Employers could see 'dramatic reduction in productivity, morale, and efficiency if implemented too quickly,' says expert
Facial emotion recognition (FER) and emotion sensing technologies (ESTs) hold strong appeal for many employers seeking innovative solutions for workplace security, attendance tracking, or even performance monitoring.
But according to Jason Walker, program director of industrial organizational psychology and applied psychology at Adler University, there isn’t enough hard evidence yet that the technology is worth the potential cost to employee morale as well as any productivity losses – plus, with the recently prorogued government pushing the pause button on any new AI-related legislation, the legal landscape in Canada will be especially unpredictable.
“Though this may seem like a win for employers eager to implement new technologies, it is inherently risky,” Walker says, pointing to the public and legal scrutiny similar technologies have received in jurisdictions such as the EU where there are stricter policies.
“Regardless of the current legal framework, Canadian employers need to start thinking carefully about what is permissible now and what the legal framework may look like in the future as far as the use of these technologies is concerned.”
Canadian employers unaware of technology laws
According to Robbie Grant, associate of privacy and data protection law at McMillan in Toronto, many Canadian employers are unaware of the complex legal and ethical implications of deploying these tools in their organizations.
“It's easy to be tempted by new, interesting technology like emotion recognition software or location tracking technology or facial recognition time clocks,” says Grant.
“But it's important to note that biometrics come with a high degree of compliance obligations, and where privacy laws apply, it may not even be permitted in the first place to deploy some of these tools.”
When Parliament was prorogued on January 6, AI and privacy legislation Bill C-27, and the Artificial Intelligence and Data Act (AIDA) along with it, died on the Order Paper.
As Grant points out, the death of this bill, which if adopted could have brought Canada in line with similar laws in the EU, means provincially regulated employers will continue to be ruled by a patchwork of provincial laws. A misunderstanding of privacy laws or failure to assess risks can expose businesses to significant liabilities under Canada's patchwork of federal and provincial privacy laws.
“Facial recognition is already regulated primarily under privacy laws, including Canada's federal privacy law, PIPEDA, and provincial privacy laws In Quebec, Alberta and British Columbia,” he explains. “For other provinces, including Ontario and Manitoba and the maritime provinces, there is no comprehensive privacy law.”
Psychological effects of facial emotion recognition on employees
Adding to the complexity is growing concern about the psychological impact these tools may have on employees, as well as the human rights risks those impacts could create.
“The rush to embrace advanced technology does leave out the human factor—how will this impact employees' autonomy and expectation of privacy?” asks Walker.
“An inherent risk to the organization could be a dramatic reduction in productivity, morale, and efficiency if implemented too quickly.”
Even when FER surveillance has positive intentions such as increased safety for employees or increased efficiency, growing evidence indicates that such strategies can cause higher levels of stress and lower job satisfaction, says Walker.
These effects can negate any benefits of the technology: “Workplace surveillance and loss of privacy could have an ‘outsized’ effect on employee trust and morale,” Walker notes, pointing to recent research by the American Psychological Association that found over half (56%) of workers surveyed feel tense or more stressed at work when they are being electronically monitored.
“That perceived monitoring can also create an environment of pressure that runs counter to the stress relief these technologies are designed to offer,” Walker adds.
“When trust is lost it's challenging to rebuild, and the psychological contract one has with the employer is strained … for many it appears that the employer doesn't trust them to make decisions. How safe these approaches are – psychologically – remains to be seen.”
Canadian privacy laws: different rules for different jurisdictions
Grant stresses the importance for employers to understand that every province has unique privacy laws in regards to employee privacy – for example, Quebec’s Act to establish a legal framework for information technology mandates employers perform Privacy Impact Assessments (PIAs) before using any technologies that process sensitive personal information, a requirement not mirrored in other provinces.
“We do see a lot of employers that come to us not fully aware of how privacy laws apply to these new technologies, and many folks have preconceived notions that these things are permitted because they see them in the news, or they know of another company that's implementing them,” says Grant.
"One of the big issues with the way privacy laws are currently in our country, outside of Quebec … is that privacy laws don't have a lot of teeth. There's not a tremendous amount of fines or orders being made," he adds, pointing out that Ontario’s recent electronic monitoring policy under the Employment Standards Act requires employers to disclose monitoring practices but doesn’t limit how it is conducted.
Walker raises additional concerns.
“The lack of a legislative framework gives employers a degree of 'free reign' to create and implement policies that may be in the organization's best interest and not necessarily for employees' legal or ethical benefit,” he says.
“So, can we trust employers to 'do the right thing' when it comes to new technology? I'm not sure.”
The disparity in regulations often leaves employers confused, Grant adds.
"You’d be surprised how many employers assume they’re subject to PIPEDA when they’re not," says Grant, noting that for employers “the starting point is to know what laws apply to you, which is sometimes trickier than you might assume."
High compliance obligations for biometric tools
Facial and emotion recognition tools collect highly sensitive biometric data, which raises the compliance bar for employers. As Grant explains, facial recognition data is considered highly sensitive because it’s biometric, and you cannot change it about yourself," he says. "Emotion recognition data might even be classified as health data, further heightening its sensitivity."
The high sensitivity of these technologies means employers must typically obtain explicit consent from employees, except in certain regulated circumstances. But compliance doesn’t end with obtaining consent. Employers must demonstrate that their use of these tools is necessary, proportionate, and appropriate.
“Existing privacy laws remain in effect, but the problem is that they are unclear regarding these new types of surveillance, which will inevitably be left to the courts to decide within the existing common law framework,” says Walker.
Privacy impact assessments for facial recognition technology compliance
Grant highlights the importance of thorough multi-factor analysis prior to any technology implementation that could compromise employee privacy:
"Is it a bona fide business interest or legitimate need? Would the device be effective in meeting that need? Is there a less invasive means that would achieve the end at comparable cost and benefits?” Grant says.
“If the privacy loss isn’t proportional to the benefits, or if a less invasive means exists, the technology likely won’t be considered appropriate."
While a PIA is only legally required in Quebec, it remains a best practice for employers across Canada, he says. Such assessments help identify risks, evaluate alternatives, and ensure compliance with applicable privacy laws.
For example, an employer seeking to monitor employee emotions for performance purposes might be better served by regular check-ins rather than deploying emotion recognition software or using key cards for workplace access may achieve the same purpose as facial recognition, while avoiding the risks associated with biometric data collection.
“The positive outcomes in trading, where heightened awareness can mitigate risky decisions, may not directly translate to general workplace settings,” Walker says.
“The reliance on AI introduces complexities such as human bias errors and lack of empathy in our choices, and these factors can easily skew the technology's utility and outcomes.”
Mitigating risks and fostering transparency
To navigate the complexities of biometric technologies, employers should adopt a strategic and transparent approach, Grant advises.
“The software won’t tell you what you need to do,” he says.
“They might say it's up to you to make sure you have all the consents and you're complying with privacy laws and how you use our product … and so as an employer, if you don't have a privacy professional on your team, or you're not getting privacy advice, you might not be aware of what those obligations are.”
Walker highlights that when employers implement biometric technology, transparency is crucial for building trust with employees.
“Ethically, asking employees to use this new technology can cause concerns, not just about the learning curve, but what this technology means for job security,” Walker says.
“Another problem is obtaining informed consent from employees; another potential issue may be that workers feel pressured into agreeing against their better judgment. These concerns point to a need for cautious step-by-step measures, while ethical standards must be implemented throughout and without exception.”
Walker also emphasizes the importance of caution, as a failure to do so can damage employee-employer relationships. He outlines three specific steps for employers to ensure legal compliance as well as sound ethical practice:
- Protect employees with ethical guidelines: Develop and enforce policies that govern the use of FER and EST technology.
- Be transparent about FER and EST use: Foster trust by including employees in decision-making.
- Assess and research: Before implementing any products, conduct ‘pilot studies’ internally and gather data on psychological effects of FER and EST use. Balance this data with any productivity gains.
“ESTs have the potential to revolutionize work and productivity, [but] their advancements must be approached with caution,” says Walker.
“Employers must critically examine these technologies' real-world efficacy and the real downsides—namely, the impact on the employee. The risks far outweigh the benefits without strong ethical standards and clearly defined legal conditions.”
