I Am Begging AI Companies to Stop Naming Features After Human Processes (www.wired.com)

Anthropic has introduced a new feature called “Dreaming” within its AI agent infrastructure at a developer conference in San Francisco. This innovative feature allows AI agents to analyze their own activity logs, facilitating self-improvement by identifying patterns from past tasks. The incorporation of “Dreaming” marks a significant development for users who deploy AI agents on complex tasks, as it enhances their capability by refining the agents' memory and performance based on insights gained between sessions. However, the feature's name has sparked a debate within the AI community about the anthropomorphizing of AI technologies. Critics argue that such terminology blurs the line between human cognitive processes and machine functions, potentially leading to misconceptions about AI capabilities and fostering misplaced trust in these tools. By likening AI agent operations to human experiences, such as “memory” and now “dreaming,” companies risk distorting users' perceptions of AI and underestimating its limitations. This ongoing trend raises important ethical questions around how language influences our understanding of AI, emphasizing the need for more precise and distanced terminology in tech marketing.