Emotion AI in the Workplace: A Solution Looking for a Problem?
- chris251714
- 7 days ago
- 5 min read
Over the past two decades, organisations have invested heavily in tools designed to improve hiring, performance management, employee engagement, and productivity. Most of these initiatives have been built around a simple principle: helping leaders make better decisions through better information.
A new generation of technology companies is now proposing something very different. Rather than measuring outcomes, performance, or behaviours, they claim to be able to measure emotions.
The premise sounds straightforward enough. By analysing facial expressions, voice patterns, eye movements, and other behavioural signals, organisations can supposedly gain insight into how employees are feeling throughout the working day. Some proponents argue this can improve wellbeing, identify burnout, strengthen engagement, and help managers support their teams more effectively.
The idea may sound appealing in theory. In practice, it raises far more questions than answers.
The first issue is one of relevance.
Most organisations assess employees based on their performance over time. They look at results, delivery, leadership, collaboration, client impact, commercial outcomes, and achievement against objectives. They do not assess employees based on whether they appeared happy, frustrated, tired, or distracted on a particular Tuesday morning.
Human beings are not machines. We do not wake up each day with identical levels of energy, motivation, concentration, or enthusiasm. We all experience good days and bad days. We arrive at work carrying the realities of everyday life with us.
An employee may appear distracted because they slept poorly the night before. Another may seem withdrawn because a family member is unwell. Someone else may be concerned about their finances, recovering from illness, dealing with a relationship issue, or simply having an off day. None of these situations necessarily have any bearing on their long-term contribution to the organisation.
Even if a system could accurately identify that someone appears unhappy, it cannot tell us why.
That distinction is critical.
Much of the discussion surrounding emotion recognition technology assumes that observable expressions are a reliable indicator of internal emotional states. Yet common experience tells us otherwise. People smile when they are nervous. They remain expressionless when concentrating. They appear calm when stressed and enthusiastic when disengaged. Human behaviour is highly contextual and often difficult to interpret even for those who know us well.
Research in this field has made significant advances in recognising facial expressions and vocal patterns. Companies such as Affectiva have spent years developing systems designed to classify observable behaviours. Academic datasets such as Microsoft's FERPlus have helped researchers train increasingly sophisticated models.
However, recognising an expression is not the same as understanding an emotion.
Perhaps more importantly, understanding an emotion is not the same as understanding a person.
One of the more interesting aspects of the research is that even human observers frequently disagree when interpreting the same facial expression. What one person sees as frustration, another may see as concentration. What one observer interprets as sadness, another may view as neutrality. If human beings struggle to reach agreement when assessing emotions, it is difficult to see how an algorithm can reliably remove that uncertainty.
The challenge becomes even greater when organisations attempt to use these systems in employment settings.
Suppose a manager receives an alert indicating that an employee appears stressed or disengaged. What exactly are they expected to do with that information?
Are they supposed to initiate a conversation?
Are they expected to document it?
Should they escalate it?
Or should they simply ignore it?
None of these options is particularly attractive.
The reality is that most managers in financial services, technology, consulting, and professional services are not trained psychologists, therapists, or counsellors. They are leaders who have been promoted because they can manage clients, build businesses, deliver projects, generate revenue, and lead teams through complex challenges.
Expecting them to interpret emotional signals generated by an algorithm creates responsibilities that many are neither qualified nor equipped to handle.
There is also a risk that organisations begin to blur the boundary between professional and personal life.
Most employees understand that their employer has a legitimate interest in their performance at work. They generally accept being assessed on results, behaviours, competencies, and contribution. What many do not accept is the idea that their emotional state should become a matter of organisational interest.
A healthy workplace culture encourages open communication. Employees should feel comfortable speaking to managers when they need support. Equally, they should have the right not to discuss personal matters if they choose not to.
That balance is important.
The danger of emotion monitoring is that it risks turning voluntary conversations into expected conversations. An employee who would otherwise have chosen to keep a personal matter private may suddenly feel pressure to explain themselves because a system has flagged a perceived change in mood or behaviour.
The result is not necessarily greater trust. In many cases, it may achieve precisely the opposite.
Trust is built when people feel respected, supported, and treated like adults. It is difficult to create that environment if employees believe they are being continuously monitored and analysed.
This concern is reflected in the growing regulatory scrutiny surrounding emotion recognition technologies. European regulators have expressed particular concern about their use in workplaces, citing questions around reliability, privacy, discrimination, and the imbalance of power between employers and employees.
That concern is understandable.
Even if these systems become technically more accurate over time, the fundamental question remains unchanged: what problem are they actually solving?
A high-performing executive may have a difficult week while continuing to deliver exceptional results. A struggling employee may appear enthusiastic while underperforming. Emotional signals rarely provide a complete picture of capability, commitment, or performance.
After more than twenty-five years working across executive search, talent acquisition, and leadership advisory, I remain convinced that the most effective assessments of people are still built around judgement, conversation, evidence, and outcomes.
The best leaders understand their teams because they spend time with them. They build relationships. They listen. They observe. They create environments where people feel comfortable speaking openly when they need support.
Technology can be a valuable tool in that process, but it should not replace human judgement.
Nor should it encourage organisations to believe they understand individuals simply because an algorithm has assigned a score to a facial expression or a voice pattern.
The workplace is not a therapy session, and managers are not clinicians. Organisations should absolutely care about employee wellbeing, but there is a significant difference between providing support and monitoring emotions.
The future of work should focus on helping people perform at their best, not analysing how they feel every moment of every day.
Just because something can be measured does not mean it should be.
Emotion recognition technology may be one of the clearest examples of that principle.
At C Graham Consulting (CGC), we advise organisations on one of the most important and often most challenging aspects of business: people.
With more than twenty-five years of experience across executive search, talent acquisition, leadership advisory, and executive coaching, we work with organisations throughout Europe, North America, the Middle East, and Asia to identify, assess, and appoint senior leadership talent.
Our focus has always been on understanding the individual behind the CV. Leadership capability, judgement, character, track record, and cultural fit cannot be reduced to a single score or algorithm. While technology can support decision-making, the most important hiring and leadership decisions still require human insight, professional judgement, and meaningful conversation.
We believe the future of work should combine the best of technology with the best of human leadership. Organisations succeed when they build cultures based on trust, accountability, communication, and performance,not surveillance.
To discuss executive search, leadership hiring, executive coaching, or talent strategy, visit www.cgrahamconsulting.com or connect with Christopher E.D. Graham on LinkedIn.
.png)
Comments