It’s a stark warning. Left unchecked, intrusive worker surveillance tech and Artificial Intelligence (AI) at work could lead quickly to discrimination, work intensification and unfair treatment, the TUC said. TUC polling results published in February 2022 revealed a clear majority of workers (60 per cent) believe they have been subject to some form of surveillance and monitoring at their current or most recent job.

SNITCH GLITCH  In 2021, Amazon announced that it would roll out ‘AI-enabled vans’, with in-vehicle cameras and a driver-tracking phone app. It works as a snitch in the cab, relaying driving behaviour directly to managers, and generating routine machine reprimands whilst driving, which could lead to a loss of bonuses or disciplinary action.

Surveillance can include monitoring of emails and files, webcams on work computers, tracking of when and how much a worker is typing or calls made and movements made by the worker, using cameras and trackable devices.  

The TUC noted there has been a notable increase in workers reporting surveillance and monitoring – up to 60 per cent in 2021, from 53 per cent in 2020.

TUC general secretary Frances O’Grady said worker surveillance “now risks spiralling out of control,” adding: “Employers are delegating serious decisions to algorithms – such as recruitment, promotions and sometimes even sackings. Workers and unions must be properly consulted on the use of AI, and be protected from its punitive ways of working.”

And AI’s punitive power could cost you much more than your job. Your health might not be part of the programme (Hazards 156).

AI is bad for you

While AI can have some positive applications at work, “algorithmic management software at work has proven to negatively impact workers’ health and safety,” academic Miriam Kullmann and European Trade Union Institute safety law specialist Aude Cefaliello note in a January 2022 paper in the journal Global Workplace Law and Policy.

“Continuous monitoring via wearables, for instance, increases work stress while affecting productivity. The way the algorithm allocates tasks and tracks workers affects the work organisation and negates workers’ right to appropriate break time, leading to severe physical and psychological stress.”

Kullman and Cefaliello warn even well-intended interventions and technologies can be used in an unintended and harmful manner, citing the example of delivery drivers. “There are already examples of AI technologies that have been developed with the intended purpose to improve the safety of the driver that are used to monitor the workers once implemented at work,” they note.

Another example is technology introduced with the aim of creating a positive environment at work, which has seen some companies installing cameras that only allow workers to enter the premises if they smile at a camera.

Canon Information Technology’s Beijing office went a step further, with AI cameras preventing a person from doing any operation unless they detect a smile on their face.
Not only is this intrusive, it can be seriously bad for your health.

Researchers from the University of Buffalo and Penn State University, in findings published in 2019 the Journal of Occupational Psychology, concluded workers who feel like they are being forced to smile are more likely to drink heavily as soon as they're off the clock.

AMAZON YUK  ‘Shameful’ figures reveal more than a thousand serious injuries at Amazon UK sites have been reported to health and safety authorities since 2016, a GMB investigation has found.  more

Amazon announced in 2021 that it would roll out ‘AI-enabled vans’ including surveillance cameras and requiring deliver drivers to download and continually run a ‘Mentor’ programme – described as ‘a digital driver safety app’ - on their phones, saying that this would “improve the safety” of the drivers.

But the cameras and app combination operate as a snitch in the cab, reporting driving behaviour directly to managers, and generating routine machine reprimands whilst driving, which could lead to a loss of bonuses.

These risks can be apparent in many jobs, but those using AI are frequently out of view and precarious by design.

Digital platform work and occupational safety and health: a review, a 2021 report from the European Union safety agency EU-OSHA, notes AI-managed jobs “are often active in sectors that are generally considered more dangerous, with higher incidence rates of (severe) occupational accidents, injuries and illnesses. Moreover, platform work involves additional tasks and/or a different combination of tasks from those associated with similar jobs in the traditional labour market.”

In Australia, a spate of deaths of food delivery workers in road traffic accidents in 2021 led to calls for a government enquiry. Concerns about the pressure on these and other platform workers was raised in The Job insecurity report, published by an Australian Senate Committee in February 2022. The report also raised concern about risks working for ride share companies, warning about a damaging ‘Uberisation’ of work.

The report noted: “Evidence from rideshare drivers indicated safety issues relating to the pandemic were not necessarily being addressed in practice, and that the use of an app to manage issues and disputes may be putting drivers and riders at risk.”

Recommendations in the report, which includes a chapter on Challenging Uberisation and the Amazon effect, included making workers’ compensation available to all workers, regardless of employment or visa status, enhance workplace injury surveillance “to capture specific information on the number, and types, of injuries and fatalities for workers engaged in the on-demand platform sector” and recommended that federal safety regulator “Safe Work Australia enhances its national data collection process to capture specific information on the number, and types, of injuries and fatalities for workers engaged in the on-demand platform sector.”

Who’s in control?

Worker safety implications – intended and unintended – should be factored in at the software design stage.

Kullman and Cefaliello say “if the AI system is intended to be used at work, the provider cannot ignore the impact on workers’ health and safety. Also, the provider should take into consideration that the AI should be designed with a view to mitigating monotonous work and work at a predetermined work-rate and to reducing their effect on health.”

They note the impact of physiological and environmental monitoring apps – checking a workers’ body functions including heart rate, skin temperature and movements, for example – depends on the motives of those doing the monitoring.

A worker showing evidence of excessive exertion or fatigue could be told they need to slow down to a more manageable work pace or take a break. Or they could be sacked or docked pay for not keeping up with the rate.

In March 2021, the TUC launched its manifesto, Dignity at work and the AI revolution, for the fair and transparent use of AI at work. It wants the UK government to follow the lead of the European Union, which published draft legislation on 9 December 2021 to give workers rights over algorithms, to stop situations where AI controls appointments, working hours or even dismissals.

Instead, workers would have the right to receive explanations for and to contest automated decisions, while companies would have to ensure workers had access to a human manager where decisions have a significant impact.

In November 2021, the UK’s All Parliamentary Group on the Future of Work called for a new Accountability for Algorithms Act, noting: “Pervasive monitoring and target-setting technologies, in particular, are associated with pronounced negative impacts on mental and physical wellbeing as workers experience the extreme pressure of constant, real-time micro-management and automated assessment” (Hazards 156).

Safe from the start

Kullman and Cefaliello say the negative applications of AI are foreseeable and can be designed out, so “the impact on operational work processes or occupational health and safety must be explicitly considered in the ‘risk management system’ required for high-risk AI systems.

“Providers can contribute to a better and fairer application of AI at work when they develop the software. For example, when they programme an AI to allocate tasks, they should guarantee that the goals are realistic – and not necessarily aiming at economic optimisation. Also, they should compute systems where these goals can be adjusted to individual capacities while avoiding risks of retaliation.”

The desire for higher production and profits, though, can push safety and employment rights down the priorities list. In a 10 March 2022 policy brief, the European Union safety agency EU-OSHA said footdragging on safety measures was because of “some voices” saying it would “put a break on innovation and cause the EU to miss the boat on the digital economy” [see: Safety risks are overlooked].

A 2021 GMB analysis of ambulance call outs to Amazon warehouses revealed injuries spike just before Prime Day, Black Friday and Christmas (Hazards 154).

EU-OSHA says the absence of unions is a factor in the AI risk excess, noting “although there are a few exceptions, digital platform workers are rarely collectively organised, which stands in the way of realising effective worker participation (through information and consultation) in the development of an effective OSH management system."

It notes “the active resistance of some digital platforms to collective organisation have hampered that development.”

Waiting for the technology to be designed, selected and installed is waiting too long.
Whatever its design purpose, AI could hand the employer the option to police, control and fire a worker, because the machine says so. It is a real and present threat to workers.

Unions can’t be an optional extra. In a high tech world more than ever, they must be the workers’ first and ever present line of defence.

Selected sources

• Fast and furious: Amazon’s high churn, high pace work template hurts us all, Hazards, number 156, December 2021. Fast and furious pin up at work poster.
• Dignity at work and the AI revolution, TUC AI manifesto, March 2021
• When AI is the boss: An introduction for union reps, TUC, December 2021.
• Seven ways platform workers are fighting back, TUC, 2021. 
• TUC examples of surveillance and monitoring of workers.
• Miriam Kullmann and Aude Cefaliello. The Interconnection between the AI Act and the EU’s Occupational Safety and Health Legal Framework, Global Workplace Law and Policy, January 2022.
• Occupational Safety and Health in digital platform work: lessons from regulations, policies, actions and initiatives, EU-OSHA, February 2022.
• Digital platform work and occupational safety and health: a review, EU-OSHA, October 2021.
• EU-OSHA digital platform resources
• Draft EU directive on improving working conditions in platform work and related 9 December speech by Commissioner Schmit.
• The Job insecurity report, The Australian Parliament Select Committee on Job Security, February 2022.

 

Over 1,000 serious injuries at Amazon sites

‘Shameful’ figures reveal more than a thousand serious injuries at Amazon sites have been reported to health and safety authorities since 2016, a GMB investigation has found. 

The union said ‘disturbingly’ the number of injuries reported to health and safety bodies is rapidly increasing, with 294 reported in the financial year 2020/21 – up from 231 the year before, or an increase of 27 per cent. This compared to 139 reports in 2016/17.

The true figure could be much higher, GMB said, with not all local authorities responding to its freedom of information enquiries.

The union findings were published on 4 January 2022. Local authority inspection reports obtained by GMB reveal multiple areas of concern, including; injuries to workers, unsafe working conditions, Covid-19 concerns, and poor health and safety record keeping or a lack of compliance with inspectors.

Separate new figures obtained by the union show that more than a thousand ambulance callouts were also made over the same period – and that callouts rose by 56 per cent during the pandemic, between 2019/20 and 2020/21.

GMB national officer Mick Rix said: “More than a thousand serious injuries at Amazon sites is a shameful statistic and one that the company must address urgently – and the true picture is probably worse.

“GMB investigations have now built up years’ worth of evidence and there can be no denying Amazon warehouses are currently dangerous, dehumanising places to work.”
He added: “We will be writing to the Health and Safety Executive to set out our findings - it is time for a proper external audit and investigation of working conditions at this highly profitable company. 

“It’s time Amazon stopped burying its head in the sand, met with GMB and worked out how to make Amazon a great, safe place to work.” 

Union campaigns

• GMB: www.gmb.org.uk/amazon
• Unite: www.actiononamazon.org

Back to main story

Safety risks are overlooked

In a 10 March 2022 policy brief, the European Union safety agency EU-OSHA has said footdragging on safety measures was because of “some voices” saying it would “put a break on innovation and cause the EU to miss the boat on the digital economy.”

It noted: “Very few of the early responses to digital platform work concerned working conditions and occupational safety and health.”

The agency identified for ‘key takeaways for policymakers and decisionmakers’:

• In spite of the recent increased attention to working and employment conditions in digital platform work, the OSH risks that digital platform workers encounter are largely overlooked and remain un(der)addressed. This includes all aspects – from prevention to management;
• Awareness should be raised among digital platform workers and digital labour platforms regarding occupational safety and health;
• Good practices in terms of OSH risk prevention and management should be shared more proactively, both nationally and internationally, to foster learning among stakeholders; and
• Efforts to develop knowledge and gather data on OSH challenges and opportunities as well as on OSH risk prevention and management in digital platform work should be reinforced.

Back to main story • Top of the page  

AI IS NOT OK

Your work van snarls ‘distracted driver!’, the biomonitor says your heart is racing but you’re off the pace, the hand-held scanner sends you scurrying over there when you’ve barely got over here. Hazards editor Rory O’Neill warns management by machine is hard at work, and it will take real union smarts to control Artificial Intelligence.