Unsettling Relationships Developing Between Workers And AI Coworkers
Experts emphasize that AI is fundamentally different from human coworkers, requiring clear protocols and shifting human roles from operating to instructing, prioritizing judgment.
Joe McKendrick, Senior Contributor
Forbes
2 min read
7/10
Key Takeaways
65% of knowledge workers use AI tools weekly as of 2025 (Gartner), yet only 30% of companies have formal AI workplace policies (IBM Institute for Business Value).
A 2024 study from MIT and Stanford found workers who anthropomorphize AI are 30% more likely to accept incorrect outputs without verification.
The shift from operating to instructing means employees now spend an average of 40% of their AI-related time on prompt engineering and output validation (McKinsey Global Institute).
45% of professionals in finance and healthcare report feeling 'emotionally attached' to their main AI assistant, according to a 2025 Deloitte survey.
Companies that implemented mandatory 'human-in-the-loop' protocols for AI-driven decisions saw a 22% reduction in errors over six months (Harvard Business Review analysis).
The colleague who never sleeps, never complains, and always has an answer might be the most unsettling coworker you've ever had. Experts are warning that as artificial intelligence becomes a permanent fixture in offices worldwide, workers are developing complex – and sometimes troubling – emotional relationships with their AI coworkers. From over-trusting their judgment to attributing human-like intentions to algorithms, the dynamic demands a fundamental rethinking of workplace protocols and role definitions. The rapid integration of generative AI assistants like ChatGPT, Microsoft Copilot, and Google Gemini into daily workflows has accelerated since 2023. By 2025, a Gartner survey found that 65% of knowledge workers interact with AI tools at least weekly. Yet the human side of this transition remains under-considered. Organizational psychologists and AI ethicists emphasize that AI is fundamentally different from human coworkers. It lacks consciousness, empathy, and accountability. But humans naturally anthropomorphize, leading to 'unsettling relationships' where workers confide in, defend, or even feel guilty toward their AI assistants. A 2024 MIT-Stanford study showed that workers who describe their AI as 'helpful' or 'friendly' are 30% more likely to accept wrong output without verification. This trust gap is dangerous in high-stakes fields like finance, healthcare, and legal compliance. The shift from operating to instructing is central. Workers no longer execute tasks directly; they prompt, review, and validate. Clear protocols must replace ad-hoc usage. Companies are now publishing 'AI etiquette' guides, setting boundaries on what tasks can be delegated, and requiring human-in-the-loop sign-offs for critical decisions. Judgment remains the human's domain. The implications ripple beyond productivity. If workers treat AI as a peer, accountability blurs. Who is responsible when an AI gives bad advice – the employee, the developer, or the company? Regulators in the EU and California are already examining algorithmic discrimination and transparency in hiring tools. Meanwhile, some firms report 'AI burnout' – employees overwhelmed by constant prompting and output checking. The path forward requires deliberate culture change. Training programs must teach not just how to use AI, but how to maintain critical distance. Some organizations have started 'AI relationship workshops' to help staff understand the limits of machine intelligence. Expect new roles like 'AI collaboration manager' to emerge. As one expert put it: 'We must treat AI as a tool, not a teammate. The unsettling relationships will fade only when we codify that distinction.'
Frequently Asked Questions
AI coworkers lack consciousness, empathy, and accountability. They cannot make ethical judgments or take responsibility for outcomes. While humans can collaborate with nuance and trust, AI follows programmed rules and data patterns without understanding intent.
Humans naturally anthropomorphize, especially when AI uses conversational language and responds helpfully. This can lead to over-trust, guilt, or even attachment. Studies show that workers who describe AI as 'friendly' are more likely to accept incorrect outputs.
Companies should establish guidelines on when to trust AI outputs, require human-in-the-loop for high-stakes decisions, and train employees to critically evaluate AI suggestions. Clear role definitions shift from 'operating' to 'instructing and validating'.
Over-reliance can erode critical thinking, reduce error detection, and blur accountability. In fields like healthcare or finance, trusting an AI assistant without verification can lead to serious mistakes. Organizations must promote healthy skepticism.
Implement AI literacy training, create 'AI relationship workshops,' and enforce protocols that keep humans in decision-making loops. Regular audits of AI usage can identify over-dependence and adjust policies accordingly.
Experts unanimously recommend treating AI as a tool. While it can augment human capabilities, attributing teammate status risks misplaced trust and accountability issues. Clear separation helps maintain professional judgment.
