The Impact of AI on Agricultural productivity and employment

in rural Zimbabwe

A Case of Mushangwe Rural Farming Community

A Dissertation Submitted

IN PARTIAL FULFILLMENT OF A
MASTER’S IN ACCOUNTING AND FINANCE

ARTHUR MCNALLY ZHOU

R1712D4017837

Supervisor: Mary Mandiringana

UU-MBA-71-ZM-2956

i
Abstract

Artificial intelligence has revolutionized many spheres of human society and Agriculture is
no exception. With a bulging global population coupled with worsening weather patterns due
to climate change there has never been a greater need for human ingenuity in Agriculture.
Agricultural production already accounts for 37,7% of the world’s surface and with little
room for expansion maximizing output with minimum inputs is key and Artificial
Intelligence (AI) has emerged as the most efficient and effective solution. The research
analyzes the impact of Artificial Intelligence on agricultural productivity and employment of
small holder rural farming communities and the impact of AI on Sustainable Agriculture.
Lastly the study explores farmer’s perception of AI as an effective tool in combating
contemporary agricultural challenges.
Mushangwe rural farming community in Zimbabwe was used as a case study because it has
similar typology with most Zimbabwean rural farming communities. Secondary research data
was used to collect data and analysis was done by a fusion of quantitative and qualitative
methods. Data was mainly derived from the internal records of Beitbridge Juicing company
which contracts the smallholder tomato farmers in Mushangwe community. Semi structured
online questionnaires, data from NGOs, Government institutions, peer reviewed journals and
academic papers are also used to satisfy the objectives of the research.
The target population consists of smallholder tomato farmers from Mushangwe and
governmental and non-governmental officials. The result of the study shows that AI has a
positive impact on productivity but leads to a decrease in employment opportunities in the
local communities. Results also show that AI can improve the sustainability of agriculture.
The study concludes that technology in agriculture can increase productivity by efficient use
of water resources, early disease detection and maximizing the quality of produce.

Keywords: Artificial intelligence, agricultural productivity, smallholder farmers

ii
Acknowledgements

My deepest and most sincere gratitude goes to my Supervisor, Mary Mandiringana for
providing her guidance, expertise, and support throughout this academic research journey.
Your contribution to the completion of this thesis is invaluable. My heartfelt gratitude also
goes to the Lecturers that have been part of this process for 2 years whose contribution has
been immense.

I would love to reserve special thanks to my family and friends who have been a pillar of
strength, motivation and encouragement throughout this journey. Thanks also goes to
Beitbridge Juicing Company for providing the necessary information without which the
research would not be possible. Respondents and governmental officials who also took their
time to assist in the study a big thank you is owed to them.

Last but not least UNICAF University for providing me with this unique opportunity to
further my studies and personal development. To the non-academic staff who played a pivotal
supporting role in enabling the seamless academic experience over the 2 years I thank you.

Arthur

November 2021

iii
ABBREVIATIONS

AI - Artificial Intelligence

UN – UNITED NATIONS

RFC – Rural Farming Community

DEWS - Drought early warning system

NGO – Non-Governmental Organizations

GDP- Gross Domestic Product

FAO- Food and Agriculture Organization

SSA- Sub Saharan Africa

iv
TABLE OF CONTENTS
Abstract…………………………………………………………………………………….ii

Acknowledgements………………………………………………………………………..iii

Abbreviations……………………………………………………………………………....iv

List of Figures……………………………………………………………………………..vii

List of Tables……………………………………………………………………………..viii
CHAPTER 1 1.1Introduction .................................................…………………………..1
1.2 Problem statement……………………………………………………..................2

1.3 Research Questions………………………………………………………………3

1.4 Research Objectives……………………………………………………………...3

1.5 Significance of Study……………………………………………………….........3

1.6 Hypothesis Testing………………………………………………….…………...4

2.0 LITERATURE REVIEW ……………………………………………………………5

2.1 Theoretical Framework ………………………………………………………….5

2.1 AI and Rural Employment……………………………………………………….6

2.2.1 Impact of AI on agricultural productivity…………….………………..............6

2.2.2 AI and rural productivity in Sub Saharan Africa ……………………………...7

2.2.3 AI and Rural Employment……………………………………………………..8

2.3 Challenges and Opportunities of implementing AI…………….……………………...9

2.4 Opportunities of implementing AI……………………………………………………..9

2.5 AI on sustainable agricultural………………………………………………………....10

3.0 RESEARCH METHODOLOGY…………………………………………...............11

3.1 Introduction……………………………………………………………………………11

3.2 Research method………………………………………………………………………12

3.3 Research Design……………………………………………………………………….12

3.3.1 Sample Size and Sampling Techniques………………………………….............12

3.3.2 Data collection tools……………………………………………………………..13

v
 3.3.3 Semi-Structured Interview……………………………………………………….13

3.2.4 Data Analysis……………………………………………………………………..13

3.4 Study Limitations and Constraints ……………………………………………………...14

3.5 Ethical Considerations…………………………………………………………………..14

4.0 CHAPTER 4 …………………………………………………………………………...15

4.1 AI and Productivity …………………….……………………………………………….15

4.2 AI and the employment rate……………………………………………………………..18

4.3 Sustainable Agriculture………………………………………………………………….20

4.4 AI useful in combating challenges………………………………………………………21

4.5 Results and Discussion…………………………………………………………………..23

4.5.1 Impact of AI on productivity………………………………………………………23
4.5.2 AI and Employment rate…………………………………………………………...23
4.5.3 AI and Sustainable Agriculture……………………………………………………24
4.5.4 AI and combating agricultural challenges…………………………………………24
5.0 Conclusion, Recommendations and Limitations.…………………………………….25

5.1 Conclusion…………………………………………………………………………...25
5.2 Recommendations on Future Research……………………………………………..25
5.3 Limitations of the study ……………………………………………………………..26

References…………………………………………………………………………………..27

Appendices…………………………………………………………………………………..32

vi
List of Tables

Table 4.1 AI and Productivity in Mushangwe……………………………………………….15

Table 4.2 Yield (kg/ha) for Maize under Command Agriculture ……………………………17

Table 4.3 Tomato Employees………………………………………………………………...18

Table 4.4 Impact of AI on Sustainable Agriculture………………………………………….20

vii
List of Figures
Fig 3.1 Location of Study area….. ……………………………………………………..11

Fig 4.1 AI and Productivity in Mashonaland Central…………………………………..16

Fig 4.2 AI and Small-Scale Commercial Farms………………………………………..17

Fig 4.3 AI and Employment rate………………………………………………………..19

Fig 4.4 Solar Powered Booster Sprayer………………………………………………...20

Fig 4.5 Farmer’s Perception of AI……………………………………………………...21

viii
CHAPTER 1

1.1 Introduction

The Industrial revolution from the 20th to the 21st century has seen humankind make
unparalleled development with the aid of Artificial intelligence. Manyika et. al (2017) defines
Artificial Intelligence as the “ability of machines to carry out cognitive functions that would
normally be associated with human intelligence such as perceiving, reasoning, learning and
solving problems”. It is the replication of human intelligence in machines by programming
machines to execute tasks that would have been done by people.

According to Zeeman and Rodriguez (2019) and Feng et al (2011) AI is being applied to
many agricultural areas including land and water use, agricultural production, machinery and
sustainable agriculture. Smith (2020) states that the availability of data, algorithms and
computational ability can have a positive impact on productivity. Population growth is
estimated to reach 10 billion by 2050 (FAO 2017) putting tremendous pressure on the
agricultural sector to meet demand. Climate change has also exacerbated the need for AI in
agricultural. It’s adverse effects on agriculture are occurring on a global scale but its impact is
unevenly distributed with Sub-Saharan Africa being the most affected region. Just like
most African countries Zimbabwe is an Agrarian economy and is heavily reliant on the
Agricultural sector.

The effects of climate change are more pronounced in rural areas where agriculture is the sole
economic activity for most. The increased vulnerability is not only caused by Climate change
but by other socio-economic issues which correlate with adaptation capacities. FAO 2016
puts the percentage of rural people relying on rainfed agriculture at 75% further increasing
their vulnerability to climate change making the introduction of smart farming more
important. However, the emergence of AI in a labor-intensive sector such as agriculture can
have a telling impact on the employment rate in farming communities.

This research paper focuses on the impact artificial intelligence has on agricultural
productivity and the relationship between AI and employment. Previous research has focused
on AI in large scale commercial farms whilst other researchers have focused on
implementation and adaptation of AI in developing countries. However, little research has
been done on what impact AI has on small holder rural farmers and the implementation of AI
on employment opportunities of the locals.

1
1.2 Problem Statement

The agricultural sector has faced many challenges due to climate change, dwindling arable
land, diseases and a rapidly growing population has shifted agriculture to a new paradigm S
Liu (2020). The increased vulnerability in Sub-Saharan Africa especially in rural areas is due
to the overreliance on agriculture and the prevailing socio-economic issues that make climate
adaptation difficult. AI has emerged as the most efficient solution in tackling challenges in
the agricultural sector.

Just like many farming communities in rural Zimbabwe, climate change has been the biggest
problem. In Zimbabwe, the weather patterns for the last 2 decades have been inconsistent and
erratic bringing about devastating weather disasters like Cyclone Elin and the droughts of
2002 and 2008. Many rural farming communities are enmeshed in a vicious cycle of poverty,
climate change, a lack of funding and insufficient government assistance. The Zimbabwean
government being of a developing country has narrow capacity for climate change policy
analysis, implementation and mitigation and adaptation strategies Ndebele and Mubaya
(2015). Mushangwe rural farming community like most rural communities has not fared well
with declines in tomato production the community’s cash crop and other supporting crops
causing food security concerns and depleted income streams thereby warranting the need for
AI as a solution to increase productivity.

AI has come in the form of automated irrigation systems, smart greenhouse technology and
machine learning algorithms. However, agriculture is labor intensive, and the advent of new
technology will have a significant impact on the employment opportunities of rural farming
communities. Input of fertilizers, tillage of land and harvesting will be automated leaving a
whole swart of labour displaced. The purpose of this research is to analyze the impact AI has
on the agricultural productivity and employment opportunities in rural communities using
Mushangwe community as a case study.

2
1.3 Research Questions

How has AI impacted agricultural productivity in rural Zimbabwe's small holder farming
communities?

What is the relationship between AI and Employment in Smallholder Rural Farming

Communities (RFC)?

What is the perception of Smallholder farmers towards AI as an effective tool in combating

agricultural challenges currently being faced?

What is the impact of AI on Sustainable Agricultural?

1.4 Objectives

To investigate the impact of AI on agricultural productivity in Zimbabwe’s Smallholder RFC

To establish a relationship between AI and employment in Zimbabwe’s smallholder RFC

To examine the perception of Smallholder farmers towards AI as an effective tool in

combating agricultural challenges.

To determine the impact of AI on Sustainable Agricultural

1.5 Significance of Study

Agriculture is the backbone of most Sub-Saharan countries accounting for 70% of

employment, 40% of exports and manufacturing sectors who are reliant on agricultural
produce as raw materials World bank (2008). This study is significant to rural communal
farmers because it seeks to investigate the impact AI has on the productivity of rural farmers.
Information generated by this research will help farmers in decision making, Banks and other
financial institution when making lending decisions.

AI in rural areas is still a novice area of research with little to no data for decision makers to
use. Findings of this research will help governmental institutions and NGOs in creating
effective policies and strategies to tackle the implementation of such technologies in rural
areas. This research will also try to fill the gap in literature on the impact of AI on the
environment and how smart technologies can aid sustainable agriculture. Whilst there has
been implementation of AI in the rural parts of Zimbabwe little attention has been paid on the
effect AI has on the employment opportunities of the locals. The study will explore the
3
relationship between implementation of AI and employment opportunities. The study can be
used by other researchers as a foundation to explore other facets of this area of study that
have not been tackled in this research.

1.6 Research Hypothesis

Most of the empirical literature on AI and productivity point to a positive relation hence this
study hypothesizes a positive relation between AI and agricultural productivity. The study
will also test the kind of relationship that exists between AI and employment rate and if AI
has a positive impact on the Sustainability of Agriculture The hypothesizes is as follows
i) AI has an impact on agricultural productivity
ii) There is a positive relationship between AI and employment
iii) AI has an impact on Sustainable Agriculture
iv) Smallholder farmers believe AI is effective in solving agricultural challenges

Depending on the results of the study the Null Hypothesis or Alternate Hypothesis will be
rejected or accepted.

4
CHAPTER 2

2.0 LITERATURE REVIEW

Artificial Intelligence has had a significant impact on all spheres of human life including the
economic, social, and political. Manyika et. al (2017) defines Artificial Intelligence as the
“ability of machines to carry out cognitive functions that would normally be associated with
human intelligence such as perceiving, reasoning, learning and solving problems”. Rusell and
Norvig (2003) builds on the definition defining Artificial Intelligence as intelligence
exhibited by an artificial agent. The literature review examines the impact AI has on the
agricultural productivity in rural communal areas. Additionally, it seeks to determine the
relationship between AI and employment in rural farming communities and the impact AI has
on the sustainability of agriculture.

2.1 Theoretical Framework

The overarching theory guiding this study is the Precision Agricultural (PA) a subgroup of
Artificial Intelligence coined Lindblom et al (2017). The PA is a conceptualized system based
on rearranging the agricultural system towards a low input high output efficiency sustainable
system using technology. Its built-on Solow’s earliest model of factors of production plus
technology leading to long term growth. Solow’s Model (1957) tries to estimate the amount
of technological growth that is directly attributable to economic growth. Similar thoughts are
shared by Talaviya (2020) stating AI based solutions have enabled farmers to produce more
output with less input. However, a study carried by FAO (2000) found that in 7 African
countries farm mechanization and technology didn’t result in greater productivity. FAO
(2000) emphasizes that poor implementation strategies were the main reasons for failure.
Chisango (2011) goes further to state that in Kenya and Zimbabwe adoption levels of AI and
mechanization are even declining.

Wolf et al (2017) is of the view that AI has a positive impact on productivity but however
there are adoption issues of new technology amongst small scale farmers because of the lack
of capacity. D Bentvoglio (2021) goes further to note that whilst adoption capacities of AI
differ currently in the long run farmers that fail to adapt will struggle to keep their farms
economically viable. Other theoretical models like the Cobb-Douglas model have been used

5
to represent the relationship between output and a set of inputs. For the purposes of this study
precision agriculture theory has been chosen because it highlights the role of technology in
increasing productivity and doing so in a sustainable manner.

2.2.1 Impact of AI on global agricultural productivity

According to Kim et al (2008) AI has managed to improve the agricultural sector by

providing efficient solutions to the challenges faced in different fields of agriculture like the
crop yield, irrigation, soil content sensing, crop- monitoring, weeding, crop establishment. Eli
Chukwu (2019) concurs as he states that AI ensures that there is proper treatment of the soil,
efficient storage of big data, and reducing the knowledge gap between farmers and
technology. With global populations soaring the pressure on the agricultural sector has never
been higher and Artificial intelligence has the potential to be the solution.

AI is a fairly novice and emerging technology in the agricultural sector and has seen
tremendous momentum in the development and commercialization of agricultural
technologies Talavaya et al (2020). Yang et al (2007) states that AI technology has enhanced
crop production, harvesting, processing, marketing and improved real time monitoring.
Nandal and Rai (1986) do not share similar sentiments arguing that failures of policy
prescriptions during farm mechanizations in the 1970s in Africa have left small holder rural
farmers with little faith in farm technology. In India the green revolution of the 1960s that
made use of machines and pesticides increased food productivity and dealt with the food
shortage crisis Y Kinha (2020). In the USA, most start-ups in agriculture are adapting AI
with the market of Agro-technologies expected to reach $1,5 billion by 2025 Gupta (2019).

2.2.2 AI and rural productivity in Sub Saharan Africa (SSA)

People who reside in rural areas are most vulnerable to challenges faced by the agricultural
hence AI application can address most of the challenges DAIA (2020). Rural SSA has been
seen to be the most likely to be affected because of its reliance on Agriculture and lack of
adaptability due to poverty. In South Africa there are over 4 million rural smallholder farmers
who are largely black and aren’t as productive as their white counterparts. The difference
being the mechanization and AI equipment of white South African farmers enabling them to
produce more on less land. AI will help reduce the difference by availing the same
technology to black rural farmers productivity P Aguera et al (2020). In Kenya AI was used
to create a genetically modified water-efficient maize seed to counter the rain inconsistency

6
Whitfield et al (2015). Simoya (2018) explains how AI has led to an increase in productivity
in rural areas of Zimbabwe through efficient distribution of farming inputs by government
programs. One could argue that success of AI in Africa is not so obvious as Aggarwal (1983)
stated that SSA has had little success with technology in farming particularly for smallholder
farmers. Panpatte 2018 stating that urbanization will force rural farmers to embrace smart
farming resulting in increased productivity. Masinde 2014 argues that one of the most
significant AI tools in agriculture is the Drought early warning system (DEWS) which
predicts when, where and the severity of a drought.

In Churumanzu a rural farming community in the Midlands Province has been severely
affected by climate change however irrigation and mechanizations schemes have led to an
increase in productivity C Gota (2016). A study carried out by Moyo et al (2016) showed that
investment in farm mechanizations and irrigation will yield an increase in productivity.
Simoya concurs as he stated that smart agriculture in the form of greenhouses has led to
higher productivity with less water resources and land used. What is missing from previous
research is a comprehensive study of how AI impacts productivity in rural areas.

2.2.3 AI and Rural Employment

P Aguera et al (2020) states that the agricultural sector is a key contributor to the GDP of
African countries and remains a pivotal provision of employment especially in rural areas.
The implementation of AI in rural areas will have a positive impact on the employment
opportunities in the area Simoya (2018). He adds that in the case of Zimbabwe the high
literacy rate makes the implementation and adoption of AI easier. Bughin (2017) also concurs
as he states that new farming technology will create new jobs that never existed. However,
Soffar (2019) argues that AI might result in unemployment with an estimated 1,5billion
people working in the agricultural sector, monotonous jobs could easily be automated leaving
a lot of people obsolete.

In South Africa were close to 9 million people are either directly or indirectly employed in
the agricultural sector the implementation of AI is likely to have a big impact on
employment. This is in convergence with Simoya who states that countries like Zimbabwe
will be able to practice winter farming because of the controlled temperature in smart
greenhouses thereby adding more employment opportunities. However, in an African context
Mellor 1984 argues that mechanized farming technology leads to a substantial displacement
of jobs leading to high unemployment. Evidence in SSA shows AI has a positive impact on
7
productivity but mechanizations of farms should not be a replacement for workers but
complementary tool that can benefit all stakeholders Gillwald et al (2019). Farkas (2003) also
argues that AI will increase productivity hence more people will be needed in the harvesting
process.

In the case of Zimbabwe, the literacy level is one of the highest in Africa. Because of the high
literacy levels in Zimbabwe the transition to machine learning and mechanized agriculture
will be smoother and the lost employment will be far less than the new jobs created. A series
of chain employment is likely to emerge after AI is implemented as training on handling farm
machinery and gathering and interpreting data will increase the quantity and quality of
employment opportunities. This paper aims to determine how employment opportunities may
rise or fall from the implementation of AI.

2.3 Challenges of implementing AI in rural areas

Whilst there is ample evidence to show the benefits of AI to smallholder farmers Jellason et
al (2021) questions the readiness of Africa’s capacity to benefit from smart agriculture
because of their rudimentary infrastructure. T Chinembiri (2020)’s view differs arguing that
modalities are already in place where smallholder farmers are already hiring agricultural
technologies. Dosso (2021) is of the view that current investment in agricultural
infrastructure does not meet the threshold for beneficiation from the high-tech revolution. In
Mkoba farming community showed that implementation of AI does not necessarily result in
productivity as prevailing socioeconomic issues such as small pot land and land tenure are
barriers to success Moyo et al (2016). The practice of intercropping and same field crop
diversity also makes adoption of AI difficult.
Secondly, Jellason also notes that farmers might be unwilling to take up new technology even
though it benefits them because of a lack of awareness and mistrust. Particularly in SSA
where agriculture is not only an economic activity, but part of culture adoption of new ways
and techniques might erase culture and knowledge Rose et al (2021). Whilst many rural
farmers have mobile phones 47% of them have no internet access. This is another significant
challenge as the usefulness of mobile technology which has been instrumental in agricultural
technology is predicated upon access to internet.

Whilst many scholars agree that the Agri- tech revolution will yield greater productivity little
consideration has been given to social and ethical implications of adopting AI in agriculture
Jellason (2021). The African continent is plagued by institutional corruption and inconsistent
8
policy decisions that have rendered previous attempts at farm mechanizations ineffective.
Finally, access to credit and funding has proven to be one of the biggest stumbling blocks in
implementing AI in rural Zimbabwe. The purpose of this study is to examine the perception
of Smallholder farmers towards AI as an effective tool in combating agricultural challenges.

2.4 Opportunities of implementing AI in rural areas

SSA faces a multi variant constraints ranging from access to credit, scale and poor
institutions however there seems to be considerable scope for AI adoption in rural SSA
Robinson (2021). One of the opportunities to be taken advantage of in rural areas is machine
learning. In Zambia over 22 000 farmers have been using mobile technology in farming. The
use of the app AgriPredict helps farmers get instant information about weather predictions
and plant diseases ahead of time BBC (2020). Jellason et al (2021) disagrees suggesting that
the implementation of AI in rural areas has significant challenges such as scale. However, in
Zimbabwe the application Ecofarmer has also been used to give farmers real time
information about inputs, fertilizer use, and weather indexed insurance CTA Report (2019).

According to the United Nation Education Statistics (2014) the literacy levels amongst adults
and youth in Zimbabwe is 88% and 90% respectively. This means that there is ample of
opportunity to exploit the abundant labor skill available. Job creation through AI will see
high level jobs such as accountants, software engineers, drone pilots and Agronomist being
filled by the highly educated youth Melaine (2001). Jellason 2021 also agrees with the vast
opportunities created by AI as he states that smallholder farmers in Zimbabwe are benefiting
from aerial chemical spraying, soil mapping and harvesting.
However, Chiputwa et al 2011 found that there was a resistance in some rural smallholder
farmers to adopt agricultural technologies which use less harrowing and mulching. Research
on the impact of AI in rural areas should not be centered around a simple cost benefit analysis
of income and yields but a multi variant one including the social and cultural impact of
modernized agriculture Aguera et al (2020).

9
2.5 Impact of AI on sustainable agricultural practices

There has been a focus to move away from high yield farming which involves excessive
amounts of chemicals and genetically modified seeds to smart based agriculture that is not
only nutritious but environmentally sustainable P Kishore (2020). This is supported by the
overarching theory of this research which is Lindblom et al (2017)’s Precision Agriculture
which emphasizes on sustainable productivity. Agriculture is the biggest consumer of water
and releases greenhouse gases and methane causing damage to an already deteriorating
environment FAO (2013). Whilst the quantification of the effects of Agriculture on the
environment are difficult to prove there is clear evidence of the negative impacts Industrial
agriculture has had Pretty et al (2001). The increasing demand for food whilst employing
ecologically friendly agricultural practices needs innovation through AI to underpin it N.P
Robinson et al (2021).

Since most smallholder farmers depend on underground water AI can determine the moisture
level of the soil and control the water supply saving up to 75% of scarce freshwater. New
smart technologies such as remote sensors and the GPS enabled specific locations of weeds
thereby avoiding uniform spraying cutting back on 65% of chemicals. J Gupta (2019)
reiterates stating that AI is shifting the way agriculture is done thereby avoiding unintended
effects on the environment with an estimated 20% decrease in emissions. However, this new
technology is very expensive, and most smallholder rural farmers cannot afford it P Kishore
(2020). Jellason et al forms a different opinion stating that technology is becoming cheaper
because of the availability of free high resolution satellite imagery. AI is set to increase
productivity whilst avoiding exacerbating contemporary environmental problems caused by
agriculture such as CO2 and greenhouse emissions, deforestation and depletion of fresh water
sources.

Hypothesis Development

2.5 AI and Agricultural productivity

Studies on Artificial Intelligence and agricultural productivity have found that AI increases
productivity by improving real time monitoring of crop production, harvesting, processing,
and marketing Yang et al (2007) and Eli Chukwu (2019). Farm technologies like drones,
aerial chemical sprays and automated irrigation systems have maximized output by the
10
efficient use of inputs Jellason (2021). The more investment made in AI the more agricultural
output to be expected. Hence

Ha. Artificial Intelligence increases agricultural productivity

However, an extensive study carried out by FAO (2000) on 7 sub-Saharan countries found
that AI and mechanization did not result in agricultural productivity. Jellason et al (2021) also
sites the rudimentary infrastructure in SSA as a barrier to successful adoption of AI. This
leads to the following hypothesis

H0 AI has no impact on agricultural productivity

2.5.2 AI and Employment

AI affects employment in 2 different ways, firstly by displacing workers from their current
jobs or by creating new jobs that did not exist before. Bughin (2017) and Simoya (2018) state
that new farming technology will create new jobs that never existed and increases in
productivity as a result will increase labour demands. Hence

Ha There is a positive relationship between AI and employment

Other studies have indicated that AI does not increase employment. Soffar (2019) argues that
AI results in displacement of workers who do not have the skill to transition to other areas of
employment.
H0 There is a negative relationship between AI and employment

2.6.3 AI and Sustainable Agriculture

Climate change has put the sustainability of agriculture into question posing serious
challenges to the sector. Lindblom et al (2017)’s precision agriculture focuses of increasing
agriculture in a sustainable manner using artificial intelligence. The following hypothesis can
be developed

Ha AI has a positive impact on Sustainable Agriculture

There is great difficulty in quantifying the effects of AI on the environment P Kishore (2020)
and the high cost of AI machinery means rural farmers cannot afford them therefore
H0 AI has no impact on Sustainable Agriculture

11
CHAPTER 3
RESEARCH METHODOLOGY
3.1 Introduction

This chapter will outline the research strategy, research method, sample size of the target
population and the type of data analysis. The study is secondary research and analyses the
impact of AI on productivity and employment opportunities of Rural farming communities in
Mushangwe farming community. Secondary research is the analysis of existing data gathered
by researchers, NGOs, governments, and other institutions for different or similar purposes
Cnossen (1997).

Fig 3.1 Study Area in Marondera, Zimbabwe

Source: Google Maps

12
3.2 Research method

Secondary data will be used to collect and synthesize the existing data and a fusion of
quantitative and qualitative research was used. Secondary research is “any further analysis of
existing data which presents interpretations, conclusions or additional knowledge that differs
to the ones presented in the first report on the inquiry as a whole or its main results” Hakim
(1982).

Advantages of Secondary Research

 It is quicker than primary data

 It saves time and its cost efficient
 Allows the researcher to generate new insights from previous research M McCaston
(2005)
 Ease of access because of the wide range of sources

Disadvantages of Secondary Research

 Determining the quality and accuracy of data can be difficult

 The original data collected may not be meant for the purposes of the current research
and therefore can be inappropriate for use and unreliable
 The author might have a bias.

3.3 Research Design

According to C.R Kothari (2004) research design is a schematic plan on which the researcher
chooses the problem to work on, reviews literature, outlines the target population with an
adequate sample size, decides on suitable methods to collect the data and analyze it.

3.3.1 Sample Size and Sampling Techniques

The area covered in this study is the Mushangwe rural farming community which specializes
in growing tomatoes located in Mashonaland East province of Zimbabwe. The area was
chosen because it resembles similar structures, environment, and general typology of rural
farming communities in Zimbabwe. Most of the farmers in the community have been
contracted to Beitbridge Juicing a subsidiary of Schweppes who use the tomato produce for

13
making tomato paste. The sampling frame used by Beitbridge Juicing (BBJ) when collecting
their internal data was random sampling. Random sampling carries less risk and reduces
chances of bias. The population size was 40 including 25 smallholder tomato farmers from
Mushangwe rural communal farming area and 5 governmental and 3 non-governmental
officials and 7 agronomists working for BBJ.

3.3.2 Data Collection Tools

Semi structured online questionnaires, data from NGOs, Government institutions and peer
reviewed journals, and academic papers were used to gather data to satisfy the objectives of
this research. For secondary data, information was obtained mainly from Beitbridge Fruit
Processing Company’s Internal records of output of tomato produce for the past 3 years.
Other sources of secondary data came from the Ministry of Agriculture, Zimbabwe Statistics
(ZimStats), Rural and land resettlement and the provincial data in Mashonaland. These
institutions give access to past data on related areas together with insights in to current and
future policy initiatives.

3.3.3 Semi-Structured Interview (SSI)

Whilst secondary data is less time consuming and cost efficient it has some downsides.
Available information may not be relevant or satisfy all the research questions. The author
adopted SSI to confirm the accuracy of the collected data by asking skilled professionals
(Government and Non-Governmental officials and Agronomists) to ensure the quality and
accuracy of the data.

3.3.4 Data Analysis

A fusion of both Qualitative and quantitative research will be used to gather and analyze the
data although much of it will be quantitative. Because the research will be largely
quantitative descriptive research will be implemented. Descriptive research is appropriate for
this research because it identifies characteristics, trends, and frequencies. Qualitative research
was used to identify the perceptions of Farmers, government, and non-governmental
officials’ view on whether AI is a good solution for Sustainable agriculture. Quantitative
research was used to measure the impact of AI both on employment statistics and agricultural
productivity.

14
3.4 Study Limitations and Constraints

The increase of data and ease of access in the modern day makes secondary research an ideal
and flexible approach to research. However, the disadvantage of this method is that
researchers must settle for the original data collection tools and methods. In this research
many of the respondents were not independent. Farmers in Mushangwe are contracted to BBJ
the company that gathered the data so questions about free will and voluntary participation
come into question. BBJ’s Agronomists also participated, and they also have a conflict of
interest.

The online questionnaire was done to confirm the data from independent sources. The
researcher also compared the data to literature review of similar farming communities in
Zimbabwe and comparing and contrasting with data from ZimStats to establish consistency.
AI in agriculture is also a novice concept in Zimbabwe and the amount of data is fairly
limited especially for rural areas. Data from different regions with similar typology was also
analyzed to increase the validity of the research.

3.5 Ethical Considerations

Ethical procedures for this research were followed. Most of the data used for this research
comes from BBJ company’s Internal data. The author got research clearance by relevant
officials to disclose the internal data for only academic purposes. The purpose of this research
has also been explained to other respondents particularly to farmers who participated
voluntarily with informed consent. Respondents were also assured of their anonymity and
confidentiality, and they will be identified with group names rather than as individuals.

15
CHAPTER 4 DATA ANALYSIS

4.0 Secondary Data Analysis

The purpose of this research was to establish if AI has any impact on productivity in
smallholder rural farmers, the relationship between AI and the employment rate, the impact
AI has on Sustainable Agriculture and farmer’s perceptions on the effectiveness of AI in
solving challenges faced by farmers.

4.1 AI and Productivity

Three data sets from Mushangwe tomato farms, case studies from similar farming
communities in Mashonaland central and results from Command agriculture were used to
answer this question. The table also shows the total output of tomato production, total yield,
and average yield per hectare in tones.

Table 4.1 AI and Productivity in Mushangwe

INVESTMEN
T IN FARM TOTAL AVE TOTAL YIELD
YEAR TECH HA YIELD/HA (Output) VALUE
$ $
2017 35,500.00 42.00 40.17 1,687.00 539,777.00
$
$ 1,456,556.0
2018 88,978.00 64.5 83.00 5,325.00 0
$
$ 2,167,028.0
2019 101,231.00 54.50 124.26 6,772.00 0
$
$ 4,163,361.0
225,709.00 96.50 247.42 13,783,550.00 0
Source: Beitbridge Juicing (BBJ) Internal Data 2020

There is a positive relationship between investment in AI equipment and the level of

productivity. From 2017-2018 the increase in farm tech investment in AI from $35 500 - $88
978 corresponds with an increase in productivity showcased by an increase in the average

16
yield from 40.17tonnes per hectare to 83tonnes per hectare. The trend continues in 2019 with
an increase in investment resulting in an increase in yields per hectare and tomato output.

4.1.2 Mashonaland Central

The graph below shows the impact of an Electrostatic sprayer and automated drip irrigation
on Agricultural productivity of 2 farming communities that are similar to Mushangwe
farming community. They are both located in the Mashonaland region and are contracted to
BBJ.

Fig 4.1 AI and Productivity in Mashonaland Central

TOMATO OUTPUT
70000 63255 60905
60000
50000
41429
Output inKgs

40000
40000
30000
20000
10000
0
2018 2019
YEARS

DELAMORE FARM-ANCHOR NHIMBE FRESH EXPORTS

Source: Beitbridge Juicing (BBJ) Internal Data 2020

Above 2 farms Delamore and Nhimbe farms experienced a significant increase in tomato
output culminating from a new drip irrigation system and a new electrostatic sprayer for
tomatoes. Both farms recorded a 50% increase in output from 2018-2019 due to 2 new AI
farming technology. AI has shown to increase productivity in both farms.

4.1.3 Command Agriculture

17
In Zimbabwe part of the implementation of AI in Agriculture has come in the form of
Command Agriculture. Command Agriculture is an Agro-industrialization concept designed
to boost production and productivity of crops using mechanization and farm technology F
Mazwi (2019).

Table 4. 2: Yield (kg/ha) for Maize under Command Agriculture and Non-
Command Agriculture by Sector
Non -Command Agriculture Command Agriculture

Land Use Production Total Yield Total Yield

Sector (Tonnes) area (Ha) (kgs)/Ha) Production area (Ha) (kgs/Ha)
of Tonnes of
surveyed surveyed
farmers farmers
CL 536 742 722 16 9 1 713
SSCF 106 76 1 382 9 3 3 014
A1 Farms 564 504 1 118 61 35 1 752
ORA 497 589 844 96 48 1 995

The Source: APM Report 2017 ZIMSTAT Ha is hectare. NB. CL is Communal Lands; SSCF is Small
Scale Commercial Farms; ORA is Old Resettlement Areas.

The table illustrates the productivity levels of different farming sectors before and after the
Implementation of smart farming. The following bar graph depicts the impact of AI on one
specific sector, the Small-Scale Commercial Farms. A simple bar graph was used because its
best when comparing sizes of components.

Fig 4.2 AI and Small-Scale Commercial Farms

18
 SSCF
3500
3014
3000

2500
Non-Command Agriculture
2000 Command Agriculture
1500 1382

1000

500

0
Yield/Ha(Kgs)

The Fig 4.2 shows that in the case of SSCF the yield per hectare has nearly tripled due to the
implementation of Command Agriculture. The yield increased from 1382kgs/ha pre
command agriculture to 3014kgs/ha post Command Agriculture. AI has shown to increase
productivity even on a different crop like maize.

Hypothesis testing will aid the analysis of the data from Table 4.1, Fig 1 and Fig 2. The null
hypothesis is represented by H0 whilst the alternate hypothesis is Ha. If there is an increase in
productivity, then the alternate hypothesis Ha. will be accepted and null hypothesis H0 is
rejected.

Reject H0 AI has no impact on agricultural productivity

Accept Ha. AI increases agricultural productivity

All 3 case studies analyzed show a consistent pattern of an increase in AI leading to an

increase in productivity. Hence, we accept the alternate research hypothesis.

4.2 What is the relationship between AI and the employment rate?

Table 4.3 is an extract from BBJ Internal reports for contracted workers working on Tomato
production for a 3-year period.

YEAR 2018 2019 2020

FARM FARM FARM
CROP LABOUR LABOUR LABOUR
Trans Planting TOMAT 110 95 90
19
 O
Fertilisers and TOMAT
Spraying O 30 15 10
TOMAT
Harvesting O 160 155 156
TOMAT
Administration O 8 17 21
TOTAL TOMATO
LABOUR 308 282 277
Source: Beitbridge Juicing (BBJ)October (2020)

A multi bar graph was chosen to represent data because the following data has multiple
categories and focuses on changes in components not totals.

Fig 4.3 below establishes the relationship between AI and Employment rate.

Labor Chat
180
160
140
Number of workers

120
100
80
60
40
20
0
FARM LABOUR FARM LABOUR FARM LABOUR
2018 2019 2020
Years

Trans Planting Fertilisers and Spraying Harvesting Administration

20
Hypothesis testing will aid the analysis of the data from Fig 3. The criterion for the
hypothesis is to accept Ha if there are positive changes in number of workers or reject Ha and
accept Null Hypothesis H0 if there is no change or a negative change in number of workers.
Accept H0 There is no or a negative relationship between AI and employment
Reject Ha There is a positive relationship between AI and employment

The null hypothesis has been accepted because there is a negative relationship between AI
and employment. As AI increases employment has decreased. Labor for transplanting has
seen a significant decrease because of efficient technology to facilitate transplanting.
Fertilizers and spraying have also seen a steep decline from 2018-2019 mainly because of the
booster sprayer which requires less people to calibrate it. Harvesting has largely remained
unchanged because the area has no combined harvester for tomatoes. The number of
administrative employees has shot up by more than 100% because of the new skilled jobs
created by AI.

Fig 4.4 shows a solar powered booster sprayer.

Sources: Forbes 2021

4.3 Sustainable Agriculture
21
Sustainable agriculture is economically viable agricultural production which does not damage
the environment in the long run OECD (2021). Technologies like the Boom sprayer,
Predictive analysis, Agri Robots, and Automated Irrigation Systems have been touted to
improve sustainable agriculture. AI on sustainable agriculture is difficult to quantify,
however, the following table will analyze what 8 farmers, 7 agronomists and 5 governmental
and non-governmental officials who work in and around the area think about the new
technology.

Does AI have an Impact on Sustainable Agriculture?

Table 4.4 Impact of AI on Sustainable Agriculture

%Response of
Perceived Impact of SA Indicators Agreement
Farmers Officials
Automated Irrigation
Systems Efficient water use
Cost Management 95% 100%
Quality of produce

Boom Sprayer
Reduction in water and air pollution 80% 95%
Better Soil Health

Drones Less labor Intensive

More effective analysis of crops 30% 75%

The criterion for acceptance is if more than 50% of the respondents perceive AI to have a
positive impact on Sustainable Agriculture.

Reject H0 AI has no impact on Sustainable Agriculture

Accept Ha AI has a positive impact on Sustainable Agriculture

Alternate hypothesis was accepted because over 50% of respondents perceive AI to be

positive. Data from both farmers, officials and agronomists largely agree to the effectiveness
of AI in reducing environmental damage. However, Agri robots and Drone technologies are
complex in nature and farmers do not have a full understanding of the functions evidenced by
the low percentage of agreements. The 75% of the respondents that agreed are government
officials and Agronomists who are well versed in Agro-technology.

4.4 Is AI useful in combating challenges faced by Smallholder rural farmers?

22
Fig 4 shows the response of farmer’s perception of AI in a pie chart. A pie chart was chosen
because it’s the most suitable for categories and frequencies.

Fig 4.5

Farmers's Perception of AI

20%

Yes
No
Not sure
55%

25%

The criterion for acceptance is if more than 50% of the respondents agree that AI is effective
in solving agricultural challenges

Reject H0 Smallholder farmers believe AI is not effective in solving agricultural challenges

Accept Ha Smallholder farmers believe AI is effective in solving agricultural challenges

More than 50% of the respondents agree to the effectiveness of solving agricultural
challenges therefore we reject the null hypothesis and accept the alternate.

More than half of the small holder farmers in Mushangwe believe that AI is useful in
combating the challenges they face. 25% did not think AI was effective in addressing
contemporary challenges faced by farmers as most felt that the technology was too expensive
to rent out and didn’t solve their problem of access to funding. The remaining 20% was not
convinced of its effectiveness, however they were not writing it off as a possible option.

23
4.5 Results and Discussion

This chapter presents and discusses the results with reference to the objective of the study
which was to analyze the impact of AI on agricultural productivity and employment. The
other 2 sub-aims are to find out famers’ perception of AI and investigate the impact AI has on
sustainable agriculture.
4.5.1 Impact of AI on productivity
Previous research by C Gota and Moyo et al (2016) demonstrated that an increase in AI will
lead to an increase in productivity. The results of this study are similar to those findings as
illustrated in Table 4.1, Fig 1 and Fig 2. All 3 findings show a positive correlation between
implementation of AI and agricultural growth. Whilst correlation does not mean causation
much if not most of the increase in productivity can be directly attributed to the
implementation of AI. Therefore, the results indicate as more investment into AI is made the
productivity increases.
4.5.2 AI and Employment rate
24
Table 4.3 summarized the findings of the impact of AI on employment by comparing the
amount of labor contracted to BBJ working in Mushangwe community for three years. The
multi bar graph Fig 2 illustrates how AI impacts the four stages of agriculture. Wilson
(2017)’s school of thought states that as AI increases more skilled jobs will be created is
venerated by the findings as Administration jobs increased significantly with many new
skilled jobs such as agronomists, accountants and software engineers were created. Other
areas like harvesting, spraying and transplanting saw a significant decrease as more
machinery was introduced. The increase in administrative jobs, however, is insufficient to
cover for the loss of jobs in other parts of the agricultural process hence there is a negative
relationship between Smart farming technology and the employment rate. There is a
discrepancy with previous research from Farkas (2003) and Simoya (2018) who argue that AI
will increase productivity hence more jobs will be created in other areas like operating
increased machinery.

4.5.3 AI and Sustainable Agriculture

Table 4.5 showed the farmers’ perception of AI in relation to sustainable agriculture. Most of
the respondents had positive views of AI particularly the boom spray and automated
irrigation systems. However, for drones only agronomists and officials responded positively
because they are educated and well versed in Agro-technology. The results show that AI does
improve Sustainable Agriculture which is also affirmed by Lindblom et al (2017)’ s Precision
Agriculture concept that views AI as key to Sustainable Agriculture.

4.5.4 Farmers’ perceptions on whether AI is useful in combating agricultural challenges

The responses in Fig 4 show that more than half the respondence believe AI is useful in
combating agricultural challenges. Only 25% do not agree with the notion with the other 20%
not being sure. This outcome is contradictory to the findings of Chiputwa et al who found
that smallholder rural farmers were resilient to adopt new technology. Similar sentiments are
shared by Salokhe and Oida (2003) stating that historic failures of the mechanization
programs have left farmers unwilling to adopt new technologies.

25
Chapter 5
5.0 Conclusion, Recommendations and Limitations
The following section will focus on conclusions of the research, identify limitations and
weaknesses of the study and recommendations for future research will be discussed.
5.1 Conclusion
As Artificial Intelligence increasingly becomes central to many industries it’s important to
understand how it impacts Agriculture especially for Sub-Saharan countries like Zimbabwe
who are heavily reliant on agriculture. This study found that AI does impact the agricultural
productivity positively with a strong correlation between the amount of AI implemented and
the corresponding increase in productivity. The purpose of the study was also to establish the
relationship between AI and employment.
Findings showed that most agricultural processes that required extensive labor experienced
significant decreases in employment after the implementation of AI. However, AI created
new administrative and skill-based jobs that were not in existence previously, but the increase
was not sufficient to cover the loss of employment in other areas. Results showed a negative
relationship between AI and employment. This is contradictory to previous about AI and
26
employment in Zimbabwe by Simoya (2018) but in line with the findings of Soffar (2019)
that AI results in loss of employment. A survey of farmers’ perception of AI as an effective
tool to combat challenges faced showed that almost 60% of farmers had a positive view of AI
with only 25% not agreeing.
There is a paucity of research on the impact of AI on agricultural productivity but only
anecdotal information on the effects of AI in the most vulnerable communities, Sub Saharan
rural smallholder farmers. This study details how employment, productivity, and the
sustainability of agriculture amongst smallholders will be affected by AI. The finding of this
research will be useful to farmers particularly smallholders, Banks and other financial
institution when making lending decisions and government and non-governmental
institutions in creating effective policies and strategies to tackle the implementation of such
technologies.

5.2 Recommendations on Future Research

This study is not exhaustive, and many areas still require further exploration like
 Future research on AI in agriculture should focus on the challenges facing smallholder
rural farmers in adopting to new technology and how prevailing socio-economic
issues affect adoption of AI.
 Furthermore, whilst this research touched on AI and its impact on sustainable
agriculture a more in-depth analysis is needed to gain insight on how AI can play a
role in reducing agricultural greenhouse emissions.
 Many farmers have insufficient knowledge of the benefits of AI on Agriculture. More
effort needs to be put in educating farmers.
 The research also found that displaced labor force had no alternative employment
opportunities and the jobs created required skills that none of the manual labor had.
Basic training of farm workers in machine operation will help ease the transition to
smart farming.

5.3 Limitations of the study

Whilst many attempts to improve the accuracy of the findings were made, they are still
many limitations to the study
 Secondary research was used to gather data and although it has many benefits like
cost effectiveness and time saving determining the quality and accuracy of data was
difficult and most of the data was not relevant to the research

27
  AI is still a novice area of research especially in Zimbabwe with little available data
on said topic. There was difficulty in obtaining reliable data from reputable sources
and the analysis was carried out using a small data set.
 There was an overreliance on one source of data which was internal data derived from
Beitbridge Juicing company. It was the main source of data because they dealt
directly with farmers in Mushangwe our study area of interest. This exposes the
research to bias and narratives of the author.
 The difficulty of measuring the Impact of AI on sustainable agriculture was another
limitation to the study.

REFERENCES

Agarwal, B. 1983 Mechanization in Indian agriculture: an analytical study based on the

Punjab.  pp.290pp.

Aguera P, T Chinembiri, N. Berglund A. Comninos A. Gillwald N. Govan-Vassen (2020)

Paving the way towards digitalizing agriculture in South Africa

A Obi,F.F Chisango 2011 Performance of Smallholder Agriculture Under Limited

Mechanization and the Fast Track Land Reform Program in Zimbabwe International Food
and Agribusiness Management Review Volume 14, Issue 4,

A Simoya (2018) How could Artificial Intelligence create employment opportunities in

Zimbabwe?
BBC News (2020) How AI is helping African farmers
https://www.bbc.com/news/av/business-55138290

28
Bentivoglio, D., Bucci, G., Belletti, M., & Finco, A. (2021). A theoretical framework on
network’s dynamics for precision agriculture technologies adoption. Revista de Economia e
Sociologia Rural, 60(4), e245721
Bughin, J. (2018,). Why AI isn't the death of jobs. MIT Sloan Management Review, 59(4),
42-46.
Cnossen, C. (1997). Secondary research: Learning paper 7. Retrieved May 15, 2017, from
https://www.coursehero.com/file/p39mt1eo/20-No-2-Cnossen-C-1997-Secondary-
ResearchLearning-Paper-7-The-Robert-Gordon/
C Gota (2016) An Assessment of Climate change Adaptation strategies for Chirumanzu
District

Chiputwa, B., Langyintuo, A. S., & Wall, P. (2011, February). Adoption of conservation
agriculture technologies by smallholder farmers in the Shamva District of Zimbabwe: A
Tobit application. In Paper accepted for the 2011 meeting of the Southern Agricultural
Economics Association (SAEA) in Texas, USA.

Daniels, C., Dosso, M., Amadi-Echendu, J., 2018 Eds.; Springer International Publishing:
Cham, Germany pp. 13–32 In Entrepreneurship, Technology Commercialization, and
Innovation Policy in Africa.

Dosso, M.; Nwankwo, C.I.; Travaly, Y (2021) The Readiness of Innovation Systems for the
Fourth Industrial Revolution (4IR) in Sub-Saharan Africa.

Decentralized AI Alliance Artificial Intelligence and Global Challenges—No Poverty by

DAIA DAIA Medium. 2020
https://medium.com/daia/artificial-intelligence-and-global-challenges-a-plan-for-progress-
fecd37cc6bda
Eli-Chukwu, 2019 N.C. Applications of Artificial Intelligence in Agriculture: A Review.
Eng. Technol. Appl. Sci. Res.9, 4377–4383.
FAO 2017 Food and Agriculture Organization of the United Nations Rome, The future of
food and agriculture Trends and challenges

FAO 2016 Food and Agriculture Organization of the United Nations The future of food and
agriculture Alternative pathways to 2050

FAO 2013 Agriculture and the environment: changing pressures, solutions and trade-offs

29
http://www.fao.org/3/y4252e/y4252e12.pdf
Farkas I, laus Gottschalk, László Nagy, (2003) Improved climate control for potato stores by
fuzzy controllers, Computers and Electronics in Agriculture
Food and Agriculture Organization. 2000. Socio-economic Impact of Smallholder Irrigation
Development in Zimbabwe: Case Studies of ten Irrigation Schemes
Feng, Y. J., Liu, Y., & Han, Z. (2011). Land use simulation and landscape assessment by
using genetic algorithm based on cellular automata under different sampling schemes. 22(4),
957–963.
Hakim C 1982 Secondary Analysis and Relationship between Official and Academic
Research

https://journals.sagepub.com/doi/abs/10.1177/0038038582016001005

H Soffar 2019, Artificial Intelligence in Agriculture, Advantages and Disadvantages

Jellason, N.P.; Robinson, E.Z.; Ogbaga, C.(2021) Agriculture 4.0: Is Sub-Saharan Africa
Ready?
J Gupta (2019) Agro-technologies and Prospective Investments.
Kothari, C.R. (2004) Research Methodology: Methods and Techniques. 2nd Edition, New
Age International Publishers, New Delhi.
Lindblom, J., Lundström, C., Ljung, M., & Jonsson, A. (2017). Promoting sustainable
intensification in precision agriculture: review of decision support systems development and
strategies. Precision Agriculture, 18(3), 309-331.
Manyika, J., Chui, M., Miremadi, M., Bughin, J., George, K., Willmott, K., Dewhurst, M.
(2017). Un futuro que funciona: automatización, empleo y productividad (p. 27).
Masinde, M. 2014 Artificial neural networks models for predicting effective drought index
Factoring effects of rainfall variability. Mitig Adapt Strateg Glob Change 19, 1139–1162
https://doi.org/10.1007/s11027-013-9464-0

Mazwi, F., Chemura, A., Mudimu, G. T., & Chambati, W. (2019). Political Economy of
Command Agriculture in Zimbabwe: A State-led Contract Farming Model. Agrarian South:
Journal of Political Economy, 8(1–2), 232–257.
 https://doi.org/10.1177/2277976019856742
McCaston, M. K. (2005). Tips for collecting, reviewing and analysing secondary data.
https://www.ands.org.au/__data/assets/pdf_file/0003/713235/
Melanie Wallendorf; (2001) Literally Literacy, Journal of Consumer Research, Vol 27, Issue
4, Pages 505–511,
30
https://doi.org/10.1086/319625
MELLOR, J.W. (1984), “Accelerated Growth in Agricultural Production and the Production
and the Intersectoral Transfer of Resources”,
Moyo, M.; van Rooyen, A.F.; Chivenge, P.; Bjornlund, H. (2016). Irrigation development in
Zimbabwe: understanding productivity barriers and opportunities at Mkoba and Silalatshani
irrigation schemes.
Nandal J And Rai (1986) “Ergonomics aspects of crop production in tropical developing
countries” International Food Policy Research Institute, DC, USA
Ndebele-Murisa, M. & Mubaya, C., 2015, ‘Climate change: Impact on agriculture, livelihood
options and adaptation strategies for smallholder farmers in Zimbabwe’, in Murisa, T. &
Chikweche, T Zimbabwe’s Prospects for Transformation, pp. 155–198.
Our Story-Alliance for a Green Revolution in Africa (AGRA). (2018)
OECD. (2021b, 2003). Glossary of statistical terms -Sustainable Agriculture
https://stats.oecd.org/glossary/detail.asp?
P Kishore (2020) Economic and Environmental Impact of AI and IOT Based Precision
Agriculture on Small and Mid-Scale Farmers

Panpatte D.G.  (2018) Artificial Intelligence in Agriculture: An Emerging Era of Research

Intutional Science, CANADA (, pp. 1-8)
Pretty J Brett C D Gee, R Hine C Mason J Morison (2001) Policy Challenges and Priorities
for Internalizing the Externalities of Modern Agriculture
Rose, D.C.; Chilvers, J. Agriculture 4.0: (2018) Broadening Responsible Innovation in an Era
of Smart Farming. Front. Sustain. Food Syst.

Rose, D.C.; Wheeler, R.; Winter, M.; Lobley, M.; Chivers, (2021) Agriculture 4.0: Making it
work for people, production, and the planet. Land Use Policy 100, 104933

Russell, Stuart J.; Norvig, Peter (2003), Artificial Intelligence: A Modern Approach  Upper

Saddle River, New Jersey: Prentice Hall, ISBN 0-13-790395-2.
Salokhe A and D. Oida. 2003. Development of appropriate agricultural machinery for
Southeast Asian Countries. Laboratory of Agricultural Systems Engineering China.

S. Liu, 2020 "Artificial Intelligence (AI) in Agriculture" in IT Professional, vol. 22, pp. 14-
15,
Smith, M.J.2020, Getting value from artificial intelligence in agriculture. Anim. Sci. 60, 46.

31
Strusani, D.; Houngbonon, 2019 G.V. The Role of Artificial Intelligence in Supporting
Development in Emerging Markets
Solow, R. 1957. Technical change and the aggregate production function.  Review of
Economics and Statistics 39, 312–20.

Tanha Talaviya, Dhara Shah, Nivedita Patel, Hiteshri Yagnik, Manan Shah, (2020)
Implementation of artificial intelligence in agriculture for optimization of irrigation and
application of pesticides and herbicides, Artificial Intelligence in Agriculture

Whitfield, S.; Dixon, J.L.; Mulenga, B.P.; Ngoma, H. (2015) Conceptualizing farming
systems for agricultural development research: Agric. Syst. 133, 54–62.

Wilson, H. J., Daugherty, P., & Bianzino, N. (2017). The jobs that AI will create. MIT Sloan
Management Review, 58(4), 14

https://search.proquest.com/docview/1916720831?accountid=13567

Wolf K, I Herrera, T Tomich (2017) Long-term agricultural experiments

inform the development of climate-smart agricultural practices
World Bank. World Development Report 2008: Agriculture for Development; The World
Bank: Washington, DC, USA, p. 390.

Y Kinha (2020) Scope and Impact of AI in Agriculture KDnuggets

https://www.kdnuggets.com/2020/07/scope-impact-ai-agriculture.html

Yang H, W. Liusheng, X. Hongli Junmin (2017) Wireless Sensor Networks for Intensive

Irrigated Agriculture, CCNC. 4th IEEE pp. 197-201
Y.J. Kim, R.G. Evans, W.M. Iversen (2008) Remote sensing and control of an irrigation
system using a distributed wireless sensor network 57 (7) pp. 1379-1387
Zeman, K. R., & Rodríguez, L. F. (2019). Quantifying farmer decision-making in an agent-
based model. ASABE Annual International Meeting.
Zimstats (2017) Zimbabwe Smallholder Agricultural Productivity Survey Report

32
Appendices

RESEARCH TOPIC: THE IMPACT OF AI ON AGRICULTURAL PRODUCTIVITY

AND EMPLOYMENT.
Appendix 1
Beitbridge Juicing Company review of Farms under contract
2018 Perfomance review

Total target Actual Actual

Farmers Ha Target yield/ha yield yield yield/ha

Old Citrus farm 8 80,000 640,000 481,250 60,156.25

Dzvova Kizito 20 80,000 1,600,000 837,345 41,867.25

Chisango Charles 2 80,000 160,000 117,115 58,557.50

Nyandoro Stanley 2 80,000 160,000 51,265 25,632.50

Rimaunga 5 80,000 400,000 108,673 21,734.60

TIME MUGUZA 2
80,000 160,000 73,110 36,555.00
TALANA FARM 2
80,000 160,000 16,200 8,100.00

PAS; Robert Mtombeni 8 80,000 640,000 376,009 47,001.13

Tiptrux Laubscher 5.8

33
 80,000 464,000 457,184 78,824.83

Flanagan, Njiri farm 10 80,000 800,000 583,571 58,357.10

Makuti Ranch 10 80,000 800,000 401,134 40,113.40

Mudanga George 3.5 80,000 280,000 330,920 94,548.57

Bonata Mapara 10 80,000 800,000 842,603 84,260.30

Bartgain Investments 8 80,000 640,000 412,596 51,574.50

Nyaruwata 3 80,000 240,000 117,049 39,016.33

Total 99.3 7,944,000 5,206,024

Appendix 2
Analysis of 2 farming communities contracted to Beitbridge Juicing in Mashonaland

2018
Total target
Farmers Ha Target yield/ha yield Actual yield Ave yield
DELAMORE 16.5
FARM-ANCHOR 80,000 1,320,000 683,571 41,428.55
NHIMBE FRESH 30
EXPORTS 80,000 2,400,000 1,200,000 40,000.00
2019
Total target
Farmers Ha Target yield/ha yield Actual yield Ave yield
DELAMORE 15
FARM-ANCHOR 80,000 1,200,000 948,820 63,255
NHIMBE FRESH 15 80,000 60,905
EXPORTS 1,200,000 913,580

Total 30 2,400,000 1,862,400

Appendix 3
Impact of Artificial Intelligence on other crops
Table 7.2: Yield (kg/ha) for Maize under Command Agriculture and Non-Command
Agriculture by Sector

34
 Non -Command Agriculture Command Agriculture

Land Production Total Yield Total Yield

Use (Tonnes) area (kgs)/Ha Production area (kgs/Ha
Sector (Ha) of ) Tonnes (Ha) of )
surveye surveye
d d
farmers farmers
CL 536 742 722 16 9 1 713
SSCF 106 76 1 382 9 3 3 014
A1 Farms 564 504 1 118 61 35 1 752
ORA 497 589 844 96 48 1 995
Ha is hectare. NB. CL is Communal Lands; SSCF is Small Scale Commercial
Farms; ORA is Old Resettlement Areas.

Appendices 4

Respondents were asked the following questions to collect data about the impact of AI on
Sustainable Agriculture.

 Does the Boom Sprayer reduce the amount of fertilizer use?

 Are automated Irrigation Systems efficient in water management?
 Is Predictive Analysis useful in Pest management?
 Do Drones and Agri Robots improve soil management?

35