sensors,
assessing what kinds of data can be acquired from the
cow, and how these data can be of value to farmers.
Data quality is thus the name of the game, and development of the technology focuses on capturing as much
meaningful actionable data to support understanding
livestock physiology and behavior.
This aligns well with the priorities of farmers, whose
main focus is to use data to understand their livestock  - tracking individual cows to know when to
inseminate, generating patterns of feed intake, and
understanding behavioral factors affecting their health
and welfare. Effectively, the farmer noted that they and
their colleagues make all decisions based on aggregated data, so their priority is to get high quality data, and
as much of it as possible.
The farmer we interviewed was relatively new to the
profession, having been a dairy farmer only for the past
three years. Of those three years, he has always used
sensor technologies, never considering it a challenge to
implement and use it, noting that " farmers have always
used technologies to understand their cows for as long
as they have existed. " The farm they work at has been
in business for over 15 years. Different kinds of technologies have come and gone since then, with farm management adopting new technologies when it allowed
them to obtain more data about their cattle. Their
requirements are to obtain more and more data, effectively to build a detailed model of each individual cow
from birth, to analyze how they will behave, and whether they will become an effective dairy cow. When
pressed on what future technologies could, or should,
hold, the answer was similar: more, detailed data,
going to the level of individual genetic information of
each cow. They explained this would be great from a
business point of view, as the cost involved in raising a
cow to adulthood is significant, and, should they not be
as productive as desired, selling them as meat cows
only allowed recouping of some of the expenses
involved in raising them.
Given the focus on data, basic cybersecurity efforts
proposed in literature (e.g., [8], [34]) are important to
consider, such as at least ensuring 1) abnormal measurement detection, 2) access control, and 3) encryption. Detection of abnormal measurements is indeed
seen as vital by the developer given the data-driven
nature of the technology and significant development
efforts go towards it. Access control is currently implemented, but limited to sensitive data, such as current
research or commercial data. However, most, if not all
data is freely shared owing to the socio-cultural origins

32

of the agriculture sector in Israel, leading stakeholders
to not perceive significant threats that a malicious actor
could carry out with livestock data. Encryption, however, is not as widely used, likely due to similar attitudes
of data being freely shared and perceptions of that it
would be counterproductive to avoid easy access to it.

Using The Technology-Not All Threats are
Perceived Equally by Vendors and Users
We further built a more accurate understanding of what
the threats the food sector in general, and digitized food
production organizations (such as precision livestock
farming) in particular, face in Israel. We interviewed a
managing research director of cybersecurity, who
focused on modeling and understanding cyberattacks
and risks, including those present in critical infrastructures, including the food sector.
The primary threat faced by food production companies is similar to that in many other contexts. There is
little evidence of malicious nation state actors or their
proxies engaged in cyberattacks, but there is real-world
evidence of attacks conducted by smaller scale cyber
criminals motivated by economic incentives (i.e., theft
of data perceived as commercially sensitive, or blackmailing). As one could expect from the smaller scale,
these do not focus on the operational technologies used
by the food sector, as developing and conducting
attacks against new technology requires an investment
of time and effort to understand the hardware, build
attacks, and find appropriate vectors to conduct them.
Instead, attackers attempt typical attacks like ransomware and phishing to affect the information technology
layer and blackmail the operators of this IT layer into
making payments to unlock the hardware or avoid
release of sensitive data. However, a more direct threat
vector has been used as well, where cyber criminals
attempt to compromise the digital infrastructure of
farms through lateral movements, again attacking the
information technology layer primarily (to the point of
overruling almost any other attack vector) through insecure configuration. That is, attackers will attempt to log
in to IT systems using standard credentials, or attempt
to abuse management systems such as auto-backup
scripts with embedded credentials. The key threat to
technologies in the food sector faced in Israel is thus
towards the information technology layer, coming primarily from an economic perspective.
From interviewing the vendor of the system we studied, these attack vectors seem particularly salient, as
only one real threat was perceived: real-time data loss,
by whatever means. Should the system crash, whether
due to software bugs, environmental factors, or cyberattacks, this would be a major problem. Given that data is
stored on the information technology layer, it thus

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

DECEMBER 2020
IEEE Technology and Society Magazine - December 2020

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2020
