As AI moves from a novelty to a necessity, HR finds itself in the “messy middle” – balancing rapid productivity gains against the risk of skill atrophy and social erosion. In part one of this two-part series, we explore the hidden human costs of automation and the downstream risks of AI facing the HR profession.

For some time now, questions about artificial intelligence in the workplace have been largely practical: How does it work? What are the use cases? Which platforms should our business adopt? But as AI embeds deeper into core work processes, the questions are becoming harder to answer. 

Ethical grey areas, governance gaps and unforeseen wellbeing impacts are dampening the initial excitement around AI, and organisations are beginning to recognise the long-term consequences of scaling AI-enabled work without paying equal attention to work design, wellbeing and governance considerations.

“We can’t solely value speed and efficiency over some of those longer-term outcomes, which are all about the impact of this technology on people,” says Dr Ben Hamer FCPHR, futurist, founder of ThinkerTank, AHRI board member and chair of AHRI’s Future of Work Advisory Panel.

Dr Emmanuelle Walkowiak, Vice-Chancellor’s Senior Research Fellow at RMIT, has spent two decades researching how the rapid adoption of technologies drives the transformation of work and organisations.

“HR is in the messy middle of AI more so than other functions, because HR must simultaneously be a user and a workplace regulator of AI systems,” she says. “HR decisions often have the most direct impact on workers – and are where the  sensitive ethical or legal questions arise.”

Despite businesses wanting to solve these challenges quickly, a rushed approach goes against the very nature of their complexities, says Hamer.

“We don’t even understand the problems deeply enough to be able to solve them. The fact that we can’t offer practical tips for these challenges is a real indication of where we are heading as a profession.

“There is no best-practice guidance for these dilemmas. We need to attract people in HR who can thrive in uncertainty and can get comfortable with ambiguity.”

With this in mind, the following three dilemmas aren’t designed to offer black-and-white guidance, but to help navigate the grey areas, with suggested thought-starters to raise with the executive or board.

Dilemma #1: Psychosocial safety and relational elements of work

Improvements to physical working conditions or workflows are often accompanied by faster, more demanding workloads, says Walkowiak.

Consider the introduction of email in the workplace. It unlocked an incredible amount of opportunity for instantaneous communication, but also immediately increased the expectation of an instantaneous response, meaning the majority of knowledge workers now spend more and more of their work hours wading through their inboxes. 

For the same reasons, AI is intensifying the rhythm of work.

“Part of my research focuses on evaluating psychosocial risk through working conditions. We look at indicators such as job demands, job control, emotional demands, unfairness at work, poor workplace relationships, conflict and job insecurity.”

One of the most pertinent protective factors, she says, is the relational dimension of work – meaning how supported, secure and connected employees feel.

“When AI systems automate processes such as performance evaluation, you might get more precise metrics, but you lose that relational element,” she says. “The social interactions that help align perceptions between employers and employees are missing – and that’s what contributes to increased emotional demands, stress about job security and uncertainty about the future.”

“We don’t even understand the problems deeply enough to be able to solve them. The fact that we can’t offer practical tips for these challenges is a real indication of where we are heading as a profession.” – Dr Ben Hamer FCPHR, futurist, founder of ThinkerTank, AHRI board member and chair of AHRI’s Future of Work Advisory Panel.

There are also implications for workplace civility. Recent research from The Michelle McQuaid Group found that civility scores were 22 per cent lower among employees who said they often use AI technologies (67 per cent) versus those rarely using AI (86 per cent).

“Asking questions about human-machine interactions and the implications on social interactions is important, but it’s quite complex,” says Walkowiak. “AI interactions may shape our communications, but not always in the way we first assume.”

She uses the example of using AI for writing assistance, such as when emailing clients or colleagues. 

“We’re creating these overly polished, formal and structured ways of communicating with each other. That means we can lose the spontaneity and authenticity that comes from unfiltered peer-to-peer interactions. All of our messages are becoming ‘optimised’, so to speak.”

This is making it harder to discern someone’s true tone, intent or emotion from their message, she says.

“Then you layer in the fact that we’re increasingly working in hybrid or remote environments where social interactions are already being reduced. Generative AI could lead to [authentic] communication being avoided.”

As a result, our tolerance levels could start to weaken.

“In real-world organisational structures, we rarely have full clarity. There are delays, escalations and negotiations – that’s normal,” says Walkowiak. “But if we start expecting the same immediacy from our colleagues that we get from AI, it could erode our patience. It also risks weakening our interpersonal skills, such as the ability to navigate ambiguity and maintain trust through human interaction.

“The disagreements and moments of tension that are essential for collaboration – we’re losing them. It’s not just about politeness. It’s about preserving the human dynamics that make work meaningful.”

HR thought starters for this dilemma:

• How is AI reshaping the rhythm and relational fabric of work within our organisation?
• What safeguards are in place to ensure the drive for productivity gains doesn’t erode the social capital and trust that underpin sustainable performance?
• Are we measuring the right indicators, such as civility, connectedness and perceived fairness, to assess the impact of AI on psychosocial safety and culture?
• How are leaders being equipped to model and maintain authentic human connection in increasingly digitised environments?

Gain clarity and stay ahead of AI-driven change with the knowledge and tools to keep your HR practices compliant and your organisation protected with this course from AHRI.

Dilemma #2: The impacts on wellbeing and capability building

High AI use could be fuelling burnout and stress in our workforce.

Dr McQuaid’s research found that those often using AI had less self-compassion (52 per cent) compared to those rarely using it (72 per cent). Those who sometimes used AI were also far less likely to report strong wellbeing (68 per cent) than those who rarely used it (98 per cent).

“When we’re working, we operate in peaks and troughs in terms of cognitive load,” says Hamer. “Knowledge work is becoming so much more intense because we’re seeing more of the boring, routine work being automated.

Listen to AHRI’s podcast with Dr Michelle McQuaid here, where she unpacks her AI research in more detail.

“People talk about how exciting that is and how it enables us to focus on more value-adding work, but those ‘lower-value’ tasks are often how our brain regulates itself.

“Think about how you feel when you come out of a high-energy workshop. You might be buzzing, but you walk out of the room feeling knackered. You might spend the afternoon tending to your emails or doing some invoicing or expenses to recover.

“When you don’t have that cognitive break, that’s what can lead to burnout. It’s not necessarily that we’re doing more work, but it’s the nature of the work that’s becoming exhausting.”

Walkowiak adds that another concern is the huge amount of work AI is able to generate within seconds.

“Instead of creating, workers often find themselves reviewing, verifying or managing information, which can lead to cognitive overload and new forms of stress,” she says.

“HR is in the messy middle of AI more so than other functions, because HR must simultaneously be a user and a workplace regulator of AI systems.” – Dr Emmanuelle Walkowiak, Vice-Chancellor’s Senior Research Fellow at RMIT

Another issue – one many organisations may be overlooking in pursuit of short-term productivity gains – is the gradual erosion of workforce capability, or “skill atrophy”.

“Some research, including from MIT, suggests that when we rely too heavily on these systems, our cognitive abilities can begin to atrophy,” says Walkowiak.

“The brain needs time and effort to process and store information. It’s much like London taxi drivers who, before GPS, developed enlarged brain areas related to spatial navigation. When we outsource too much of that mental work, we risk losing some of those capabilities over time.”

Hamer highlights the example of a leader he worked with who had the opportunity to automate his company’s call centre, but chose not to.

“He recognised that some of the organisation’s strongest middle managers had come from that environment. They were people who understood customers deeply, communicated effectively and knew how to handle sensitive conversations. From a succession pipeline perspective, it didn’t make sense to automate that part of the workforce,” he says.

“We’re often too fixated on productivity and efficiency. Many organisations are aiming to grow over the next 12 to 24 months without increasing headcount, which inevitably means more AI. But we’re thinking about the end result without considering the implications of how we get there.

“HR are going to be the ones dealing with the downstream implications of this in five or 10 years’ time, so they have a vested interest in getting on the front foot of it now,” says Hamer.

HR thought starters for this dilemma:

• Are our wellbeing strategies keeping pace with the ways AI is changing the cognitive demands of work in our organisation?
• What safeguards or design principles do we have in place to ensure AI enhances, rather than erodes, human energy, creativity and capacity for deep thinking?
• How are leaders being equipped to recognise and respond to new forms of burnout emerging from AI-enabled work patterns?

Dilemma #3: Loss of judgement skills

There’s growing evidence that AI use can affect our judgement skills. 

One study, cited by Zivit Inbar FCPHR, founder and CEO of DifferenThinking, followed medical specialists performing colonoscopies. It found that, over time, those assisted by AI became less skilled at identifying cancer indicators themselves, as they became too reliant on the technology. 

It’s worth noting that the researchers highlight the limits of these outcomes due to the observational nature of their study, and call for further research to be conducted in this area.

“These findings point to a critical issue: when we rely too heavily on AI, we risk weakening the very judgment and expertise that make human work valuable,” says Inbar, who is also the  facilitator of AHRI’s two new AI courses – one on what HR needs to know to stay compliant and another on embedding responsible AI.

“Think about how technology has already changed us. Twenty years ago, before smartphones, I could remember everyone’s phone number. Now, I don’t even know my own children’s numbers because we’ve become so reliant on devices to store information for us.

“The same risk applies to AI. The more we depend on it, the more skills we risk losing – not just cognitive ones, but judgement-based ones too.”

Hear more from Dr Inbar in AI governance dilemmas in part two of this article.

HR thought starters for this dilemma:

• How do we preserve and strengthen human judgment, expertise and ethical reasoning in an environment increasingly shaped by machine recommendations?
• What checks and escalation pathways exist to support employees to question or override AI-generated decisions appropriately – and are people genuinely empowered to use them?
• How are we monitoring for cognitive or skill atrophy caused by over-reliance on AI tools, particularly in roles where professional judgment is critical?

Navigating the messy middle and risks of AI

Ultimately, navigating the “messy middle” of AI isn’t about finding a final destination, but about HR becoming the architect of a new, more intentional workplace. 

As we move from asking what AI can do to what it should do, the goal is to ensure that in our race for digital efficiency, we don’t accidentally automate away the very human judgment and connection that give our organisations their competitive edge.

A longer version of this article was originally published in the Dec/Jan 2026 edition of AHRI’s HRM Magazine. Read part two of this article here.