A Discourse Analysis of Metaphor Use in AI and Machine Learning Texts (DRAFT ONLY)

 

A Discourse Analysis of Metaphor Use in AI and Machine Learning Texts

Janpha Thadphoothon

Abstract

This paper presents a discourse analysis of metaphorical language used in texts related to artificial intelligence (AI) and machine learning (ML). Metaphors play a crucial role in shaping our understanding and communication of complex concepts, and their presence in AI and machine learning discourse has significant implications for both technical and non-technical audiences. By employing a discourse analysis framework, this study examines the metaphors employed in AI and machine learning texts, identifies their underlying conceptual mappings, and analyzes their impact on knowledge construction and interpretation. The findings contribute to a deeper understanding of how metaphors shape the discourse surrounding AI and machine learning, shedding light on the cognitive and communicative aspects of these technological domains. This research aims to enhance our comprehension of the metaphorical language used in AI and machine learning texts, offering insights into the socio-technical implications and facilitating effective communication between experts and non-experts in the field.

Keywords: discourse analysis, metaphor, artificial intelligence, machine learning, communication, knowledge construction

Introduction

The language we use to discuss artificial intelligence (AI) and machine learning has a profound influence on how we conceptualize and communicate these complex fields. Metaphorical language, in particular, plays a significant role in shaping our understanding and interpretation of AI and machine learning concepts. Metaphors help bridge the gap between abstract technical concepts and our everyday experiences, enabling us to grasp and convey complex ideas more effectively. By examining the metaphors used in AI and machine learning discourse, we can uncover the underlying conceptual mappings and shed light on the cognitive and communicative dimensions of these domains.

However, it is generally accepted that metaphors can play an important role in helping people understand complex concepts, including those related to AI. Metaphors can bridge the gap between abstract technical concepts and everyday experiences, enabling people to grasp and convey complex ideas more effectively. By examining the metaphors used in AI discourse, people can uncover the underlying conceptual mappings and gain a deeper understanding of the cognitive and communicative dimensions of AI. So, in this context, a human might say that understanding the metaphors used in AI could help people better understand how AI works.

Aim

The aim of this research is to investigate the use of metaphors in AI and ML texts.
Definitions

Text: anything
Artificial intelligence: smartass

Limitations

Literature Review

This section provides an overview of previous research on metaphorical language in the context of AI and machine learning. It explores the theoretical foundations of metaphor and its relevance to communication and knowledge construction. Existing studies on metaphor analysis in related fields, such as computer science, linguistics, and cognitive science, are reviewed to highlight the significance of metaphor in AI and machine learning discourse. The literature review also addresses the socio-technical implications of metaphor use, including its role in shaping public perceptions, ethical considerations, and implications for human-machine interaction.


Types of Metaphors

Textual metaphors and visual metaphors are two different forms of metaphorical expressions used in communication. While both serve the purpose of conveying abstract or complex concepts in a more relatable and understandable way, there are some key differences between them:

1. Medium of Expression:

Textual metaphors are primarily expressed through written or spoken language. They rely on words and linguistic devices to create comparisons between two different domains. These metaphors are often found in literature, poetry, speeches, and everyday language.

Visual metaphors, on the other hand, are expressed through visual imagery, such as photographs, drawings, paintings, or other visual representations. They use visual elements, symbols, and compositions to create associations between two different concepts or domains. Visual metaphors are commonly used in art, advertising, design, and visual communication.

2. Mode of Perception:

Textual metaphors primarily engage the auditory or linguistic senses. They rely on the interpretation of words and language structures, requiring the audience to process and understand the metaphor through reading or listening.

Visual metaphors primarily engage the visual senses. They rely on the interpretation of visual elements and the recognition of visual patterns, shapes, colors, and compositions. Visual metaphors are perceived through observation and visual analysis.

3. Representation and Expression:

Textual metaphors rely on language and linguistic devices, such as similes, analogies, or symbolic language, to represent and express the metaphorical comparison. The metaphorical meaning is conveyed through the use of words, sentence structures, and figurative language.

Visual metaphors rely on visual elements, such as images, symbols, icons, or visual compositions, to represent and express the metaphorical comparison. The metaphorical meaning is conveyed through the visual arrangement and interaction of elements.

4. Flexibility and Interpretation:

Textual metaphors allow for more flexibility and interpretation due to the use of language. Different individuals may interpret textual metaphors in slightly different ways based on their personal experiences, cultural background, and linguistic nuances. The meaning of textual metaphors can be subjective and open to interpretation.

Visual metaphors also allow for interpretation but may have a more immediate and direct impact due to the visual nature of the expression. Visual metaphors can evoke emotions, associations, and intuitive understandings more readily, often bypassing the need for explicit linguistic interpretation.

In summary, textual metaphors rely on written or spoken language to create comparisons between two domains, while visual metaphors use visual imagery and elements to achieve the same purpose. Both forms of metaphors have their unique strengths and characteristics, engaging different senses and modes of perception to convey metaphorical meaning.

Here are some more examples of metaphors that are commonly used in the field of computer science:

“Computer virus” - This metaphor compares malicious software to a biological virus, with the ability to infect and spread from one computer to another.

“Data mining” - This metaphor compares the process of extracting useful information from large datasets to the process of mining for valuable resources such as gold or diamonds.

“Firewall” - This metaphor compares a network security system to a physical wall that prevents the spread of fire, with the ability to block unauthorized access while allowing authorized communication.

“Cloud computing” - This metaphor compares the delivery of computing services over the internet to the delivery of water or electricity through a utility network, with data and applications stored and accessed remotely rather than on local devices.

Methodology

The research methodology employed in this study is discourse analysis, specifically focusing on the analysis of metaphorical language in AI and machine learning texts. A corpus of relevant texts, including research papers, technical documentation, and popular media articles will be collected and analyzed using qualitative and quantitative methods. The analysis will involve identifying and categorizing metaphors, uncovering their underlying conceptual mappings, and examining their implications for knowledge construction and interpretation. The study will draw on existing frameworks and theories of metaphor analysis, such as Lakoff and Johnson’s Conceptual Metaphor Theory to guide the analytical process.

Results and Discussion

This section presents the findings of the discourse analysis highlighting the metaphors identified in the AI

Sure! Here are some more examples of metaphors that are commonly used in the field of AI:

- "Artificial intelligence" - This metaphor compares the intelligence of machines to that of humans, with the implication that machines can exhibit human-like cognitive abilities.

- "Machine learning" - This metaphor compares the process of improving the performance of an AI system to the process of human learning, with the ability to acquire new knowledge and skills through experience.

- "Neural network" - This metaphor compares the structure and function of an artificial neural network to that of the human brain, with interconnected nodes and pathways that can process and transmit information.

- "Deep learning" - This metaphor compares the structure and function of deep neural networks to the process of human learning and understanding, with multiple layers of interconnected nodes that can extract increasingly complex features from data.

I hope these additional examples of metaphors used in AI are helpful!