Platform work and AI in the Global South Flashcards

1
Q

Market despotism

A

Burawoy (1983)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Temporal flexibility trend in all employment, enabling firm flexibility without necessitating firm reliance on contingent workers
Unions are ineffective at compensating for a lack of institutional and structural power

A

Wood (2016)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Zero-hour contracts favour employers, and very rarely favour employees
Other than those who do not require a fixed number of working hours, which is a minority

A

Pennycook et al (2013)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Automation will not inevitably replace human labour - it will change how we work, and what work we do
Technology is not neutral and needs to be regulated to ensure that exploitation does not occur
A key development is the digitisation not of management but of decisions concerning workers, leading to arbitrary implementation undermining workers rights

A

Aloisi & de Stefano (2022) Your Boss is an Algorithm

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

The governance of crowdsourcing.
Looks at how new practices of work are normalised (rationalities)
Scale of crowdsourcing is difficult to ascertain because it is relatively new and growing repidly
Decentralisation of labour, wageless work and precarisation
Self-governance of crowds is an old and efficient form of governmentality
The institutionalisation of informal work on a global scale

A

Ettlinger (2016)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

CPE approach, looking at the materialities and discourses of gig work development in Japan
The move to gig work further legitimises the flexibilisation of labour markets
Abe government has leveraged specific Japanese discourses - such as overworking - to promote the benefits of gig work - such as flexibility - in response to structural pressures that demand new opportunities for economic growth

A

Shibata (2020) Discourse of autonomy

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Problematises the idea that gig workers have flexibility. Study of Uber’s algorithmic management which finds that the app uses ‘soft control’ to encourage self-responsibilisation so that the decisions made align with firm’s
Workers don’t really get to choose where/when they work

A

Rosenblat & Stark, 2016

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Platform work is starting to become more and more like full-time jobs in China, with a full schedule but none of the traditional job security.
Proposes concept of ‘sticky labour’ to understand management as static.
Identifies the tension between flexibility as a motivation for workers to stay, and flexibility as a mode of regulation benefiting employers

A

Sun et al (2019)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Flexible work is not generalisable. Flexibility is relational, relative, culture-and-context dependent’

A

Sun et al (2022)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Across six countries and varied platforms, finds that algorithmic control is central to the operation of online labour platforms.
Most effective form of control was the rating system.
Flexibility myth - workers could not convert purported flexibility into reality, which translated to long working hours and working in isolation
Job quality was determined by individual marketing bargaining power, which prevents worker solidarity by engendering competition. Skills and platform reputation were found to be the most important indicators

A

Wood et al (2019) Good gig, bad gig

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Platforms are not middle men but exploit regulatory loopholes and reproduce workforce precarity
Various forms of nudging (so similar to Rosenblat & Stark, 2016) align worker decisions with company interests
However, workers become familiar with company policies (‘qualculacion’). They tolerate these as long as they align with their own interests - high rates of worker turnover

A

Shapiro (2018) Between autonomy and flexibility

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Fragile art of multi-apping. This is not a collective form of resistance but an individual coping mechanism which further reifies the platform’s power
Other forms of everyday resilience include using shortcuts, postponing orders and upgrading vehicles
Differentiates between resistance, resilience and re-working, and posits the snapping/resilience cycle
Companies still use the same management strategies e.g. a piece-rate pay system (Burawoy, 1983) but new ones as well such as info asymmetry
Demonstrates an exploitative work arrangements which leads to ever-more intensive work regimes

A

Popan (2024)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Finds that the power of worker agency is constrained by work organisation
Agency is predominantly entrepreneurial - low-level expressions aimed at materially improving individual conditions
In some cases this reinforces capital accumulation - e.g. multi-apping allows platforms to claim workers are independent contractors

A

Barratt et al (2020) Australian case study of Uber Eats & Deliveroo

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Multi-apping is argued to be a form of resistance with subversive potential.
Actually, I think this is a tenuous claim since this only temporarily prevents control of the platform and it actually benefits from it anyway. To be sure, this is an agency that should be acknowledged but I don’t think this is incompatible with also seeing that this agency occurs within fundamentally oppressive wider structures

A

Tironi & Albornoz, 2022

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Studies platform work in Jakarta to propose the driver’s view from within, which is more relational and experiential than the abstract vision of the algorithm
This manifests materially in unpaid labour on the part of drivers, who expend energy on figuring out unanticipated hurdles, shortcuts, shelter, police etc
Worker subversions to make money (e.g. automatically accepting any order) can align with platform goal of maximising income

Spatiotemporal rhythms of the city

A

Qadri & D’Ignazio (2022)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Women in the gig economy - this gendered dynamic is often unexplored but plays a role in shaping the operation, outcomes and experiences of digital labour platforms.
For example the constraints of motherhood and care (Which are far from novel for feminist scholars) constrain the freedom of remote female workers, despite rhetoric that remote work breaks down gender barriers - e.g. needing to watch the children means you cannot accept more lucrative jobs with tight turnarounds; losing clients while pregnant and having to build up a base again
Digital labour platforms further reinscribe pre-existing gendered labour market inequalities in new ways

A

James (2022)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Going ‘Karura’ in Kenya - an analysis of Uber drivers, who went on strike in July 2018.
‘Karura’ refers to the forests, which are associated with political resistance against British colonialism and later state development policies. This conjured up the continuity of colonial, political and corporate dynamics of exploitation and demonstrates the politicised nature of the supposedly unbiased digital economy
This forefronts a tension between the ‘subjectification’ from above (where Uber construes workers as independent contractors) and labour subjectivities created from below (how workers actually understand their identity) - the former is closely linked with collective agency, embedded in the Kenyan context

A

Iazzolino (2023)

18
Q

An overview of unpaid labour in cloud-work, defined as remotely performed labour mediated by digital labour platforms
Technically geographically untethered labour has geographically contingent outcomes with cloud-work often functioning as an engine of South-North value extraction
Southern workers do more unpaid labour on average than their Northern counterparts
So unpaid labour is a systematic mechanism of how platforms make money + shift risk onto workers + dodge traditional responsibilities in three ways: unpaid wages for done work; unpaid time searching for jobs; unpaid aspirational labour (upskilling)
99 Designs: design tasks are posted as contests and workers must create briefs that may not even be used

A

Howson et al (2022)

19
Q

Case study of socio-spatial networks as coping mechanism for lack of state/platform support in migrant drivers, India
Platforms promise better incomes but expose workers to the same precarity as the informal economy
Important to remember that this was the standard beforehand
New ‘digitally organised informality’

A

Ray (2024)

20
Q

Platformisation - the introduction of internet apps that consolidate dispersed small-scale producers under the idea of entrepreneurial self-employment
‘The penetration of infrastructures, economic processes and governmental frameworks of digital platforms in spheres of life, and the reorganisation of cultural practices around these platforms’
Driven primarily by US-based companies
Platforms are defined by their digital nature, their mediation between end-users and complementors, and their systematic collection of data and algorithmic processing

A

Poell et al (2019)

21
Q

New regime of ‘organised informality’
In the GS, platform apps are a further extension of pre-existing neoliberal forms of economic and labour governance in the GS
Embedded in historical dynamics of social hierarchies - reproducing inequalities of caste, class race and gender in the contemporary labour market
Facilitated by the systematic recruitment of the most deprived to serve in the low-end service economy - although training is formalised (often with state co-operation), terms of employment remain precarious
‘The emotional proletariat’ refers to how employers control not only the physical labour but also the emotions of their workers to develop the right skills and attitude - the affective expression and corporeal embodiment of servility

A

Gooptu (2013)

22
Q

March 2024 launch of Platforms and Society journal - showing how recent and ongoing this work on platforms is.
Platform studies are interdisciplinary
Most work is in Global North - so lots of work to be done on the unequal material and spatial repercussions of platformisation, especially re gendered, racialised and (post)colonial trajectories
AI is rapidly turning into another platform industry

A

Chen et al (2024)

23
Q

Integrating the digital into global production value networks
GPN is a key framework that has been used to understand the relations between economic production and its actors - but how can it be used to understand growing role of the digital
The digital is not just an infrastructure to be added, but is itself central to the management, make-up and monitoring of global dispersed networks

A

Foster & Graham (2017)

24
Q

The micro-work that goes into supporting AI is divided into three parts: AI prep, AI verification and AI impersonation.
There is a gap in the literature on the place of this labour in AI, and the nature of its embeddedness in data companies.
The paper finds that, rather than human labour being an initial form of support that will no longer be needed once technology ‘matures’, it is rather a structural component of production processes because data availability will never reach a steady state - verification will always have to be performed.
This will lead to further marginalisation and precariousness (heteromation) of human workforces - AI sells a techno-utopian vision but is in fact a labour-intensive, poorly paid and low security sector.

A

Tubaro et al (2019)

25
Q

Heteromation - automation replacing human labour is a myth. Rather, humans will operate on the margin of machines and demand for labour will remain high (need to ask where)

A

Ekbia & Nardi (2017)

26
Q

The paradox of automation’s last mile is the incessant creation of residual human tasks as a result of technological progress

A

Gray & Suri (2017)

27
Q

OECD reports on digitalisation of labour - the first report found the impact to be ambiguous as tech can be labour-saving, but not to the point of replacing old jobs
The second report found that digitalisation did not pose a real risk of destroying any significant number of jobs in the foreseeable future - but the nature of tasks would change significantly

A

OECD (2016, 2018)

28
Q

France-Madagascar case study looking at how/if AI changes pre-existing labour relations.
Finds that rather than AI replacing jobs, it has led to a lengthening of the externalisation chain since France outsources AI tasks to Malagasy workers - performing the kind of work outlined by Tubaro et al (2019)
Division of labour, constant updating required, deep embeddedness combined with invisibility - what is sold is a production process, not a finished algorithm
AI as a sociotechnical device

A

Ludec et al (2023)

29
Q

The ‘digital factory’ - evoking a historic and contemporary site of strenuous and menial work.
The age of automation is actually a new era of exploitation because human labour remains an indispensable precondition for today’s global capitalism.
New digital regimes bear many similarities of an early, Taylorist age e.g. tendencies twoards de-skilling and surveillance of workers.
So digital capitalism is not the end of the factory - it is the factory’s explosion beyond its traditional walls, its multiplication and its spatial reconfiguration
Case study of World of Warcraft and gold farmers in China who are paid to mine gold.

A

Altenreid (2022)

30
Q

Case study of women in crowdwork. Many Northern American women on Mechanical Turk, because spatial/temporal flexibility is well-suited to people performing unpaid care and housework.
Gendered division of reproductive labour - both home-based labour and its devaluation is now new (Marx referencing lace-making as an external department of the factory)
The myth of the housewife who sews/microtasks for fun and thus does not require a proper wage
Platforms are not neutral: they assume many of the socio-spatial functions of the traditional factory with a high degree of standardisation

A

Altenried (2020)

Link to James (2022); Fairwork (2022); Rodriguez-Moron et al (2022)

31
Q

Digital decolonial turn.
Pitfalls of colonialism narrative: trivialising/generalising colonial pasts/geographies (many former colonial empires are not the most demanding whilst US has switched focus); predigital/digital dualism; orientalism that takes away agency
‘Coloniality’ - long-standing patterns of power that emanated from colonialism, but that has impacts beyond limits of colonoial administrations. Also manifests through norms, collective identities and aspirations

Ludec et al (2023)

Colonialism? GN/GS divide re data/tasks being bought/sold. Racialised/gendered minorities as cheap sources of work. Parallels with past outsourcing dynamics.

A

Casilli (2017)

32
Q

A range of risks and costs for workers in the GS (SSA and SE Asia) - bargaining power, economic inclusion, intermediated value chains and upgrading
Some workers can thrive through individualisation discourses, however this perspective neglects structural issues such as imperfect info and alientation which harm workers that are unable to navigate the workplace
Bargaining power is undermined by size and scope of the global market
Governments are increasingly viewing digital labour as a mechanism for escaping poverty and achieving economic growth
Digital work has distinct geographies

A

Graham et al (2017)

33
Q

Case study

Facebook. AI has been offered up as the solution for vast amounts of content moderation. However, the overwhelming majority of what is being automatically identified e.g. as hate speech are copies of content that have already been reviewed by a human
Facebook has used statistics to obscure the role of its human content moderators by implying that AI is accurately spotting new instances of content, not just variants of old ones

A

Gillepsie (2020)

34
Q

Vendors are found to strategically disclose information about how, where and for whom AI services have been produced to their clients, depending on commercial interests
However, when vendors co-produce AI with clients, they must explain in clear terms the mundane efforts and limits of labour
Exemplifies the tension between the techno-utopian vision of AI and its reality of arduous and menial labour

A

Newlands (2021)

Link to Gillepsie (2020) on hidden AI work, Facebook

35
Q

Crowdwork defined as a type of digital labour and ‘micro’ work - firms requiring services that crowdworkers can sign up to complete.
This is both a territorialisation of online work and its deterritorialisation.
Precarity is imposed whilst firms save money on space and wages etc
Work is routine and repetitive, analogous to traditional factory manufacturing work (‘digital sweatshops’)
The institutionalisation of informal work
The new digital infrastructure is a labour regime which reconfigures pre-existing capital labour relations to institutionalise informal work whilst eliminating traditional costs
Workers freely self-enrol due to the absence of alternatives - however, the informalisation of work makes it unamenable to traditional forms of protest

A

Ettlinger (2017) Problems and paradoxes of online work platforms

36
Q

An example of strikes in Greece - targeting platform’s network effects, which is part of its monopoly

Links to Davies et al (2023)

A

Tsardanidis (2024)

37
Q

In favour of digital colonialism. Using Marxist theory to argue that digital division of labour is an extension of historical imperialism and work in the peripheries is crucial to maintaining corporate empire

A

Fuchs (2016)

Rebut with Casilli (2017)

38
Q

The power of platforms: precarity and place

An overview of platforms.
Underexplored: implications for inequality, civil society and role of place
Platforms derive their value from organising the production of goods/services
They have both hard economic and political power and softer forms of cultural power
Regulatory backlash using both traditional and targeted regulation
Platform co-operativism as an alternative
Digital platforms avoid responsibility but this may be changing

A

Davies et al (2023)

39
Q

In China, platform economy is more embedded in local life due to stronger state interventionist approach

A

Chu et al (2019)

Link to Krishnan (2023)

40
Q

Top-down platform interventions in Varanasi, Utter Pradesh, India. Disproportionate impact of surveillance technolgoies on women’s groups and health care workers

A

Krishnan (2023)

Links to Chu et al (2019)

41
Q

Gender equality across digital labour platforms in Europe.
38% of platform workers are women in Europe, and there are pervasive gender inequalities.
Women platform workers engage more frequently in domestic and care work platforms.
They work with lower intensity and have less capacity to allocate time to platform work - reducing their earnings - due to their personal caring responsibilities.
Flexibility cannot be fully realised.

A

Rodriguez-Modrono et al (2022)