Hostname: page-component-586b7cd67f-t7czq Total loading time: 0 Render date: 2024-11-27T13:08:31.286Z Has data issue: false hasContentIssue false

Effects of Artificial Intelligence and Robotics on Human Labour: A Systematic Review

Published online by Cambridge University Press:  30 October 2024

Rights & Permissions [Opens in a new window]

Abstract

This systematic literature review paper, written by Channarong Intahchomphoo, Jason Millar, Odd Erik Gundersen, Christian Tschirhart, Kris Meawasige and Hojjat Salemi, examines academic research publications to learn about the effects of artificial intelligence (AI) and robotics on human labour. Papers were collected from three academic databases: Scopus, Web of Science and ABI/INFORM Collection. From 710 papers, 159 papers were included. The article finds that the effects of AI and robotics on human labour can be categorised as: (i) positive effects, (ii) negative effects, and (iii) neutral or unsure effects. The positive effects have five reasons regarding AI and robotics’ potential to: do dangerous work, do tedious work with high efficiency and accuracy, do some aspects of computing work, do work that human labour does not want to do and be used to deal with the labour shortage, and help to reduce business production and maintenance costs. The negative effects are based on two reasons, that AI and robotics will take over human labour in part or entirely, thereby creating unemployment crises, and will not only replace manually repetitive jobs from human labour but also cognitive jobs, causing human labour to fear that their jobs will be replaced by AI and robotics. The neutral and unsure effects are based on various unique arguments. The findings of this review are used to suggest future research for academic communities and practical recommendations to legal professionals and policy makers.

Type
International Perspectives
Copyright
Copyright © The Author(s), 2024. Published by British and Irish Association of Law Librarians

1. INTRODUCTION

The purpose of this systematic literature review is to establish the current state of knowledge in the academically indexed literature on the effects of AI and robotics on human labour. The use of AI and robots in the workplace in various industries is often referred to as the “fourth industrial revolution” and this is an era we are living in.Footnote 1 Increasingly, human labour has to share their workplace or work collaboratively with AI and robotics. Therefore, it is important to conduct this review to better understand the automation and the future of human work. Also, in the past several decades, AI and robotics technologies have rapidly and qualitatively advanced in numerous ways, including ChatGPT, the latest text-based generative AI by OpenAI launched in November 2022. As such, there will be many effects and changes that will directly and indirectly impact human labour. Some jobs may change, and some may vanish. Based on our search of databases, to date on February 2022, there is currently no systematic literature review documenting the effects of AI and robotics on human labour.Footnote 2

The structure of this review consists of the following sections. Section 2 contains the methods of this review. It describes data sources, search strings, selection criteria, and the included and excluded studies. Section 3 presents the results of the included studies and how they answer the review question. Findings are presented based on thematic similarity. Section 4 includes our discussion and suggestions for future research and practical recommendations for legal professionals and policy makers. Section 5 consists of the conclusion and limitation of this review.

2. METHODOLOGY

REVIEW QUESTION

  • What are the effects of artificial intelligence and robotics on human labour?

DATA SOURCES AND SEARCH STRINGS

Three academic databases were used to search for scholarly publications for this review: SCOPUS, Web of Science, and ABI/INFORM Collection. Authors accessed the databases via the University of Ottawa Library, Canada. SCOPUS houses academic collections in many subject areas, whereas the Web of Science contains research publications in engineering and pure and applied sciences. ABI/INFORM Collection is known for having human resource management and business administration publications. The variety of subject content in these repositories is important because we wish to retrieve information from interdisciplinary sources in order to find a variety of viewpoints. Authors performed title and abstract searches in all three selected databases on the same day, (February 1, 2022), with the same search strings (Figure 1). The search expression was designed around three main concepts: artificial intelligence, human labour, and effects. The terms of machine learning and deep learning (a structured machine learning technique) were also included because they are currently heavily used in designing AI algorithms. Abbreviations and synonyms of the three concepts were also included. The Covidence web-based software platform was used to streamline the production of this review including citation screening, abstract and full-text review, and extraction and export of data and references.Footnote 3 This platform was used through the University of Ottawa's institutional subscription.

Figure 1: Search strings designed for this systematic literature review

SELECTION CRITERIA

The criteria for selecting the primary studies to be included in the sample and subsequently analysed included type of publication, date of publication, relevance to the review question, language, subject, and duplicates, as detailed below.

Type of publication: The studies must have been published in peer-reviewed journals or conference proceedings. The content also had to be available in full text format. The criterion of peer-review offers some level of quality control, as does the fact that they are indexed in SCOPUS, Web of Science or ABI/INFORM Collection.

Date of publication: The studies must have been published between January 2011 to February 2022, the time this review was conducted. The date of publication parameters is due to the many new AI and robotics technology developments in the past 10 years.

Relevance to the review question: The studies must be relevant to the review question. Queries in a search engine only retrieve documents that contain the search terms and the presence of a term or even a combination of terms is not a sufficient condition for the relevance of the study to the research question. The relevance of a study must be evaluated by a human reviewer, which is completed by reading the article, starting with abstract screening and then full-text screening.

Language: The studies must have been written in English.

Subject: Studies could come from any academic discipline so there is no constraint in this area. The absence of disciplinary constraints ensures that the search expressions retrieve as many relevant articles as possible (maximum recall). Furthermore, academic discipline should not be restricted as research on this question spans multiple disciplines.

Duplicates: If the searches on SCOPUS, Web of Science and ABI/INFORM Collection produced duplicate results, only one occurrence of the document would be counted.Footnote 4

INCLUDED AND EXCLUDED STUDIES

Figure 2 details the total records identified from each of the three databases searched, records removed due to duplication, wrong type, date or language, articles screened for eligibility and removed, leading to the final total of 159 studies included in this literature review.

Figure 2: Studies identified by the search strings and selection criteria

3. RESULTS

Included studies were clustered according to thematic similarity. The findings indicate that AI and robotics affect human labour in three ways: (i) AI and robotics create positive effects for human labour mainly by supplementing human labour in the workplace, (ii) AI and robotics create negative effects by replacing human labour in the workplace, and (iii) AI and robotics create neutral or unsure effects.

3.1 AI AND ROBOTICS CREATE POSITIVE EFFECTS FOR HUMAN LABOUR

AI and robotics create positive effects for human labour mainly because they supplement human labour in the workplace and reduce workplace risks. This phenomenon creates a new collaborative work between human labour and AI and robotics. The following are the reasons and examples as to how AI and robotics create positive effects for human labour.

AI AND ROBOTICS COULD DO DANGEROUS WORK

In our sample, 24 papers discuss how AI and robotics could do dangerous work. They are used to supplement human labour, because the workplaces are dangerous, hazardous, or in harsh environments and often with high chemical usage, particularly in manufacturing factories.Footnote 5,Footnote 6 Human labour's safety is put in jeopardy.Footnote 7 Examples include iron and steel manufacturing where they use chemicals and high temperatures in the production, but now AI and robotics systems can detect and classify surface defects of the metal. This task was previously performed by human labour exclusively.Footnote 8 Now there are robotic forklifts, hands, arms, and grippers that can lift large or heavy items, and they can handle general goods, waste disposal, and also chemical hazards and toxic substances in factories.Footnote 9,Footnote 10,Footnote 11 Automotive and electronics manufacturing also have robots doing work such as welding, deburring, grinding, printing, and assembling.Footnote 12,Footnote 13,Footnote 14 Those examples demonstrate how, in some cases, AI and robotics liberate human labour by doing dangerous tasks while at the same time helping to improve and protect the health of human labour and by this increasing both safety and efficiency.

There are similar uses of AI and robotics in high risk workplaces including in the oil and gas sector,Footnote 15 in agriculture where spraying pesticide and insect repelling tasks are involved,Footnote 16,Footnote 17 in fish-processing plants where human labour need to be very careful about bacterial contamination spoiling the foods, so using AI and robotics in the cleaning system is much better because they help to minimise contamination,Footnote 18 and in waste management activities where AI image recognition systems are able to classify and monitor garbage, while also reducing human labour needed for dirty recycling and waste disposal tasks.Footnote 19

Furthermore, AI and robotics are used to supplement human labour in dangerous infrastructure maintenance and inspection. Robots are made to clean glass façades and windows of high-rise building instead of human labourFootnote 20,Footnote 21 and are also used to maintain building exterior walls and detect faulty pipelines and ceilings.Footnote 22,Footnote 23 AI and robotics systems can also do work like underwater manipulation tasks when installed in remotely operated underwater vehicles in maritime industry.Footnote 24

Beyond the dangers in workplaces mentioned above, AI and robotics are being used to help humans in rescue missions.Footnote 25 After catastrophic situations like a tsunami, disaster assistance robots are used in the search and rescue process when it is still too risky for human rescuers to do the work. Moreover, robotics drones are used to collect aerial data to plan the rescue place and locate obstacles in the water and on the land.Footnote 26,Footnote 27 During the Covid-19 pandemic, several AI and robotics tools were created to work on tasks in diagnosis, screening, disinfection, surgery, telehealth, patient care, logistics, manufacturing, and others because they could not be infected and spread the virus, unlike the human health workers.Footnote 28

AI AND ROBOTICS COULD DO TEDIOUS WORK WITH HIGH EFFICIENCY AND ACCURACY

In our sample, 33 papers discuss how AI and robotics can do tedious work. They can handle tedious works to supplement the limited physical capacities of human labour. Tedious work includes tiring, time-consuming, labour-intensive, and repetitive tasks.Footnote 29,Footnote 30 Over a certain period, those tasks will be at risk of error-prone situations and can lead to inconsistency of work, and subsequent damages could be costly to a given company. Examples of AI and robotics doing tedious work with high efficiency and accuracy includes: cleaning spaces,Footnote 31 AI algorithms evaluating credit risk,Footnote 32 AI in the document processing automation in the financial and accounting industry,Footnote 33 AI in large scale land-use mapping of the Guangdong Province of China,Footnote 34 AI technology in the fault diagnosis of rotating machinery tested with centrifugal pump and motor bearing (a time-consuming and labour-intensive task if performed by human labour),Footnote 35,Footnote 36,Footnote 37 AI in product quality inspection on a production line i.e., ball bonding inspection in the manufacturing process of semiconductor devices and integrated circuits – a task impossible for human labour to always perform accurately because of eye-fatigue and other health conditionsFootnote 38,Footnote 39 – while AI and robotics are also being implemented in inspection of finished metal surfaces in the lapping and levelling process.Footnote 40 For infrastructure maintenance, AI and robotics could efficiently inspect the condition of sewer pipeline systems, detect potholes on streets and other types of road damage and detect defects in pipeline and concrete structures. Importantly, AI advancement is getting better at doing highly complex inspection and maintenance management tasks.Footnote 41,Footnote 42,Footnote 43,Footnote 44,Footnote 45,Footnote 46

In the academic community, AI is being developed to review research papers submitted to journals and conferences, as the manual review process is time-consuming.Footnote 47 In the agricultural sector, AI with image processing ability is used to classify apples and other fruits and sort them automatically based on their skin colour, determining maturity and ripeness,Footnote 48,Footnote 49 classifying hazelnut varieties,Footnote 50 and monitoring for whiteflies in tomato greenhouses.Footnote 51 In medicine, AI and robotics are made specifically to do tedious work that requires processing a large number of medical records in a short amount of time (i.e., AI classification of patient records with thyroid conditions),Footnote 52 AI is employed in the sending and retrieving of surgical instruments during the patient operation process, tasks normally done by nurses,Footnote 53 and AI is used for screening diabetic retinopathy from retinal images as it takes less time for AI to process the data and millions of people need to be screened for diabetes.Footnote 54 In the retail sector, AI systems are installed on shopping carts to detect and monitor for out-of-stock products and process refill orders.Footnote 55

Beyond the ability of AI and robotics to do tedious works, there is also the ability to perform tasks with higher efficiency and accuracy than human labour in terms of scale, speed, and self-learning. For example, AI analyses the development of corrosion in pipe systems reducing the need for engineers to engage in repetitive tasks,Footnote 56 automotive plants use AI to diagnose technical failures, anticipate maintenance needs, and order replacement parts,Footnote 57 and AI and robotics systems accurately remove rust areas on the surface of a steel material.Footnote 58 In chemical factories, the implementation of AI and robotics is shown to cause fewer errors during the packaging process than human labour.Footnote 59,Footnote 60 In the mining sector, robots could assist in the 3D reconstruction process in geological exploration.Footnote 61

AI AND ROBOTICS COULD DO SOME ASPECTS OF COMPUTING WORK

In our sample, nine papers discuss how AI and robotics can do some aspects of computing work. They can do repeatable computing tasks to decrease the burdens on human labour and they could do more precise work in computing than human labour in general.Footnote 62 For example, AI systems are used in the automatic extraction of data for input in and output from databases,Footnote 63 also, AI can automatically label computerised axial tomography (CAT) scans,Footnote 64 2D and 3D X-ray microtomography and other health imaginal scans which normally requires manual review and label creation for each data item by human health experts.Footnote 65 Generally, AI application developments need many data points to be labelled and used as the training data to build an AI model. These tasks normally require a lot of human labour if the data labelling is done manually, but AI systems are now designed to do automatic data labelling and training and it helps to scale up computing project.Footnote 66,Footnote 67

Moreover, there are several AI applications in social media usage. AI can help in fact-checking and detecting fake news and rumours on social media before they spread widely online and cause individual and societal problems. Due to the enormous volume of data on social media, it is not efficient to only have human labour to conduct fake news detection.Footnote 68,Footnote 69 Another example is that AI can be used in the automatic news or text classification with unstructured content from the web and social media; it is practically impossible for human labour to do text classification for all data on the Internet without help from AI.Footnote 70 Therefore, AI could reduce human burdens in certain computing activities.

AI AND ROBOTICS COULD DO WORK THAT HUMAN LABOUR DOES NOT WANT TO DO AND DEAL WITH THE LABOUR SHORTAGE

In our sample, 13 papers discuss how AI and robotics could do work that human labour does not want to do and at the same time deal with the labour shortage. AI and robotics simply add more labour-equivalent productivity into developed economies that are facing problems with aging populations and declining birthrates,Footnote 71 especially in Japan, where service robots have been developed for use in convenience stores to deal with the labour shortage. These robots perform tasks such as displaying and arranging products on shelves, disposing of expired food items, and interacting with customers.Footnote 72 Additionally, AI and robotics could help solve the human labour shortage problem in the areas far from cities that often have problems with finding staff.Footnote 73

There were examples of how AI and robotics could help to deal with the shortage of human labour in several sectors. In the home care sector, there are cleaning robots for indoor spaces and service robots to grab objects and open a door, Footnote 74,Footnote 75 and robots with hand gestures for elder care, as well as robots that are able to quickly recognise if elders suddenly fall.Footnote 76,Footnote 77

The manufacturing sector also faces similar problems regarding the labour shortage and AI and robotics could be helpful in, for example, ship hull maintenance and inspection services, where robots could do many operational tasksFootnote 78 and AI and robotics can assist in the trimming process of faucet manufacturing, without having human labour manually grasp and load materials into the trimming machine.Footnote 79

In the agricultural sector, AI and robotics are used in the cultivation of apple and pear orchards and other farmlands for plant seeding, pruning, thinning, spraying, weeding, harvesting, mobile navigating and providing security to protect the land from intruders and wild animals. They also help to mitigate the problems related to the shortage of seasonal farm workers.Footnote 80,Footnote 81,Footnote 82 Moreover, AI and robotics can even match skilled workers with companies by using public data to build decision-making models.Footnote 83

AI AND ROBOTICS COULD HELP TO REDUCE PRODUCTION AND MAINTENANCE COSTS

In our sample, 25 papers discuss how AI and robotics help to reduce the business production and maintenance costs, particularly in advanced industries with machinery workload.Footnote 84,Footnote 85 In the farming industry, AI and robotics are used in automatic processes for detection and classification of apples and other fruits and crops for their ripeness and for harvesting and to prune trees by using mobile robots with near-ground sensors to measure the soil's electrical conductivity and soil moisture content, which optimises farming practices and increases crop yields, as well as solving the main problem of increasing human labour costs.Footnote 86,Footnote 87,Footnote 88 Similarly, in the evaluation of citrus plant conditionFootnote 89 and the cane sugar crystallisation process, AI can perform automatic image classification during the processing of cane sugar.Footnote 90 Additionally, AI and robotics could be applied to a variety of agricultural activities including AI with wireless sensors in irrigation projects to monitor water and energy consumption in the farming fields to save on human labour costs, thereby maximising profits.Footnote 91

In academia and medicine, AI can play a role in the process of citation screening for systematic literature reviews to lower the operational costs. This task typically relies on high-cost human labour.Footnote 92 Moreover, AI systems can also help to diagnose cardiovascular disease and disorders of the brain tissues and nervous systems by automatically interpreting magnetic resonance imaging (MRI) scans and electrocardiogram waveforms. Normally, a clinician must deal with hundreds of patients per day who need to be tested. In this case, AI systems will help to save healthcare operational costs and reduce the diagnostic workload.Footnote 93,Footnote 94,Footnote 95 Another aspect is how AI could auto-score sleep stages in humans for sleep studies, if this is done manually it will be costly, as it is labour intensive.Footnote 96

In the construction sector, drones equipped with AI systems can reduce the cost of maintenance of solar panels by automatically detecting defective solar cells while at the same time minimising human labour costs.Footnote 97 Another example is when robotic arms at construction sites are tested to assemble a structure from predefined discrete building blocks and an AI simulation system can be used to train a human controller in the placing of building blocks.Footnote 98 Relying less on highly skilled and expensive human labour to do those jobs is helpful to firms. This also applies to the defect detection of power apparatus based on infrared photos taken and analysed by AI and robots.Footnote 99 Also, in the automobile industry, where AI and robotics are used in automobile repair shops to load and weld steel parts, the cost to buy them and implement them into work tasks are declining, so companies could increase profits by using AI and robotics and having less human labour related costs.Footnote 100,Footnote 101 Moreover, autonomous robots can now construct bridges with minimised human labour, as a solution for businesses to deal with the high human labour costs in bridge construction, it is hoped robots will help to lower the cost of production by replacing human labour.Footnote 102

In waste management, AI is developed to identify recyclable garbage in real-time based on AI automatic classification systems and image recognition technologies; this is also helping to reduce the huge costs of human labour in recycling tasks.Footnote 103,Footnote 104

In the manufacturing sector, AI and robotics are developed to perform in assembly lines to reduce the number and cost associated with human labour.Footnote 105 AI and robotics are experimented with for the purpose of painting ceramic plates, with the aim of reducing production costs and maximising productivity.Footnote 106 In another example, AI and robots are designed to undertake the gluing process for footwear, an area in which this industry normally uses a large volume of skilled and highly-compensated human labour.Footnote 107 For small businesses, AI chatbots with machine learning and natural language processing technologies, are being used more and more to provide customer support, to reduce the human labour costs.Footnote 108

3.2 AI AND ROBOTICS CREATE NEGATIVE EFFECTS FOR HUMAN LABOUR

AI and robotics create negative effects for human labour mainly due to the automation of work, which directly affects the human labour market and can lead to the replacement of human labour in a wide range of occupations. Below are the reasons and examples as to how AI and robotics create negative effects for human labour.

AI AND ROBOTICS WILL TAKE OVER HUMAN LABOUR IN PART OR ENTIRELY AND WILL THEN CREATE AN UNEMPLOYMENT CRISIS

In our sample, 21 papers discuss how AI and robotics will take over from human labour, in part or entirely, and will subsequently create an unemployment crisis.Footnote 109,Footnote 110,Footnote 111,Footnote 112,Footnote 113,Footnote 114 Workers are concerned about potential job displacement due to the automation of jobs by AI and robotics.Footnote 115 This concern is discussed in various occupational areas including in manufacturing,Footnote 116 service tasksFootnote 117 and in the software industry with regards to business process automation.Footnote 118 Also, in the mining industry where AI and robotics will require less human labour in the new mining operations, for instance mine trucks could be driven and controlled from a remote office.Footnote 119

Other examples explore the use of AI and robotics in coconut farms for de-husking and loading coconuts and for monitoring the coconut plants’ health by processing digital images of the plants, thereby replacing farm workers.Footnote 120 In date fruit farming, there will be less and less reliance on human labour expertise for grading and sorting of fruits when AI computer vision technology can perform those tasks.Footnote 121 In the service sector, including restaurants, the amount of human labour in the workforce is being reduced by robots, which are now performing as waiters to take orders and serve food to customers.Footnote 122

Furthermore, the Organisation of Economic Co-operation and Development (OECD) indicated in its report that robots will displace human labour and will also decrease human labour's work compensation over time, particularly in the manufacturing sectors among OECD's member countries.Footnote 123 China is also worried about the same issue.Footnote 124

Beyond the concerns that AI and robotics is likely to entirely take over much of human labour, thus creating an unemployment crisis, there is a worry that AI and robotics are also not good for social equality and could create more income inequality.Footnote 125,Footnote 126,Footnote 127 The effect could be that only management and skilled workers would be protected from the negative effects of AI and robotics as their work is unlikely to be replaced by AI and robots.Footnote 128 Therefore, it has been suggested that governments should introduce a ‘robot tax’ on companies, because AI and robots are believed to be increasingly taking away human labour work.Footnote 129,Footnote 130

AI AND ROBOTICS WILL NOT ONLY REPLACE MANUALLY REPETITIVE JOBS BUT ALSO COGNITIVE JOBS

In our sample, 15 papers discuss how AI and robotics cause problems in the labour market; they change and disrupt business models by not only replacing manually repetitive jobs from human labour but also cognitive and creative jobs, causing a feeling of vulnerability among human labour.Footnote 131,Footnote 132,Footnote 133 Repetitive human labour tasks will be the first to be replaced, but now AI and robotics can do the jobs that were more technically difficult in the past, including medical image processing of magnetic resonance imaging scans of the human body, rail inspection, monitoring crops in apple orchards, counting the number of trees, and going through warehouse aisles to pick up and store items. Those cognitive tasks were normally done by human labour but now AI and robots can do these jobs more accurately than any human labourer or technician.Footnote 134,Footnote 135,Footnote 136,Footnote 137,Footnote 138,Footnote 139

Even the IT sector cannot avoid this negative effect. AI and robotics systems have been replacing human IT service positions.Footnote 140 Another tangible example is that AI could be used to monitor and identify illegal FM radio broadcasting signals.Footnote 141 Soon, there might be less need for human labour when various AI systems become efficient at providing more IT services.

Moreover, several studies have shown that workers would prefer that their position were replaced by other human, not by AI or robots.Footnote 142 Overall, human labour feels fear that jobs may be replaced by AI and robotic automation.Footnote 143

Conversely, other studies emphasise concerns that when robots replace human labour, most of the power and wealth could become even more concentrated in a small group of people and this will deepen the socio-economic gap and inequality.Footnote 144 Additionally, this points to another concern about human labour in low-income countries in the Global South, people who are doing a lot of the data sorting and sifting tasks manually for AI projects because AI systems rely on learning from a large volume of data, and this data needs to be cleaned in the countries that have low human labour costs. This could create another global income inequality problem.Footnote 145

3.3 AI AND ROBOTICS CREATE NEUTRAL AND UNSURE EFFECTS FOR HUMAN LABOUR

In our sample, 28 papers discuss how AI and robotics create neutral and unsure effects for human labour. They could both substitute and complement human labour.Footnote 146,Footnote 147,Footnote 148 Some researchers believe that the total replacement of human labour by AI and robotics is not yet feasible and will not fully happen for a while, as the AI and robotics technologies are not quite there yet.Footnote 149,Footnote 150 They are still currently considered a very new thing in many industries including retail and health care, where AI and robotics are still not yet being used in core operations but they hold a wide range of promising real-world applications.Footnote 151,Footnote 152 In addition, consumers still prefer to have human labour to offer products and services to them, particularly when they are purchasing goods with high symbolic value and uniqueness.Footnote 153 Even with the rapid advancements of AI and robotics in construction, human labour is still needed in complex crafting processes.Footnote 154 Therefore, AI and robotics would not cause an unemployment crisis for human labour. It is very important for human labour to learn to collaborate or work alongside AI and robotics.Footnote 155,Footnote 156,Footnote 157,Footnote 158,Footnote 159,Footnote 160,Footnote 161

The questions now are more about how to find the right balance between the use of AI and robotics and maintaining human labour employment.Footnote 162,Footnote 163 To achieve the right balance, it is essential to position human labour at the forefront, thereby boosting their trust in AI and robotics and change perceptions of these so that they are seen as more of a source of self-improvement and positive employment change.Footnote 164,Footnote 165 In some cases, AI and robotics will even allow human labour to dedicate themselves to doing more higher-level tasks, letting AI and robotics to do the groundwork.Footnote 166

Another topic of discussion is about how AI is better at making predictions than human labour. Though true, humans are still better at making judgements.Footnote 167 The only way for AI and robotics to substitute human labour is if they can show that they can help to reduce business operational costs.

Some researchers go further by arguing that AI and robotics will negatively impact both those who consider themselves as ‘skilled and technical labour’ and the ‘unskilled’.Footnote 168 However, the AI and robotics impact on wage growth might be more significant for low-income occupations compared to middle and high-income technical jobs.Footnote 169

Other neutral and unsure effects of AI and robotics on human labour include: (a) The belief that AI and robotics have both positive and negative effects (complementing or replacing human labour) depending on how they are implemented in a given workplace.Footnote 170 (b) Whether or not it is possible for both human labour and AI and robotics to equally share the tasks. And (c), whether AI and robotics will end up creating more productivity gains and make companies and societies wealthier, and this will be considered as some form of shared prosperity and inclusive economy between AI and robotics and human labour.Footnote 171

Lastly, societies will continue to have a split view on AI and robotics effects on human labour.Footnote 172 We need to keep monitoring the workplace transformation and the increase of AI and robotics automated work environments and how they impact human labour.Footnote 173

4. DISCUSSION AND SUGGESTIONS FOR FUTURE RESEARCH AND PRACTICAL RECOMMENDATIONS FOR LEGAL PROFESSIONALS AND POLICY MAKERS

Ultimately the findings of this review demonstrate that the positive effects of AI and robotics on human labour received more research attention than the negative effects. Compared to 103 studies focusing on positive effects, only 37 studies focused on the negative effects. To achieve more balance, we think that researchers should give more attention to the negative effects of AI and robotics on human labour. Interestingly, there are 28 studies that discussed the neutral and unsure effects of AI and robotics on human labour. It shows that some groups of researchers identify there are many unknowns with regards to emerging technologies like AI and robotics, especially regarding whether there will be full adoption in human societies and, particularly, in the workforce. We also still need more research in the area of neutral and unknown effects of AI and robotics.

Looking deeper at the themes of this review's findings, starting with the positive effects of AI and robotics on human labour, doing dangerous and tedious work and doing the work that human labour does not want to do, are both considered positive and, overall, sensible applications. Performing some aspects of computing work is becoming increasingly prevalent, especially when dealing with big data analysis, data organising, and data cleaning. AI and robotics help to reduce business production and maintenance costs as businesses are always looking for ways to save money. But how does the introduction of AI and robotics look from the perspective of employees who have to do their jobs while seeing the technology being introduced into their workplace? This could be an interesting study.

For the review's findings on the negative effects of AI and robotics, the main concern is about how they can replace human labour in various job categories. The studies that are mentioned discuss how human labour fear that their jobs will be replaced by AI and robotics. We think that this is an interesting research topic, to look at how we might reduce the fear felt by human labour, and to find ways to encourage human labour to see its adoption as an opportunity for them to improve their skills and leverage these technologies to their advantage. The research area on trust and trustworthiness of AI and robotic systems and products is important, because if human end users refuse to adopt the new technologies, even if they are made to the best standards, if they are not being used at all there is almost no point in creating the technology.

Finally, our findings show that there are some concerns about an unemployment crisis caused by AI and robotics. We think that there should be more serious studies with government policy makers and legal professionals to explore whether we should have laws that sets limits on how much businesses and employers could implement AI and robotics into their operations, especially if they could show that there are not a lot of productivity gains in having AI and robotics do the work compared to having human labour doing the same jobs.

Furthermore, what happens to people that were doing garbage collection (as an example) but had a meaning in their life by working and keeping busy. Now, even if they get paid, there is no meaning in their life, we can't expect everyone to become an AI expert or whatever the jobs of the future turn out to be. Another example is the job that a hi-tech executive has that is decision making, once a machine will make better decisions, what happens to she or he? How does meaning come to their life?

Additionally, we think there should be more AI and robotics research with under-served communities, particularly indigenous peoples, to understand how they perceive AI and robotics, and to delve into the intersections between indigenous ways of knowing and being and the utilisation of AI and robotics in a workforce; their ways of life, cultures, languages, and relationships with the land. These findings may not be published in academic databases but rather in conversations and interviews with indigenous peoples living in these communities, including urban centres. Does AI and robotics align with indigenous ways of knowing, or does it work against it? Or is there a middle ground?

These suggested tasks need complex legal and public policy thinking and involvement by people from business, labour unions, governments, academia, and from consumers. We believe such legal and policy studies are important and should be explored.

5. CONCLUSION AND LIMITATION

The future is not set in stone. How AI and robotics will affect society is up to us to decide. We can do this by making policies that reduce the negative effects on society and increase the positive ones. For society to make informed decisions about the effect of AI and robotics, getting a good overview of the findings reported in the literature is important. Our hope is that this literature review will help in the effort of making policies that ensure that society as a whole gains from the introduction of AI and robotics.

In this systematic literature review, we identify the positive, negative, and neutral and unsure effects of AI and robotics on human labour in the existing literature. We also discus key gaps in the literature and give some suggestions for future research and practical recommendations for legal professionals and policy makers. This review has a limitation, as it did not include non-scientific literature such as policy reports, working papers, and government and non-government organisation documents.

References

Endnotes

1 Xu, Min, David, Jeanne M., and Kim, Suk Hi, ‘The Fourth Industrial Revolution: Opportunities and ChallengesInternational Journal of Financial Research 9, no. 2 (2018) 9095CrossRefGoogle Scholar.

2 We started conducting this review in February 2022, before the public launch of the ChatGPT generative AI tool. As a result, no ChatGPT-related studies were included in this review. However, we are currently conducting another systematic review on ChatGPT's ethical issues, including its effects on human labour.

3 Co-author Channarong Intahchomphoo conducted databases search, reviewed studies, and created the codebook for the themes for included studies. Co-authors Jason Millar, Odd Erik Gundersen, Christian Tschirhart, Kris Meawasige, and Hojjat Salemi commented on the overall research methodology, data analysis and discussion.

4 For transparency of this review, the metadata of included and excluded studies from all databases search can be found at https://github.com/ChannarongIntahchomphoo/Effects-of-AI-and-robotics-on-human-labour

5 Siwawong Buthgate and Somyot Kaitwanidvilai, ‘Development of a New Base of Robot Arm Using Finite Element Analysis’ 2018 International Electrical Engineering Congress, (2018) 1–4.

6 Pauliková, Alena, Babeľová, Zdenka Gyurák, and Ubárová, Monika, ‘Analysis of the Impact of Human–Cobot Collabourative Manufacturing Implementation on the Occupational Health and Safety and the Quality RequirementsInternational Journal of Environmental Research and Public Health 18, no. 4 (2021) 1927CrossRefGoogle ScholarPubMed.

7 Mohamad Yuden, Mohamad Adzeem, Md Ghazaly, Mariam, Amran, Aliza Che, Jamaludin, Irma Wani, Yee, Khoo Hui, Yaacob, Mohd Rusdy, Abdullah, Zulkeflee, and Kiat, Yeo Chin, ‘Positioning Control Performances of a Robotic Hand SystemJurnal Teknologi 79, no. 1 (2017) 123Google Scholar.

8 Kostenetskiy, Pavel, Alkapov, Rustem, Vetoshkin, Nikita, Chulkevich, Roman, Napolskikh, Ilya, and Poponin, Ostap, ‘Real-Time System for Automatic Cold Strip Surface Defect DetectionFME Transactions 47, no. 4 (2019) 765774CrossRefGoogle Scholar.

9 Ji, Wei, Wang, Yuquan, Liu, Hongyi, and Wang, Lihui, ‘Interface Architecture Design for Minimum Programming in Human-Robot CollabourationProcedia CIRP 72, (2018) 129134CrossRefGoogle Scholar.

10 Abdul Mutolib, Rina Mardiati, Edi Mulyana, Aan Eko Setiawan, and Ahmad Fathonih, ‘Design of Automatic Goods Carrier Robot System Based on Line Sensor and Fuzzy Logic Control Mamdani’ 6th International Conference on Wireless and Telematics, (2020) 1–4.

11 Achilli, Gabriele Maria, Logozzo, Silvia, Valigi, Maria Cristina, Salvietti, Gionata, Prattichizzo, Domenico, and Malvezzi, Monica, ‘Underactuated Soft Gripper for Helping Humans in Harmful WorksProceedings of I4SDG Workshop 2021: IFToMM for Sustainable Development Goals 108, (2022) 264272CrossRefGoogle Scholar.

12 Veerawuth Punlum, Jiraphon Srisertpol, and Sorada Khaengkam, ‘The Application of Double Arms SCARA Robot for Deburring of PCB Support Plate’ International Conference on Circuits, Devices and Systems, (2017)1–5.

13 Ulewicz, Robert, and Mazur, Magdalena, ‘Economic Aspects of Robotization of Production Processes by Example of a Car Semi-Trailers ManufacturerManufacturing Technology 19, no. 6 (2019) 10541059CrossRefGoogle Scholar.

14 Garcia, Carlos A., Montalvo-Lopez, William, and Garcia, Marcelo V., ‘Human-Robot Collabouration Based on Cyber-Physical Production System and MQTTProcedia Manufacturing 42, (2020) 315321CrossRefGoogle Scholar.

15 Nugraha, F. P., Setiati, R., and Fattahanisa, A., ‘Comparison of Gas Initial in Place Calculation Using MBAL and Artificial Intelligence MethodsJournal of Physics: Conference Series, 1402, no. 5 (2019)Google Scholar.

16 Aletdinova, Anna, Kravchenko, Maxim, and Tsybina, Yana, ‘Prospects of the Social Agricultural Robot CreationAdvances in Computer Science Research 72, (2017) 188192Google Scholar.

17 Ege Ozgul and Ugur Celik, ‘Design and Implementation of Semi-Autonomous Anti-Pesticide Spraying and Insect Repellent Mobile Robot for Agricultural Applications’ 5th International Conference on Electrical and Electronic Engineering, (2018) 233–237.

18 Lars Andre Langøyli Giske, Emil Bjørlykhaug, Trond Løvdal, and Ola Jon Mork, ‘Experimental Study of Effectiveness of Robotic Cleaning for Fish-Processing Plants’ Food Control 100, (2019) 269–277.

19 Wen-Tsai Sung, Ihzany Vilia Devi, S. Hsiao, and Fathria Nurul Fadillah, ‘Smart Garbage Bin Based on AIOT’ Intelligent Automation & Soft Computing 32, no. 3 (2022) 1387–1401.

20 MA Viraj J. Muthugala, Manuel Vega-Heredia, Ayyalusami Vengadesh, G. Sriharsha, and Mohan Rajesh Elara, ‘Design of an Adhesion-Aware Façade Cleaning Robot’ IEEE/RSJ International Conference on Intelligent Robots and Systems, (2019) 1441–1447.

21 Sung Min Moon, Chang Yeob Shin, Jaemyung Huh, Kyeong Won Oh, and Daehie Hong, ‘Window Cleaning System with Water Circulation for Building Façade Maintenance Robot and Its Efficiency Analysis’ International Journal of Precision Engineering and Manufacturing-Green Technology 2, no. 1 (2015) 65–72.

22 Prasanthi Ambati, K. M. Suman Raj, and A. Joshuva, ‘A Review on Pipeline Inspection Robot’ AIP Conference Proceedings 2311, no. 1. (2020) 06002–7.

23 M. A. Viraj J. Muthugala, Koppaka Ganesh Sai Apuroop, Saurav Ghante Anantha Padmanabha, S. M. Bhagya P. Samarakoon, Mohan Rajesh Elara, and Raymond Yeong Wei Wen, ‘Falcon: A False Ceiling Inspection Robot’ Sensors 21, (2021) 5281.

24 Mohd Shahrieel Mohd Aras, Tan Pheng Chaing, Muhamad Khairi Aripin, Mohd Khairi Mohd Zambri, Alias Khamis, Mohd Bazli Bahar, and Mohd Zamzuri Ab Rashid, ‘Monitored and Controlled Underwater Scissor Arm Manipulator Using Pixy Camera’ Indian Journal of Geo Marine Sciences 48, no.7 (2019) 1120–1131.

25 Dibyendu Roy, Madhubanti Maitra, and Samar Bhattacharya, ‘Adaptive Formation-Switching of a Multi-Robot System in an Unknown Occluded Environment Using BAT Algorithm’ International Journal of Intelligent Robotics and Applications 4 (2020) 465–489.

26 Hardefa Rizky Putu Rogonondo, Son Kuswadi, Anhar Risnumawan, and Eny Kusumawati, ‘Tsunami Post Disaster Robot: Simulation and Implementation of Aerial Path Planning to Explore Unknown Environment’ AIP Conference Proceedings 2366, no. 1 (2021) 1–10.

27 Saeed Saeedvand, Hadi S. Aghdasi, and Jacky Baltes, ‘Robust Multi-Objective Multi-Humanoid Robots Task Allocation Based on Novel Hybrid Metaheuristic Algorithm’ Applied Intelligence 49, (2019) 4097–4127.

28 Yang Shen, Dejun Guo, Fei Long, Luis A. Mateos, Houzhu Ding, Zhen Xiu, Randall B. Hellman et al, ‘Robots Under COVID-19 Pandemic: A Comprehensive Survey’ IEEE Access 9, (2021) 1590–1615.

29 Jiří Bláha, Lucjan Klimsza, Aleš Lokaj, and Lech Nierostek, ‘Multidimensional Analysis of Ethical Leadership for Business Development’ European Journal of Sustainable Development 10, no. 1 (2021) 290–290.

30 Supra, note 12, Alena Pauliková, Zdenka Gyurák Babeľová, and Monika Ubárová.

31 Tsung-I Liao, Song-Shyong Chen, Chun-Chieh Lien, Hung-Cheng Lin, Meng-Fang Lu, and Wen-Ping Chen, ‘Development of a High-Endurance Cleaning Robot with Scanning-Based Path Planning and Path Correction’ Microsystem Technologies 27, no. 4 (2021) 1061–1074.

32 Bichen Zheng, ‘Financial Default Payment Predictions Using a Hybrid of Simulated Annealing Heuristics and Extreme Gradient Boosting Machines’ International Journal of Internet Technology and Secured Transactions 9, no. 4 (2019) 404–425.

33 Martin Holeček, ‘Learning from Similarity and Information Extraction from Structured Documents’ International Journal on Document Analysis and Recognition 24, no. 3 (2021) 149–165.

34 Xuan Yang, Zhengchao Chen, Baipeng Li, Dailiang Peng, Pan Chen, and Bing Zhang, ‘A Fast and Precise Method for Large-Scale Land-Use Mapping Based on Deep Learning’ IEEE International Geoscience and Remote Sensing Symposium, (2019) 5913–5916.

35 Jiedi Sun, Changhong Yan, and Jiangtao Wen, ‘Intelligent Bearing Fault Diagnosis Method Combining Compressed Data Acquisition and Deep Learning’ IEEE Transactions on Instrumentation and Measurement 67, no. 1 (2017) 185–195.

36 Yuantao Yang, Huailiang Zheng, Yongbo Li, Minqiang Xu, and Yushu Chen, ‘A Fault Diagnosis Scheme for Rotating Machinery Using Hierarchical Symbolic Analysis and Convolutional Neural Network’ ISA Transactions 91 (2019) 235–252.

37 Hosameldin O.A. Ahmed and Asoke K. Nandi, ‘Connected Components-Based Colour Image Representations of Vibrations for a Two-Stage Fault Diagnosis of Roller Bearings Using Convolutional Neural Networks’ Chinese Journal of Mechanical Engineering 34, (2021) 1–21.

38 Kit Yan Chan, Ka Fai Cedric Yiu, Hak-Keung Lam, and Bert Wei Wong, ‘Ball Bonding Inspections Using a Conjoint Framework with Machine Learning and Human Judgement’ Applied Soft Computing 102, (2021) 107115.

39 Kai Su, Qiangfu Zhao, and Po Chun Lien, ‘Product Surface Defect Detection Based on CNN Ensemble with Rejection’ IEEE 17th International Conference on Dependable, Autonomic and Secure Computing, IEEE 17th International Conference on Pervasive Intelligence and Computing, IEEE 5th International Conference on Cloud and Big Data Computing, IEEE 4th Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech), (2019) 326–331.

40 Junaid Dar, Dinuka Ravimal, ChaBum Lee, and Sun-Kyu Lee, ‘Field Surface Roughness Levelling of the Lapping Metal Surface Using Specular White Light’ The International Journal of Advanced Manufacturing Technology 119, no. 5–6 (2022) 2895–2909.

41 Syed Ibrahim Hassan, L. Minh Dang, Irfan Mehmood, Suhyeon Im, Changho Choi, Jaemo Kang, Young-Soo Park, and Hyeonjoon Moon, ‘Underground Sewer Pipe Condition Assessment Based on Convolutional Neural Networks’ Automation in Construction 106, (2019) 102849.

42 Supra, note 28, Prasanthi Ambati, K. M. Suman Raj, and A. Joshuva.

43 Dirk Meijer, Lisa Scholten, Francois Clemens, and Arno Knobbe, ‘A Defect Classification Methodology for Sewer Image Sets with Convolutional Neural Networks’ Automation in Construction 104, (2019) 281–298.

44 Gioele Ciaparrone, Angela Serra, Vito Covito, Paolo Finelli, Carlo Alberto Scarpato, and Roberto Tagliaferri, ‘A Deep Learning Approach for Road Damage Classification’ Advanced Multimedia and Ubiquitous Engineering, (2018) 655–661.

45 Shushu Fan, Kouichi Takeya, Eiichi Sasaki, ‘Evaluation Method of Defects in Concrete Structures Using Hammer Test by Time-Frequency Analysis and Neural Networks’ IABSE Congress: Structural Engineering for Future Societal Needs, (2021) 955–964.

46 Dharneeshkar J., Soban Dhakshana V., Aniruthan S. A., Karthika R., and Latha Parameswaran, ‘Deep Learning Based Detection of Potholes in Indian Roads Using YOLO’ International Conference on Inventive Computation Technologies, (2020) 381–385.

47 Youfang Leng, Li Yu, and Jie Xiong, ‘DeepReviewer: Collabourative Grammar and Innovation Neural Network for Automatic Paper Review’ International Conference on Multimodal Interaction, (2019) 395–403.

48 Muladi Muladi, Dyah Lestari, and Dedy Tri Prasetyo, ‘Classification of Eligibility Consumption of Manalagi Apple Fruit Varieties Using Backpropagation’ 2019 International Conference on Advanced Mechatronics, Intelligent Manufacture and Industrial Automation, (2019) 75–79.

49 Raymond Erz Saragih, Dessy Gloria, and Albertus Joko Santoso, ‘Classification of Ambarella Fruit Ripeness Based on Color Feature Extraction’ ICIC Express Letters 15, no. 9 (2021) 1013–1020.

50 Alper Taner, Yeşim Benal Öztekin, and Hüseyin Duran, ‘Performance Analysis of Deep Learning CNN Models for Variety Classification in Hazelnut’ Sustainability 13, no. 12 (2021) 6527.

51 Tomáš Tureček, Pavel Vařacha, Alžběta Turečková, Václav Psota, Peter Janků, Vít Štěpánek, Adam Viktorin et al. ‘Scouting of Whiteflies in Tomato Greenhouse Environment Using Deep Learning’ Agriculture Digitalization and Organic Production: Proceedings of the First International Conference, (2021) 323–335.

52 Dongyup Shin, Hye Jin Kam, Min-Seok Jeon, and Ha Young Kim, ‘Automatic Classification of Thyroid Findings Using Static and Contextualized Ensemble Natural Language Processing Systems: Development Study’ JMIR Medical Informatics 9, no. 9 (2021) e30223.

53 Fica Aida Nadhifatul Aini, Ahmad Zatnika Purwalaksana, and Istas Pratomo Manalu, ‘Object Detection of Surgical Instruments for Assistant Robot Surgeon Using KNN’ 2019 International Conference on Advanced Mechatronics, Intelligent Manufacture and Industrial Automation, (2019) 37–40.

54 Gahyung Ryu, Kyungmin Lee, Donggeun Park, Sang Hyun Park, and Min Sagong, ‘A Deep Learning Model for Identifying Diabetic Retinopathy Using Optical Coherence Tomography Angiography’ Scientific Reports 11, no. 1 (2021) 23024.

55 Dario Allegra, Mattia Litrico, Maria Ausilia Napoli Spatafora, Filippo Stanco, and Giovanni Maria Farinella, ‘Exploiting Egocentric Vision on Shopping Cart for Out-Of-Stock Detection in Retail Environments’ Proceedings of the IEEE/CVF International Conference on Computer Vision, (2021) 1735–1740.

56 Andika Rachman and RM Chandima Ratnayake, ‘Corrosion Loop Development of Oil and Gas Piping System Based on Machine Learning and Group Technology Method’ Journal of Quality in Maintenance Engineering 26, no. 3 (2019) 349–368.

57 Eduardo Plastino and Mark Purdy, ‘Game Changing Value from Artificial Intelligence: Eight Strategies’ Strategy & Leadership 46, no.1 (2018)16–22.

58 Yang Tian, Guoteng Zhang, Jie Ma, and Shugen Ma, ‘Automated Rust Detection via Digital Image Recognition during Grinding Work Process’ 2018 IEEE International Conference on Information and Automation, (2018) 318–323.

59 Mohamed T. Eraky, Dmitry V. Zubov, and Konstantin S. Krysanov, ‘Investigation of Inverse Kinematics Software Program of KUKA Manipulator Robot and Creation of Optimal Trajectory Control for Quality Evaluation within Chemical Production Lines’ International Journal of Engineering Research and Technology 11, no. 12 (2018) 2135–2158.

60 Mohamed T. Eraky, Dmitry V. Zubov, and Konstantin S. Krysanov, ‘Investigation of Forward Kinematics Software Program and Control of 3-DOF Manipulator Robot Using a New Developed Geometrical Approach Method for Improvement of Quality Food and Chemical Industries’ International Journal of Engineering Research and Technology 12, no. 9 (2019) 1381–1388.

61 Yang Lei, Li En, and Liang Zize, ‘An Automated Robot-Assisted 3D Reconstruction System for Adit’ 2017 Chinese Automation Congress, (2017) 1767–1772.

62 Shivali Agarwal, Vishalaksh Aggarwal, Arjun R. Akula, Gargi Banerjee Dasgupta, and Giriprasad Sridhara, ‘Automatic Problem Extraction and Analysis from Unstructured Text in IT Tickets’ IBM Journal of Research and Development 61, no. 1 (2017) 4–41.

63 Nicolas Hengartner, Leticia Cuellar, Xiao-Cheng Wu, Georgia Tourassi, John Qiu, Blair Christian, and Tanmoy Bhattacharya, ‘CAT: Computer Aided Triage Improving upon the Bayes Risk through ε-Refusal Triage Rules’ BMC Bioinformatics 19, no. 18 (2018) 3–8.

64 Matias Benitez, James Tian, Mark Kelly, Vignesh Selvakumaran, Matthew Phelan, Maciej Mazurowski, Joseph Y. Lo, Geoffrey D. Rubin, and Ricardo Henao, ‘Combining Deep Learning Methods and Human Knowledge to Identify Abnormalities in Computed Tomography (CT) Reports’ Medical Imaging 2019: Computer-Aided Diagnosis 10950, (2019) 229–235.

65 Reed Kopp, Joshua Joseph, and Brian L. Wardle, ‘Automated Segmentation of In Situ X-ray Microtomography of Progressive Damage in Advanced Composites via Deep Learning’ AIAA Scitech 2021 Forum, (2021) 1–23.

66 Bharat Prakash, Mohit Khatwani, Nicholas Waytowich, and Tinoosh Mohsenin, ‘Improving Safety in Reinforcement Learning Using Model-Based Architectures and Human Intervention’ Association for the Advancement of Artificial Intelligence, (2019).

67 Lionel J. Wong, Anthony Vassiliou, and Rodney Stromberg, ‘AI-Complemented First-Break Picking for Field Low-S/N Seismic Data’ The Leading Edge 40, no.6 (2021) 460–463.

68 Amir Pouran Ben Veyseh, My T. Thai, Thien Huu Nguyen, and Dejing Dou, ‘Rumor Detection in Social Networks via Deep Contextual Modeling’ Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, (2019) 113–120.

69 Albert Sawczyn, Jakub Binkowski, Denis Janiak, Łukasz Augustyniak, and Tomasz Kajdanowicz, ‘Fact-Checking: Relevance Assessment of References in the Polish Political Domain’ Procedia Computer Science 192, (2021) 1285–1293.

70 Adrita Barua, Omar Sharif, and Mohammed Moshiul Hoque, ‘Multi-Class Sports News Categorization Using Machine Learning Techniques: Resource Creation and Evaluation’ Procedia Computer Science 193, (2021) 112–121.

71 Harumo Sasatake, Ryosuke Tasaki, Takahito Yamashita, and Naoki Uchiyama, ‘Imitation Learning System Design with Small Training Data for Flexible Tool Manipulation’ International Journal of Automation Technology 15, no. 5 (2021) 669–677.

72 Mayuka Shii, and Kenichi Ohara, ‘Stable Manipulation with Visual Servo Control for the Disposal of Objects in Retail Stores’ 2020 IEEE/SICE International Symposium on System Integration Proceedings, (2020) 972–975.

73 Supra, note 19, Robert Ulewicz, and Magdalena Mazur.

74 Supra, note 37, Tsung-I Liao, Song-Shyong Chen, Chun-Chieh Lien, Hung-Cheng Lin, Meng-Fang Lu, and Wen-Ping Chen.

75 Yuya Nakanishi and Yuki Inoue, ‘A Study on Active Wheel Level Control of Robots with Steering Moving Omnidirectionally’ 45th Annual Conference of the IEEE Industrial Electronics Society 1, (2019) 6821–6826.

76 Akitoshi Sato, Joo Kooi Tan, and Yuta Ono, ‘Development of a Human-Robot Cooperative System Based on Visual Information’ International Workshop on Advanced Image Technology 11049, (2019) 793–798.

77 Roy Chaoming Hsu, Po-Cheng Su, Jia-Le Hsu, and Chi-Yong Wang, ‘Real-Time Interaction System of Human-Robot with Hand Gestures’ IEEE Eurasia Conference on IOT, Communication and Engineering, (2020) 396–398.

78 M.A. Viraj J. Muthugala, S.M. Bhagya P. Samarakoon, and Mohan Rajesh Elara, ‘Toward Energy-Efficient Online Complete Coverage Path Planning of a Ship Hull Maintenance Robot Based on Glasius Bio-Inspired Neural Network’ Expert Systems with Applications 187, (2022) 1–9.

79 S. Jantanalach and S. Vongbunyong, ‘Gripper Design for Automatic Trimming System in Forging Process’ IOP Conference Series: Materials Science and Engineering 639, no. 1 (2019) 012013.

80 Novian Habibie, Aditya Murda Nugraha, Ahmad Zaki Anshori, M. Anwar Ma'sum, and Wisnu Jatmiko, ‘Fruit Mapping Mobile Robot on Simulated Agricultural Area in Gazebo Simulator Using Simultaneous Localization and Mapping’ 2017 International Symposium on Micro-NanoMechatronics and Human Science, (2017) 1–7.

81 Nithin, V., Shivam Mishra, P. Devarubiny, and S. Muthulakshmi, ‘IoT Enabled Farming Assist and Security Using Machine Learning’ ARPN Journal of Engineering and Applied Sciences 14, No.9 (2019) 1809–1819.

82 Yaqoob Majeed, Jing Zhang, Xin Zhang, Longsheng Fu, Manoj Karkee, Qin Zhang, and Matthew D. Whiting, ‘Deep Learning Based Segmentation for Automated Training of Apple Trees on Trellis Wires’ Computers and Electronics in Agriculture 170, (2020) 105277.

83 Harri Ketamo, Anu Passi-Rauste, Peter Vesterbacka, and Sanna Vahtivuori-Hänninen, ‘Accelerating the Nation: Applying AI to Scout Individual and Organisational Human Capital’ ICIE 6th International Conference on Innovation and Entrepreneurship, (2018) 198–205.

84 Zixiang Li, Nilanjan Dey, Amira S. Ashour, and Qiuhua Tang, ‘Discrete Cuckoo Search Algorithms for Two-Sided Robotic Assembly Line Balancing Problem’ Neural Computing and Applications 30, no. 9 (2018) 2685–2696.

85 Palash Dutta and Gourangajit Borah, ‘Robot Selection Problem via Fuzzy TOPSIS Method Using Novel Distance and Similarity Measure for Generalized Fuzzy Numbers with Unequal Heights’ New Mathematics and Natural Computation 18, no. 03 (2022) 657–702.

86 Bingjie Xiao, Minh Nguyen, and Wei Qi Yan, ‘Apple Ripeness Identification Using Deep Learning’ International Symposium on Geometry and Vision, (2021) 53–67.

87 Merrick Campbell, Keran Ye, Elia Scudiero, and Konstantinos Karydis, ‘A Portable Agricultural Robot for Continuous Apparent Soil Electrical Conductivity Measurements to Improve Irrigation Practices’ 17th International Conference on Automation Science and Engineering (2021) 2228–2234.

88 Rafaël Verbiest, Kris Ruysen, Tanja Vanwalleghem, Eric Demeester, and Karel Kellens, ‘Automation and Robotics in the Cultivation of Pome Fruit: Where Do We Stand Today?.’ Journal of Field Robotics 38, no. 4 (2021) 513–531.

89 Dat Do, Frank Pham, Amar Raheja, and Subodh Bhandari, ‘Machine Learning Techniques for the Assessment of Citrus Plant Health using UAV-Based Digital Images’ Proceedings of SPIE, (2018) 189–200.

90 Jinlai Zhang, Yanmei Meng, Jianfan Wu, Johnny Qin, Tao Yao, and Shuangshuang Yu, ‘Monitoring Sugar Crystallization with Deep Neural Networks’ Journal of Food Engineering 280 (2020) 109965.

91 Khongdet Phasinam and Thanwamas Kassanuk, ‘Supervised Machine Learning in Precision Agriculture,’ International Journal of Mechanical Engineering 7, no. 1 (2022) 1621–1625.

92 Anjani Dhrangadhariya, Roger Hilfiker, Roger Schaer, and Henning Müller, ‘Machine Learning Assisted Citation Screening for Systematic Reviews’ Studies in Health Technology and Informatics 270, (2020) 302–306.

93 Riddhish Bhalodia, Anupama Goparaju, Tim Sodergren, Alan Morris, Evgueni Kholmovski, Nassir Marrouche, Joshua Cates, Ross Whitaker, and Shireen Elhabian, ‘Deep Learning for End-to-End Atrial Fibrillation Recurrence Estimation’ 2018 Computing in Cardiology Conference 45, (2018) 1–4.

94 Yuki Shimada, Tetsuya Tanimoto, Masataka Nishimori, Antoine Choppin, Arie Meir, Akihiko Ozaki, Asaka Higuchi, Makoto Kosaka, Yuki Shimahara, and Naoyuki Kitamura, ‘Incidental Cerebral Aneurysms Detected by a Computer-Assisted Detection System Based on Artificial Intelligence: A Case Series’ Medicine 99, no. 43 (2020).

95 Ahmad Mohsin and Oliver Faust, ‘Automated Characterization of Cardiovascular Diseases Using Wavelet Transform Features Extracted from ECG Signals’ Journal of Mechanics in Medicine and Biology 19, no. 01 (2019) 1940009–19.

96 Linda Zhang, Daniel Fabbri, Raghu Upender, and David Kent, ‘Automated Sleep Stage Scoring of the Sleep Heart Health Study Using Deep Neural Networks’ SLEEPJ 42, no. 11 (2019) 1–10.

97 Kyungmin Kim, Yoojin Choi, Keonyoung Shim, Hyewon Jeon, Jane Commerford, and Eric T. Matson, ‘Analyzing the Range of Angles of a Solar Panel to Detect Defective Cells, Using a UAV’ Third IEEE International Conference on Robotic Computing, (2019) 471–476.

98 Boris Belousov, Bastian Wibranek, Jan Schneider, Tim Schneider, Georgia Chalvatzaki, Jan Peters, and Oliver Tessmann, ‘Robotic Architectural Assembly with Tactile Skills: Simulation and Optimization’ Automation in Construction 133, (2022) 104006.

99 Chenxi Li, Jun Wang, Yuyan Li, Yufeng Huang, Bo Li, and Zhong Zheng, ‘Infrared Image Defect Diagnosis Through LAB Space Transformation’ 2nd International Conference on Information Systems and Computer Aided Education, (2019) 126–130.

100 Alain Patchong and Laoucine Kerbache, ‘Transiting Toward the Factory of the Future: Optimal Buffer Sizes and Robot Cell Design in Car Body Production’ 2017 IEEE International Conference on Industrial Engineering and Engineering Management, (2017) 1596–1601.

101 Supra, note 19, Robert Ulewicz, and Magdalena Mazur.

102 Paul Gauvreau, ‘Robotics in Construction and the New Era of Efficient Concrete Bridges’ IABSE Congress: The Evolving Metropolis, (2019) 182–186.

103 Qianqian Luo, Guohua Yang, and Xiaofeng Zhao, ‘An Edge-Cloud Framework Equipped with Deep Learning Model for Recyclable Garbage Detection’ Eighth International Conference on Advanced Cloud and Big Data, (2020) 248–252.

104 Qianqian Luo, Zhenzhou Lin, Guohua Yang, and Xiaofeng Zhao, ‘DEC: A Deep-Learning Based Edge-Cloud Orchestrated System for Recyclable Garbage Detection’ Concurrency and Computation: Practice and Experience 35, no. 13 (2023) e6661.

105 J. Mukund Nilakantan, Izabela Nielsen, S. G. Ponnambalam, and S. Venkataramanaiah, ‘Differential Evolution Algorithm for Solving RALB Problem Using Cost-and Time-Based Models’ The International Journal of Advanced Manufacturing Technology 89, no. 1 (2017) 311–332.

106 Nuntiya Kruethi, Seksan Chaijit, Wachirapond Permpoonsinsup, Dechathanat Thongkot, A. H. Ismail, and Pramot Srinoi, ‘A Simulation Model to Improve the Efficiency of Painting Robots and Applied an Engineering Economic for Project Selection’ 17th International Conference on ICT and Knowledge Engineering, (2019) 1–6.

107 Kevin Castelli, Ahmed Magdy Ahmed Zaki, Yevheniy Dmytriyev, Marco Carnevale, and Hermes Giberti, ‘A Feasibility Study of a Robotic Approach for the Gluing Process in the Footwear Industry’ Robotics 10, no. 1 (2020) 6.

108 Rupesh Singh, Manmath Paste, Nirmala Shinde, Harshkumar Patel, and Nitin Mishra, ‘Chatbot Using TensorFlow for Small Businesses’ 2018 Second International Conference on Inventive Communication and Computational Technologies, (2018) 1614–1619.

109 Juha Saukkonen, Pia Kreus, Nora Obermayer, Óscar Rodríguez Ruiz, and Maija Haaranen, ‘AI, RPA, ML and Other Emerging Technologies: Anticipating Adoption in the HRM Field’ ECIAIR 2019 European Conference on the Impact of Artificial Intelligence and Robotics, (2019) 287–296.

110 Zoltán Rajnai and István Kocsis, ‘Labour Market Risks of Industry 4.0, Digitization, Robots and AI’ 2017 IEEE 15th International Symposium on Intelligent Systems and Informatics, (2017) 343–346.

111 Roland W. Scholz, Eric J. Bartelsman, Sarah Diefenbach, Lude Franke, Arnim Grunwald, Dirk Helbing, Richard Hill et al, ‘Unintended Side Effects of the Digital Transition: European Scientists’ Messages from a Proposition-Based Expert Round Table’ Sustainability 10, no. 6 (2018): 2001.

112 Supra, note 90, Zixiang Li, Nilanjan Dey, Amira S. Ashour, and Qiuhua Tang.

113 Chia-Hui Lu, ‘The Impact of Artificial Intelligence on Economic Growth and Welfare’ Journal of Macroeconomics 69, (2021) 103342.

114 Debraj Ray and Dilip Mookherjee, ‘Growth, Automation, and the Long-Run Share of Labour’ Review of Economic Dynamics 46 (2022) 1–26.

115 Stanislav Ivanov, Mihail Kuyumdzhiev, and Craig Webster, ‘Automation Fears: Drivers and Solutions’ Technology in Society 63, (2020) 101431.

116 Ting-Sheng Chen, and Jen-Yuan Chang, ‘Robot Manipulator Vibration Resistance Control Based on Physical Model’ Information Storage and Processing Systems 83600, (2020) V001T03A002.

117 Ming-Hui Huang, and Roland T. Rust, ‘Artificial Intelligence in Service’ Journal of Service Research 21, no. 2 (2018) 155–172.

118 Andrzej Sobczak, ‘Building a Robotic Capability Map of the Enterprise’ Problemy Zarządzania – Management Issue 17, no. 5(85) (2019) 132–153.

119 Dusan Paredes and David Fleming-Muñoz, ‘Automation and Robotics in Mining: Jobs, Income and Inequality implications’ The Extractive Industries and Society 8, no. 1 (2021) 189–193.

120 Anu B. Titus, Thejas Narayanan, and Gautham P. Das, ‘Vision System for Coconut Farm Cable Robot’ 2017 IEEE International Conference on Smart Technologies and Management for Computing, Communication, Controls, Energy and Materials, (2017) 443–450.

121 Haya Alaskar, Shaikah Alhewaidi, Bayan Obaid, Ghadah Alzahrani, Aisha Abdulahi, Zohra Sbai, and Thavavel Vaiyapuri, ‘Dates Fruit Classification Using Convolution Neural Networks’ Proceedings of Sixth International Congress on Information and Communication Technology 3, (2022) 757–775.

122 Tajim Md. Niamat Ullah Akhund, Md. Abu Bakkar Siddik, Md. Rakib Hossain, Md. Mazedur Rahman, Nishat Tasnim Newaz, and Mohd. Saifuzzaman. ‘IoT Waiter Bot: A Low Cost IoT Based Multi Functioned Robot for Restaurants’ in IEEE 8th International Conference on Reliability, Infocom Technologies and Optimization: Trends and Future Directions (2020) 1174–1178.

123 Fabiano Compagnucci, Andrea Gentili, Enzo Valentini, and Mauro Gallegati, ‘Robotization and Labour Dislocation in the Manufacturing Sectors of OECD Countries: A Panel VAR Approach’ Applied Economic 51, no. 57 (2019) 6127–6138.

124 Yi Xu and Xin Ye, ‘Technology Upgrading and Labour Degrading? A Sociological Study of Three Robotized Factories’ The Journal of Chinese Sociology 8, no. 1 (2021) 1–23.

125 Spyros Makridakis, ‘High Tech Advances in Artificial Intelligence (AI) and Intelligence Augmentation (IA) and Cyprus’ Cyprus Review 30, no. 2 (2018) 159–167.

126 Cynthia Estlund, ‘What Should We Do After Work? Automation and Employment Law’ The Yale Law Journal 128, no. 2 (2018) 254–326.

127 ÖndeR Nomaler and Bart Verspagen, ‘Perpetual Growth, the Labour Share, and Robots’ Economics of Innovation and New Technology 29, no. 5 (2020) 540–558.

128 Andrew Berg, Edward F. Buffie, and Luis-Felipe Zanna, ‘Should We Fear the Robot Revolution? (The Correct Answer is Yes)’ Journal of Monetary Economics 97, (2018) 117–148.

129 Oliver Bendel, ‘Are Robot Tax, Basic Income or Basic Property Solutions to the Social Problems of Automation?’ AAAI 2019 Spring Symposium: Interpretable AI for Well-being, (2019).

130 Supra, note 133, ÖndeR Nomaler and Bart Verspagen.

131 Alex Liebergesell, ‘Design in the Age of Autonomous Machines: Modeling Inclusion, Dialogue, and Behavior’ The International Journal of Technology, Knowledge, and Society 15, no. 1 (2019) 27–37.

132 Henrique Barros Lopes, Flávio Vinícius Cruzeiro Martins, Rodrigo T. N. Cardoso, and Vinícius Fernandes dos Santos, ‘Combining Rules and Proportions: A Multiobjective Approach to Algorithmic Composition’ 2017 IEEE Congress on Evolutionary Computation, (2017) 2282–2289.

133 Josep Lladós-Masllorens, ‘Surfing the Waves of Digital Automation in Spanish Labour Market’ The International Research & Innovation Forum, (2019) 451–458.

134 Jhonata E. Ramos, Hae Yong Kim, and F. B. Tancredi, ‘Automation of the ACR MRI Low-Contrast Resolution Test Using Machine Learning’ 11th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, (2018).

135 Yunus Santur, Mehmet Karaköse, and Erhan Akin, ‘Big Data Framework for Rail Inspection’ 2017 International Artificial Intelligence and Data Processing Symposium, (2017) 1–4.

136 Pinmo Tong, Pengcheng Han, Suicheng Li, Ni Li, Shuhui Bu, Qing Li, and Ke Li, ‘Counting Trees with Point-Wise Supervised Segmentation Network’ Engineering Applications of Artificial Intelligence 100 (2021) 104172.

137 Pinmo Tong, Xishan Zhang, Pengcheng Han, and Shuhui Bu, ‘Point In: Counting Trees with Weakly Supervised Segmentation Network’ 2020 25th International Conference on Pattern Recognition, (2021) 9546–9552.

138 Hanan M. Hameed, Abdulmuttalib Turky Rashid, and Khairia A. Al Amry, ‘Automatic Storage and Retrieval System Using a Single Mobile Robot’ Proceedings of the 2nd International Conference on Electrical, Communication, and Computer Engineering, (2020) 1–6.

139 Miao Zhao, Yankun Peng, Long Li, and Xin Qiao, ‘Detection and Classification Manipulator System for Apple Based on Machine Vision and Optical Technology’ ASABE 2020 Annual International Virtual Meeting, (2020) 1–8.

140 Supra, note 68, Shivali Agarwal, Vishalaksh Aggarwal, Arjun R. Akula, Gargi Banerjee Dasgupta, and Giriprasad Sridhara.

141 M. Huang, D. R. Yang, D. Zhu, M. X. Yang, and J. J. Yang, ‘FM Broadcast Monitoring Using Artificial Intelligence’ Radio Science 55, no. 4 (2020) 1–6.

142 Armin Granulo, Christoph Fuchs, and Stefano Puntoni, ‘Psychological Reactions to Human Versus Robotic Job Replacement’ Nature Human Behaviour 3, no. 10 (2019) 1062–1069.

143 Supra, note 121, Stanislav Ivanov, Mihail Kuyumdzhiev, and Craig Webster.

144 Mika Westerlund, ‘The Ethical Dimensions of Public Opinion on Smart Robots’ Technology Innovation Management Review 10, no. 2 (2020) 25–36.

145 Benedetta Catanzariti, Srravya Chandhiramowuli, Suha Mohamed, Sarayu Natarajan, Shantanu Prabhat, Noopur Raval, Alex S. Taylor, and Ding Wang, ‘The Global Labours of AI and Data Intensive Systems’ Companion Publication of the 2021 Conference on Computer Supported Cooperative Work and Social Computing, (2021) 319–322.

146 Frank Fossen and Alina Sorgner, ‘Mapping the Future of Occupations: Transformative and Destructive Effects of New Digital Technologies on Jobs’ Foresight and STI Governance 13, no. 2 (2019) 10–18.

147 Dawei Zhang, Gang Peng, and Yuliang Yao, ‘Artificial Intelligence or Intelligence Augmentation? Unravelling the Debate through an Industry-level analysis’ Pacific Asia Conference on Information Systems Proceedings 68, (2019).

148 Supra, note 119, Chia-Hui Lu.

149 Supra, note 131, Spyros Makridakis.

150 João Reis, Nuno Melão, Juliana Salvadorinho, Bárbara Soares, and Ana Rosete, ‘Service Robots in the Hospitality Industry: The Case of Henn-na Hotel, Japan’ Technology in Society 63, (2020) 101423.

151 Felix Dominik Weber and Reinhard Schütte, ‘State-of-the-Art and Adoption of Artificial Intelligence in Retailing’ Digital Policy, Regulation and Governance 21, no. 3 (2019) 264–279.

152 Andrea Paesano, ‘Artificial Intelligence and Creative Activities Inside Organizational Behavior’ International Journal of Organizational Analysis 31, no.5 (2021) 1694–1723.

153 Armin Granulo, Christoph Fuchs, and Stefano Puntoni, ‘Preference for Human (vs. Robotic) Labour is Stronger in Symbolic Consumption Contexts’ Journal of Consumer Psychology 31, no. 1 (2021) 72–80.

154 Soomeen Hahm, ‘Augmented Craftsmanship: Creating Architectural Design and Construction Workflow by Augmenting Human Designers and Builders’ ACADIA, (2019) 448–456.

155 Melanie Swan, ‘Philosophy of Social Robotics: Abundance Economics’ International Conference on Social Robotics, (2016) 900–908.

156 Zedong Hu, Chenguang Yang, Wei He, Zhijun Li, and Shunzhan He, ‘Modeling and Simulation of Hand Based on OpenSim and Leap Motion’ 2017 Chinese Automation Congress, (2017) 4844–4849.

157 Charlynne Bolton, Veronika Machová, Maria Kovacova, and Katarina Valaskova, ‘The Power of Human-Machine Collabouration: Artificial Intelligence, Business Automation, and the Smart Economy’ Economics, Management, and Financial Markets 13, no. 4 (2018) 51–57.

158 Weitian Wang, Rui Li, Yi Chen, Z. Max Diekel, and Yunyi Jia, ‘Facilitating Human–Robot Collabourative Tasks by Teaching-Learning-Collabouration from Human Demonstrations’ IEEE Transactions on Automation Science and Engineering 16, no. 2 (2018) 640–653.

159 Alina Tausch and Annette Kluge, ‘The Best Task Allocation Process is to Decide on One's Own: Effects of the Allocation Agent in Human–Robot Interaction on Perceived Work Characteristics and Satisfaction’ Cognition, Technology & Work 24, no. 1 (2022) 39–55.

160 Yujiao Cheng, Liting Sun, Changliu Liu, and Masayoshi Tomizuka, ‘Towards Efficient Human-Robot Collabouration with Robust Plan Recognition and Trajectory Prediction’ IEEE Robotics and Automation Letters 5, no. 2 (2020) 2602–2609.

161 Nancy Velásquez Villagrán, Patricia Pesado, and Elsa Estevez, ‘Cloud Robotics for Industry 4.0: A Literature Review’ Cloud Computing, Big Data & Emerging Topics: 8th Conference JCC-BD&ET, (2020) 3–15.

162 Oscar Afonso, ‘The Oscar Goes to Horizontal Ellipsis Robots or Humans? Competition in a Directed Technical Change Model with Monetary Policy’ Economic of Innovation and New Technology, (2021).

163 Chiara Natalie Focacci, ‘Technological Unemployment, Robotisation, and Green Deal: A Story of Unstable Spillovers in China and South Korea (2008–2018)’ Technology in Society 64, (2021) 101504.

164 Divya Agarwal and Pushpendra S. Bharti, ‘Computation of Cause and Effect Relationship for Acceptance of Autonomous Mobile Robots in Industries’ Journal of Statistics and Management Systems 22, no. 2 (2019) 237–256.

165 Sihem Amer-Yahia, Senjuti Basu Roy, Lei Chen, Atsuyuki Morishima, James Abello Monedero, Pierre Bourhis, François Charoy et al, ‘Making AI Machines Work for Humans in FoW’ ACM Sigmod Record 49, no. 2 (2020) 30–35.

166 Felipe Tobar and Rodrigo González, ‘On Machine Learning and the Replacement of Human Labour: Anti-Cartesianism Versus Babbage's Path’ AI & SOCIETY 37 (2022) 1459–1471.

167 Ajay Agrawal, Joshua S. Gans, and Avi Goldfarb, ‘Exploring the Impact of Artificial Intelligence: Prediction Versus Judgment’ Information Economics and Policy 47 (2019): 1–6.

168 Chia-Hui Lu, ‘Artificial Intelligence and Human Jobs’ Macroeconomic Dynamics 26, no. 5 (2022) 1162–1201.

169 Edward W. Felten, Manav Raj, and Robert Seamans, ‘The Effect of Artificial Intelligence on Human Labour: An Ability-Based Approach’ Academy of Management Proceedings 2019, no. 1 (2019).

170 Tekin Birinci, ‘The Role of Artificial Intelligence and ICT on Economic Growth of G7 Countries’ International Journal of Innovative Technology and Exploring Engineering 8, no.8 (2019) 3251–3253.

171 Katya Klinova and Anton Korinek, ‘AI and Shared Prosperity’ Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, (2021) 645–651.

172 Andrea Gentili, Fabiano Compagnucci, Mauro Gallegati, and Enzo Valentini, ‘Are Machines Stealing Our Jobs?’ Cambridge Journal of Regions, Economy and Society 13, no. 1 (2020) 153–173.

173 Supra, note 175, Edward W. Felten, Manav Raj, and Robert Seamans

Figure 0

Figure 1: Search strings designed for this systematic literature review

Figure 1

Figure 2: Studies identified by the search strings and selection criteria