As artificial intelligence (AI) becomes ever more integrated into our daily lives, an intriguing question arises: Are we imbuing AI with a sense of humanity through the language we use? Words such as ‘think,’ ‘understand,’ and ‘remember’ are typically associated with human cognitive processes, yet these terms are increasingly used to describe AI. This linguistic choice could lead us to anthropomorphize AI systems, falsely attributing human-like qualities to them.
In a recent study, researchers Jo Mackiewicz and Jeanine Aune from Iowa State University delved into this issue. Their work, titled ‘Anthropomorphizing Artificial Intelligence: A Corpus Study of Mental Verbs Used with AI and ChatGPT,’ published in Technical Communication Quarterly, examines how language shapes public perception of AI. The study reveals subtle yet significant ways that our verbal descriptions of AI can skew our understanding of these technologies.
Main Findings
-
Infrequent Use of Mental Verbs: The study found that in written news, terms like ‘AI’ and ‘ChatGPT’ are less frequently paired with mental verbs than in everyday speech. This discrepancy might be due to intentional editorial choices that aim to prevent the attribution of human emotions or thoughts to AI.
-
Significance of Context: When AI is described using verbs such as ‘needs,’ context plays a vital role. Typically, these references are to technical capacities or functions, such as ‘AI needs data,’ rather than implying AI has desires or preferences.
-
Anthropomorphization as a Spectrum: The research highlights that the portrayal of AI can vary widely. While some descriptions can erroneously edge towards assigning human-like qualities to AI, most narratives stop short of suggesting AI possesses genuine thoughts or emotions.
Conclusion and Implications
The findings underscore the critical role language plays in shaping public perceptions of AI. Using mental verbs may unintentionally suggest AI has human-like capabilities, thus distorting how we view and understand these technologies. Accurately representing AI as a sophisticated tool, rather than an independent entity, is crucial. It keeps the focus on the human designers and operators who ultimately guide AI’s application and governance.
Looking ahead, further research could delve into the nuanced impacts of specific language choices on public perceptions of AI. As AI technology progresses, so too must our approach to discussing it, ensuring our language evolves to meet the demands of these complex systems accurately. This study is a powerful reminder of how pivotal word choice can be in shaping public discourse and expectations surrounding AI, highlighting the need for deliberate and mindful communication in the age of AI.