WIRED: Stop Naming AI Features After Human Processes
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic unveiled a new feature called "dreaming" at its developer conference in San Francisco, part of its AI agent infrastructure that reviews completed agent activity logs to extract insights and improve future performance.
- Anthropic described the feature in a blog post: "Together, memory and dreaming form a robust memory system for self-improving agents," with memory capturing learnings during work and dreaming refining them between sessions.
- OpenAI joined the anthropomorphic naming trend in 2024 with its first "reasoning" model, described at launch as "designed to spend more time thinking before they respond," alongside numerous startups that market chatbot "memories" as humanlike nuggets of information.
- Anthropic's own constitution states the company discusses Claude "in terms normally reserved for humans (e.g., 'virtue,' 'wisdom')" because it expects Claude's reasoning to draw on human concepts, and the company employs a resident philosopher to shape the bot's values.
- A study published in the AI & Ethics journal found that anthropomorphism "distorts moral judgments about AI, such as those concerning its moral character and status, as well as judgments of responsibility and trust," per the source.
- WIRED columnist Reece Rogers framed the critique around Philip K. Dick's Do Androids Dream of Electric Sheep?, arguing that AI leaders appear "unable, or at least unwilling, to accept the limitations of their own inhuman tools."
Why it matters: Anthropic's own constitution admits to deliberately framing Claude in human terms like 'virtue' and 'wisdom,' and the cited AI & Ethics research links that kind of framing to distorted moral judgments and misplaced trust — meaning the naming convention isn't just marketing flair, it's a documented mechanism that can mislead users about what these tools actually are.
Ask SkimNews
