Workplace AI surveillance is not only on the rise, it is also discriminating

Working Women Queensland assists women who are vulnerable to excessive workplace surveillance. Wotton Kearney’s Emma Campbell was recently seconded to WWQ to look at how data is obtained improperly and used in a discriminatory way.

Eloise Dalton,
Director,
Working Women Queensland

Emma Campbell,
Associate,
Wotton Kearney [seconded to Working Women Queensland in 2024

In Australia, organisations like Working Women Queensland (WWQ)1 are seeing surveillance evidence weaponised by employers against working women who attempt to exercise their rights as workers or who make complaints against management.

Recognising that the use of artificial intelligence (AI) tools to conduct workplace surveillance and monitoring stands to exacerbate these existing issues, Wotton Kearney2 has partnered with WWQ to explore potential legal safeguards.

Grace is 25, works in retail and is of Asian ethnicity. As she is young, female, of colour and low paid, she is 49% more likely than her counterparts to be subject to workplace surveillance. If she had no union representation, this figure would increase to 69%.3

Working Women Queensland seeks to help women like Grace who are vulnerable to excessive workplace surveillance. The use of AI tools raises concerns that consent to the collection and use of data obtained through surveillance or monitoring has not been properly obtained, or that the data will be used in a discriminatory way.

This partnership is a case study in how targeted pro-bono assistance can enable community legal organisations to engage in proactive advocacy.

The rise and prevalence of workplace surveillance

The COVID-19 global pandemic accelerated the adoption of emerging surveillance technologies by employers. The number of employers using monitoring tools to monitor their employees in 2022 was estimated at 60%, having doubled since the beginning of the pandemic.4

In a survey conducted by global law firm Herbert Smith Freehills, 90% of respondents indicated they use software to monitor their employees’ location when they work remotely,5 82% of respondents said they planned to use digital tools to monitor workers’ wellbeing in the future, and 43% already use technology such as sentiment analysis software to detect and address wellbeing issues at work.6

Submissions in response to a recent Australian state parliamentary inquiry into workplace surveillance indicate excessive workplace surveillance and monitoring spans industries and employment types.7 The evidence and data collected from these surveillance and monitoring practices is often used for the purposes of unjustly disciplining and terminating employees.8

Traditional methods of surveillance range from CCTV cameras to GPS tracking and monitoring of emails. However, AI-driven surveillance can also include collection of biometric data, use of facial recognition technology in CCTV cameras, logging keystrokes and taking routine photographs through computer webcams.

AI software allows companies to monitor the apps and websites visited by employees as well as taking periodic screenshots of employees’ screens. AI systems can analyse performance metrics, evaluate emotional states through video or voice data, and even infer intentions or ‘internal states’, such as the likelihood of leaving a job, union membership or political orientation.

Are workers really giving their consent to how their data is collected and used via AI surveillance tools?

The issue of consent is woven into both the carrying out of workplace surveillance and the use and control of data collected. In Australia, the laws governing workplace surveillance are inconsistent across the federation of states.

In fact, such laws only exist in three jurisdictions out of seven, making them overall less effective and undermining the privacy of those being surveilled.9

The laws may need to catch up to effectively regulate more sophisticated monitoring such as wearable devices which measure biological signals such as respiration, heart rate, skin-conductance levels and speech.

The existing laws often emphasise giving notice the surveillance is happening, rather than asking for consent before it happens. This means employees do not have genuine power to influence the conduct of the surveillance and may feel coerced into agreeing to intrusive monitoring practices.

Issues of informed consent, well developed in Australia’s privacy law, need to be considered in the workplace surveillance context.

Firstly, workers are unable to truly provide meaningful consent to the use of their personal and sensitive information without a clear idea of how the AI system uses the data obtained.

Secondly, the experience of function creep – where data is used beyond its original purpose – means workers may not know how their information will be handled. They may consent to its collection for the original purpose, for example health data to enable targeted adjustments or supports, but not for a secondary purpose, such as profiling that may lead to biased decision-making about their work capacity.

Requiring transparency about how the AI system works is critical to allow workers 1) to request corrections to personal information collected or generated, and 2) to understand and challenge decisions made with this information.

Even with full transparency, the power asymmetries which exist within the workplace means the consent provided may not be genuine, for example if employees like Grace fear repercussions from withholding consent.

“The use of AI systems to monitor employees and the collection of vast quantities of employee data (both incidentally and intentionally) increases the potential for employers to use the information in a punitive or discriminatory way.”

Implications for working women: function creep and discrimination

Function creep is almost unavoidable when implementing AI systems, given AI’s ability to generate personal information and make inferences from seemingly de-personalised data, such as health, emotions/mood or political orientation (often without the employee’s knowledge).

The use of AI systems to monitor employees and the collection of vast quantities of employee data (both incidentally and intentionally) increases the potential for employers to use the information in a punitive or discriminatory way.

For example, working women may be required to use wearable devices with biological sensors, unintentionally giving employers information about their health status, including early indicators of pregnancy or cognitive decline. Due to the persistent cultural issues and gender bias present in workplaces, there is a significant risk such information will be used against working women as a basis for unfair decisions such as withholding promotions or terminating employment.

What safeguards should we be pushing for?

Bringing together our casework experience and legal research, including the protections being developed in other jurisdictions globally such as the United States [with varying state-level privacy protections], and Europe, which employs broad regulations like the General Data Protection Regulation, WWQ is advocating for:

  1. Workplace surveillance laws: Unifying and strengthening Australia’s workplace surveillance laws to ensure proper regulation of excessive surveillance practices and ensure the monitoring is proportional to its purpose.

  2. Requiring appropriate consent: Ensuring privacy laws clarify consent should be voluntary, informed, current, specific and unambiguous and ensuring employers are in practice restricting collection of information to that which is necessary for or related to the employer’s business activities.

  3. Bias audits: Requiring AI systems to undergo regular bias audits to avoid further increasing the risk of discrimination based on race, gender, age, or other protected characteristics.

  4. Union Involvement: Involving unions in decision-making about surveillance tools to ensure deployment aligns with employee interests and to alleviate the burden on employees to understand information handling practices.

1 Working women centres have been established around Australia to provide free legal and industrial advice, information, support and representation to vulnerable, non-unionised workers about their rights at work.
2 Wotton Kearney is Asia Pacific’s insurance & risk legal business. With a unique team, WK partners with forward thinking insurers and corporates to solve their greatest challenges. More than just a law firm, we’re a community of big thinkers, trail blazers and impactful humans that’s committed to making a difference. Together, with our clients, our people and our communities, we’re creating a legacy that redefines the world of insurance and risk. Together, we are greater.
3 Institute for Public Policy Research, ‘Watching me watching you: Worker surveillance in the UK after the pandemic’, March 2023, p20.
4 Gartner, ‘The Right Way to Monitor Your Employee Productivity’, 9 June 2022 <https://www.gartner.com/en/articles/the-right-way-to-monitor-your-employee-productivity#:~:text=According%20to%20Gartner%20research%2C%20the,within%20the%20next%20three%20years.>
5 Herbert Smith Freehills, ‘Remote/Controlled: The Future of Work Report 2021’, p30.
6 Herbert Smith Freehills, ‘Remote/Controlled: The Future of Work Report 2021’, p18.
7 United Workers Union, ‘Submission to Parliament of Victoria Inquiry into Workplace Surveillance’, July 2024.
8 United Workers Union, ‘Submission to Parliament of Victoria Inquiry into Workplace Surveillance’, July 2024.
9 Australian Law Reform Commission, ‘Final Report 123: Serious Invasions of Privacy in the Digital Era’, June 2014.

Quicklinks

Choose the page you wish to view

PSMG LLP

The White Collar Factory
1 Old Street Yard
London
EC1Y 8AF

Tel: 020 39481087

Managing Partner: gail.jaffa@psmg.co.uk

Share this brochure through your social media channel!
Chose from: