How to talk about "AI" without adding to the anthropomorphization (buttondown.com)

In a recent op-ed for Tech Policy Press, researchers Emily M. Bender and Nanna Inie argue against the anthropomorphizing language often used to discuss "AI" technologies. They highlight how such language complicates understanding and discussion around these systems, recommending a shift towards terminology that reflects actual functionalities rather than human-like qualities. Key suggestions include replacing terms like "artificial intelligence" with "probabilistic automation" and avoiding phrases that suggest emotional agency when discussing software operations. This approach emphasizes the human role in shaping technology, encouraging clearer communication about what these systems can and cannot do. The significance of this shift lies in fostering a more precise understanding of AI and machine learning technologies, ultimately benefiting both developers and users. By categorizing and identifying anthropomorphizing language, the authors hope to create a new standard that encourages thoughtfulness in discussions surrounding these tools. This effort not only better represents the reality of algorithmic processes but also promotes critical engagement with technology, urging conversations to focus on the true nature of the systems rather than exaggerated representations.