For many employees, asking AI for help has become as routine as opening an inbox. A difficult email, a half-formed idea or a problem that needs another perspective can now be handed to a chatbot within seconds, raising a new question about what happens to the human conversations that once shaped working life.
A workplace study by the Human Connection Workplace Index found that 76% of employees use AI tools for professional advice, while 52% use them for brainstorming and 37% say they use AI for what the research describes as ‘workplace companionship.’
Dr Cindy Friedman-van der Westhuizen, a lecturer in AI Ethics at Stellenbosch Business School Executive Development, says AI’s convenience can make it particularly easy for employees to turn to software before seeking another person’s perspective.
‘Relying on AI as your default collaborator is very tempting, and many do rely on AI for once highly collaborative tasks such as brainstorming and generating ideas. These might seem like small interactions, but they’re the moments that have always made workplaces work,’ she says.
The appeal is straightforward, she adds. AI is available immediately, can process information quickly and does not appear to judge the person using it.
‘One of AI’s greatest strengths is that it’s always available, incredibly fast and instant with its reply. AI responds with a complete lack of judgment and doesn’t mind if you ask it to review a document five times, brainstorm basic ideas, explain a concept or help you draft a difficult email.’
That accessibility, however, can alter how employees interact with the people around them.
‘Every time you ask someone for a second opinion, challenge an idea together or work through a problem, you’re doing more than completing a task. You’re building trust, learning how other people think and strengthening relationships.’
Friedman-van der Westhuizen argues that the shift is not simply about whether AI can complete a task. It is also about what may be lost when an interaction that once happened between colleagues is moved entirely into a digital environment.
‘AI is becoming the first place we go when we need to think something through or need to think through a problem,’ says Friedman-van der Westhuizen. ‘That’s enormously useful. But it’s also replacing the five-minute desk-side conversations where ideas are tested, problems are shared, and relationships are built. The typical water cooler moments are now being replaced with a ‘Hi Claude, please help me, I’m stuck’.’
The problem with an always-helpful collaborator & why AI can feel easier to trust
Human colleagues can bring different experiences, perspectives and forms of disagreement to a discussion. AI, by contrast, is designed to respond helpfully within the parameters of a user’s request.
‘AI doesn’t become impatient, and it doesn’t disagree based on its own lived experience,’ says Dr Friedman-van der Westhuizen.
‘It functions less like a true collaborator and more like a very sophisticated yes-man. It may help refine your thinking or fill in gaps, but it is unlikely to challenge the premises you start with fundamentally.’
That distinction becomes particularly relevant when AI is used for matters that extend beyond routine administration.
‘While work is a site of production, it is a critical hub for human social interaction. Relying on an algorithm to generate ideas, navigate sensitive workplace politics or plan for a difficult conversation strips away the vital micro-interactions that build corporate cultures.’
The concern is not that employees are using AI, but that repeated reliance on it could reduce opportunities to practise interpersonal skills that are developed through everyday workplace exchanges.
‘If our first instinct is always to ask AI instead of another person, we may gradually lose the everyday opportunities to practice negotiation, persuasion, constructive disagreement and collaborative problem-solving,’ warns Dr Friedman-van der Westhuizen.
‘Those interactions create social capital inside organisations. If AI starts replacing them, the impact may only become visible years later.’
Another part of the shift is psychological, as AI systems can communicate in a fluent, conversational manner and respond in ways that mirror the emotional tone of a user. That can make an interaction feel more personal than the technology itself suggests.
‘AI models speak with human fluency and mirror emotional registers, making it incredibly easy for users to mistake AI encounters for genuine human empathy. This powerful illusion is driving employees to navigate sensitive workplace politics and personal anxieties entirely through software.
‘Employees find it easier to trust an unbiased algorithm with their raw, unpolished ideas and corporate anxieties than to risk showing vulnerability to a human colleague or manager.’
Research cited in the supplied material found that 57% of global employees actively hide their AI use, while almost half acknowledged using AI in ways that breach company policies, including uploading sensitive company information to free public AI platforms.
The same research found that 66% of employees relied on AI-generated information without evaluating its accuracy, and 56% said they had made mistakes at work because of AI.
Only 47% said they had received AI training, while 40% reported that their workplace had a policy or guidance on generative AI use.
The case for AI with boundaries
Friedman-van der Westhuizen does not advocate abandoning AI. Instead, she argues that organisations need to be more deliberate about where the technology belongs.
‘The efficiency gains of AI are real. AI can be beneficial to remove repetitive administrative work, summarise information and improve productivity. However, organisations must recognise that human collaboration is not just about getting work done but about maintaining the social and intellectual skills required to work together.’
That distinction could mean treating certain workplace activities as fundamentally human.
‘There needs to be an awareness that not every task that AI can do is one it should do. Mentoring, collective problem-solving, coaching, conflict resolution and brainstorming should be protected as human spaces where the value lies in the human interaction, not only the output.’
For organisations, that could translate into clearer rules around approved AI applications, stronger employee training and deliberate opportunities for people to work together without relying on automated assistance.
‘AI should be used as a tool, not as a partner. Organisations need to make it clear that AI can never be a substitute for human engagement. If AI is used for administrative purposes, then in theory that should free up cognitive and social bandwidth for the richer, messier work of human collaboration.
‘To avoid being deskilled by technology, organisations need to invest in developing certain skills intentionally, rather than assuming they will develop organically. This could look like meetings and workshops that are explicitly AI-free, designed to practise negotiation, disagreement and creative problem-solving together.’
Six ways workplaces can set boundaries
The guidance outlined in the supplied material draws on the human-centred principles of the African Union’s Continental Artificial Intelligence Strategy and adapts them to workplace use.
1. Establish a clear purpose for AI
Businesses should identify the problem a particular AI system is intended to solve before introducing it into workplace processes. Adoption should have a defined purpose rather than simply following the availability of new technology.
2. Set boundaries
Organisations should establish which AI applications are permitted, restricted or prohibited. Activities such as mentoring, peer feedback, conflict resolution and collective brainstorming can be explicitly protected as human-led processes.
3. Explain AI’s limitations
Employees need to understand not only what AI can do, but also where its limitations lie. This includes recognising that fluent responses do not necessarily represent genuine understanding, lived experience or independent challenge.
4. Give someone responsibility for AI governance
AI policies need practical oversight. The proposed approach includes establishing an AI governance committee or assigning specific people to monitor, approve and periodically review how AI is being used.
5. Build broader AI literacy
Training should cover more than prompts and productivity. Employees should also understand risks such as overreliance, anthropomorphism, inaccurate outputs and the potential erosion of collaborative skills.
6. Make it possible to raise concerns
Employees should have clear channels for reporting problematic AI outputs and raising ethical concerns, including instances where technology appears to be displacing meaningful human interaction.
As businesses continue integrating AI into their workflows, the challenge identified by Friedman-van der Westhuizen is to retain those human interactions while using automation where it can genuinely improve productivity.
The technology may be able to provide an answer in seconds. What remains less easily automated is the workplace relationship built when two people sit down, challenge an idea, share experience and work out the answer together, or, in this case, simply have someone to vent to over coffee.
Be the first to know – Join our WhatsApp Channel for content worth tapping into! Click here to join!
Also read:
Picture: Pexels
