John Cutler Warns Against Treating AI Agents as Human Colleagues

In a recent LinkedIn post, John Cutler argues that treating AI agents as equivalent to people can encourage a mechanistic view of work and employment. This article reports on Cutler’s post and examines the progression he describes; it is journalistic coverage, not an article written by Cutler.

“Making ‘agents’ ontologically equivalent to humans, and imagining them as anything other than tools, is the most predictably stupid slippery slope I’ve ever seen.”

Cutler’s concern is not simply about the language used to describe artificial intelligence. He objects to framing agents as coworkers or managers, a framing he says may make workplace decisions appear like neutral task allocation. In his view, that metaphor risks obscuring the people and power structures behind those decisions.

From managing agents to being managed by them

Cutler traces what he sees as a progression: people first learn to manage agents, then work alongside them, and eventually encounter agents making decisions about human workers. He writes:

“It starts with ‘humans managing agents,’ then ‘humans working alongside agents,’ then ‘agents managing you on behalf of some manager’ (we all have one), and eventually ‘agents deciding whether humans are worth employing at all.’”

That sequence is Cutler’s warning about how workplace assumptions could shift as organizations adopt AI systems. He connects the argument to the experience of job seekers, saying that “we’re already there.” The post does not detail particular hiring systems or provide evidence about their use; instead, Cutler uses the observation to underscore his concern about automated judgments of workers’ value.

The workplace as a game board

According to Cutler, imagining people and agents as interchangeable players makes it easier to see work as a matter of assigning and optimizing resources. He describes the metaphor as comfortable and convenient for investors, executives, and managers. His criticism is that it can normalize treating human work as a mechanical process.

Technology and existing assumptions

Cutler’s post also makes a broader point about technology: tools do not create every workplace belief from scratch, but they can give existing beliefs new justifications. He concludes:

“The technology doesn’t just enable this worldview. It gives people a new way to rationalize it.”

That distinction is central to his argument. Cutler is not only questioning whether AI agents should be described as humanlike. He is asking readers to consider how those descriptions might shape decisions about management, accountability, and employment. His post presents a caution about the assumptions embedded in workplace technology, rather than a detailed account of any one system or policy.