MIT Technology Review: Why Calling AI Agents "Colleagues" Is a Mistake
Companies are giving AI agents names such as "Alex" and "Maya" and introducing them to staff as "colleagues." MIT Technology Review warns that this creates false expectations. People start projecting human qualities onto AI (responsibility, judgment, empathy) that these systems do not have. In medicine, law and finance, this is no longer a theoretical problem.
AI-processed from MIT Technology Review; edited by Hamidun News
Large companies have started giving AI agents names such as "Alex," "Maya" and "Sam," and introducing them to employees as "new colleagues." MIT Technology Review looks at why this move changes how people treat these tools, and not always for the better.
What it looks like in practice
Imagine coming in to work and learning that you have a new subordinate named Alex. Alex is not a person but an AI tool owned by your company. Yet management deliberately presents it that way: "a new team member," "your digital assistant," "a colleague you can assign tasks to."
This scenario is no longer science fiction. Large corporations around the world have embraced anthropomorphizing AI tools. Research shows that employees accept change much more readily when a system has a name and at least the appearance of a personality. "Automation System v2.3" triggers anxiety and a sense of being replaced. "Alex" feels familiar and neutral.
The phenomenon already has its own vocabulary in corporate settings: "agents." That is what companies such as Microsoft, Salesforce and dozens of others call their AI systems, deliberately choosing a term that stresses autonomy. The next step is names. After names come "workstations" and "roles."
The problem with expectations
Names never arrive alone; assumptions come with them. When an employee hears the word "colleague," they automatically project human qualities onto the AI:
- responsibility for the outcome and its consequences
- the ability to explain its behavior in human terms
- empathy for those affected by its decisions
- the ability to handle ambiguous situations
- moral judgment in conflict cases
But an AI agent does not have these qualities in the sense we expect from a human. It can give plausible answers and imitate understanding, but it does not bear responsibility for consequences, does not understand context at the level of social norms, and is not capable of real moral judgment. When something goes wrong, there is no one to hold accountable: an AI is not subject to disciplinary action.
Marketing as a management tool
MIT Technology Review stresses that when a company gives an AI agent a human name and calls it a "colleague," this is not a technical description. It is a management decision. The goal is to reduce resistance to rollout, ease fear of automation and make the technology seem less threatening.
In the short term, this works very well. The long-term consequences raise questions. Employees who see AI as a full partner start delegating judgment to it in situations where that is dangerous. This is especially noticeable in high-stakes fields: medicine, law and financial advice. An AI error here is no longer a technical glitch but the result of misunderstanding the nature of the tool.
"Imagine coming in to work to learn that a new underling will report to you.
The worker is not a person but an AI tool" — MIT Technology Review sees the core of the problem in precisely this paradox.
What this means
The trend toward anthropomorphizing AI agents will keep growing: companies have found it an effective change-management tool. For employees, it is important to stay clear-eyed: a name and a "personality" do not change the nature of the system. "Alex" is an AI with specific capabilities and specific limitations, not a colleague with judgment and accountability.
Understanding this is not skepticism toward technology. It is professional literacy in the age of AI.
- Nov 21, 2026Microsoft outage: nearly 83,000 Outlook complaints on Downdetector, Teams and Copilot down too
- Nov 21, 2026MLPerf Training Adds Nvidia Rubin and TPU Ironwood in First Appearance: 153 Systems
- Nov 21, 2026Microsoft Turns Copilot Into an "Operating System for Work": Word and Excel Become Agent Plugins
- Nov 21, 2026Google to split Gemini chat history from model training, so you can opt out without losing your history
Key Questions Answered
What it looks like in practice?
Imagine coming in to work and learning that you have a new subordinate named Alex. Alex is not a person but an AI tool owned by your company. Yet management deliberately presents it that way: "a new team member," "your digital assistant," "a colleague you can assign tasks to." This scenario is no longer science fiction. Large corporations around the world have embraced anthropomorphizing AI tools.
What are the key details on The problem with expectations?
Names never arrive alone; assumptions come with them. When an employee hears the word "colleague," they automatically project human qualities onto the AI: responsibility for the outcome and its consequences the ability to explain its behavior in human terms empathy for those affected by its decisions the ability to handle ambiguous situations moral judgment in conflict cases But an AI agent does not have these qualities in the sense we expect from a human.
What are the key details on Marketing as a management tool?
MIT Technology Review stresses that when a company gives an AI agent a human name and calls it a "colleague," this is not a technical description. It is a management decision. The goal is to reduce resistance to rollout, ease fear of automation and make the technology seem less threatening. In the short term, this works very well. The long-term consequences raise questions. Employees who see AI as a full partner start delegating judgment to it in situations where that is dangerous.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.