By
Stuart Kerr, Technology Correspondent, LiveAIWire
The gig economy predates artificial intelligence, but AI has
become its central nervous system. The algorithms that assign delivery
routes, rate driver performance, set dynamic prices, and deactivate accounts
without explanation are AI systems, and they now govern the working lives of
tens of millions of people across the world. The McKinsey Global Institute
estimated in 2023 that independent platform workers number approximately 150
to 200 million globally, and these workers are, in a meaningful sense,
managed by artificial intelligence rather than by human
supervisors.
Algorithmic Management: What It Means in
Practice
Algorithmic management refers to the use of automated systems to
assign tasks, monitor performance, set pay rates, and make decisions that in
traditional employment relationships would be made by a human manager. A
delivery rider on a platform like Deliveroo or Just Eat receives order
assignments through an app. The algorithm takes into account the rider’s
location, their historical acceptance and completion rates, their customer
ratings, and dozens of other variables. The rider cannot inspect the
algorithm, cannot know why a particular assignment was offered or not
offered, and cannot appeal a decision to a human manager who can explain the
reasoning.
Research from Oxford University’s Fairwork project has found that pay
transparency, the ability to understand how earnings are determined, is among
the lowest-scoring indicators across the platforms it assesses globally. The
Eurofound
research on automation and the future of work identifies
algorithmic management as one of the labour market trends most urgently
requiring regulatory attention.
What This Means for You
If you use platform services, the people who serve you are subject
to algorithmic management systems whose design choices directly affect their
earning potential, working conditions, and job security. Consumer pressure
has contributed to improved conditions in some cases: several UK delivery
platforms improved pay guarantees following media coverage of rider incomes,
and the UK Supreme Court ruling establishing Uber drivers as workers entitled
to minimum wage and holiday pay was accompanied by significant public
attention.
As LiveAIWire has examined in coverage of the
shadow workforce of AI, the structural features of platform labour
that concentrate risk with workers and reward with platforms are consistent
across sectors. The gig economy is not a peripheral phenomenon but a central
feature of how AI-enabled economic organisation is
developing.
Deactivation: Termination by Algorithm
The most consequential algorithmic decision in platform work is
deactivation: the removal of a worker’s access to the platform, effectively
terminating their income without notice, explanation, or meaningful right of
appeal. Deactivation decisions are made automatically based on algorithmic
assessment of performance metrics, at a rate and speed that human HR
processes cannot match. The Trades Union Congress found in UK research that a
significant proportion of gig workers surveyed had experienced unexpected
account restrictions or deactivation, with many unable to identify the reason
or access an effective appeal process.
Legal Challenges and the Worker Classification
Debate
The central legal battleground of the gig economy is worker
classification. Platforms classify their workers as independent contractors,
exempting them from employer obligations including minimum wage guarantees,
sick pay, and protection from unfair dismissal. Courts in multiple
jurisdictions have challenged this classification. The EU Platform Work
Directive, agreed in principle in 2024, will create a rebuttable presumption
of employment status for platform workers across member states. The International
Labour Organisation’s work on platform governance provides a
framework for international standards, though enforcement depends on national
governments.
The Future of Platform Labour Governance
Regulatory frameworks are being updated to address algorithmic
management specifically, requiring platforms to provide workers with
information about the systems that manage them, to ensure human review of
significant automated decisions, and to demonstrate that monitoring systems
do not discriminate against protected groups. Trade union organising in the
gig economy is developing, with unions including the GMB and IWGB in the UK
having successfully organised delivery and ride-hailing workers and secured
improved conditions through both litigation and collective
bargaining.
As LiveAIWire has explored in coverage of AI
displacement of traditional employment in agriculture, the
structural question of how economic gains from AI automation are distributed
between those who own the platforms and those whose labour makes them
function is one of the central political economy questions of the current
decade. The gig economy is where it is being contested most visibly, and the
outcomes of current legal and regulatory battles will set precedents that
extend far beyond the delivery sector.
Technology, Power, and the Future of Work
The gig economy debate is often framed as a question about labour
classification, but the deeper question is about power: who has it, who lacks
it, and what mechanisms can rebalance it when the imbalance is harmful.
Algorithmic management concentrates decision-making power in the platforms
that design and operate the algorithms, while distributing risk to the
workers subject to them. The legal and regulatory battles currently underway
are attempts to use public authority to rebalance an arrangement that private
contracting has produced.
Technology is not neutral in this context. The specific design
choices embedded in algorithmic management systems, what is measured, how it
is weighted, what triggers deactivation, and how transparency is or is not provided,
are choices made by people whose interests are not identical to those of the
workers the systems manage. Requiring those design choices to be made more
transparently, and making platform operators accountable for their
consequences, is a different approach from requiring worker reclassification
but pursues the same underlying goal of ensuring that people who work for
platform companies are not subject to arbitrary and unaccountable algorithmic
authority.
The broader question of what AI-enabled platform labour means for
the future of work extends beyond the gig economy. As AI systems become
capable of managing more complex tasks and more diverse workforces, the
patterns established in delivery and ride-hailing are likely to extend to
other sectors. The governance frameworks that are developed or not developed
in response to current gig economy challenges will shape the context within
which that broader transformation occurs.
What Workers Can Do
Workers in the gig economy have more options than the power
imbalance might suggest, though exercising them effectively requires
organisation and collective action that platform labour makes structurally
difficult. Understanding the appeal and dispute processes available on
specific platforms, documenting interactions and performance metrics
independently, and connecting with worker organisations that provide advice
and representation are all practical steps. In the UK, unions including the
IWGB provide support specifically for gig workers navigating platform
disputes and deactivation processes.
The legal landscape for gig workers has improved significantly in
the UK following the Uber Supreme Court ruling and subsequent Employment
Tribunal decisions affecting other platforms. Workers who believe they have
been misclassified as contractors when they should be recognised as workers
are entitled to bring claims to the Employment Tribunal, and several worker
organisations provide support and representation for those doing so. The
limitation period for such claims is three months from the act complained of,
which means that workers who experience unlawful deactivation or pay
practices need to act promptly.
Collective action beyond formal legal proceedings has also proved
effective in specific contexts. Strike action by delivery riders in several
UK cities in 2021 and 2022 produced improved pay terms from some platforms,
demonstrating that even workers without formal employment status retain the
practical ability to withdraw their labour collectively. The effectiveness of
collective action depends on sufficient worker organisation and on the
platform’s inability to immediately substitute non-striking workers,
conditions that are more favourable in some markets and at some times than
others.
About the Author
Stuart Kerr is the Technology Correspondent at LiveAIWire, covering
artificial intelligence across society, policy, and industry. About
LiveAIWire.