Guest author: Javad Taheri
In this post, I briefly discuss contemporary approaches to Artificial Intelligence (AI), metaphor, and ethics, and how AI can enhance ethical systems and offer innovative solutions to the moral challenges faced by today’s society, although we must also be aware of the shortcomings of AI.
Artificial Intelligence
Artificial Intelligence refers to the “simulation of human intelligence in machines.” A machine, in this sense, is programmed to function similarly to human cognition, such as learning, problem-solving, and pattern recognition. AI can be broadly categorized into cognitive intelligence and practical intelligence. Cognitive intelligence involves the ability to think, learn, and understand, while practical intelligence includes skills like performing actions wisely and communicating effectively. Generative AI falls into the latter category, as it generates text as well as images, audio, and videos. For instance, ChatGPT, developed by OpenAI, generates linguistic content and interacts with humans through language.
AI’s development has advanced from basic rule-based systems to more complex models capable of learning from data and enhancing their performance over time. These advancements have enabled AI to manage tasks that require a level of understanding and flexibility once thought to be exclusive to humans. For example, AI’s ability to analyze vast amounts of data and identify patterns allows it to assist humans in medical, financial, social, artistic, and communicative activities.
Metaphor
The contemporary notion of metaphor is defined as understanding or reason about one domain (target domain) by using the framework of another domain (source domain). Metaphors are characteristic of both cognition and language, indicating that metaphorical mappings provide new interpretations and showcase the potential of language to creatively and effectively convey novel ideas. Research indicates that human cognitive systems and language are mutually interactive, with language originating from cognition and, in turn, influencing it. This interaction is exemplified in social contexts where language is more developed, indicating that individuals in these settings exhibit enhanced cognitive abilities compared to those in more isolated communities. This cognitive tool not only aids in communication but also in the generation of new ideas and perspectives.
Common-sense Ethics
Common-sense ethics (CSE) refers to the intuitive and pre-theoretical moral judgments that ordinary people use to navigate everyday situations. Unlike formalized ethical theories, CSE can arise from cultural traditions, social norms, and personal experiences, reflecting a community’s collective wisdom. Common-sense ethical principles are characterized by their practicality, accessibility, and adaptability, allowing individuals to adjust their moral judgments based on the context. One of the key strengths of CSE is their flexibility, rooted in the idea that moral understanding and motivations are accessible to everyone, guided by a natural conscience that helps distinguish right from wrong.
The Concept of Comparison
Comparison is a fundamental process involving at least four elements: the items being compared, the comparer, the similarities between the compared items, and the new knowledge generated from the comparison. It starts with a general understanding of the different sides to be compared, followed by identifying similarities and differences in their conceptual foundations, structures, and meanings. This dialogue between sides leads to the detection of new patterns of thought that transcend the simple sum of their parts. This emergence of new knowledge is due to the metaphorical nature of comparison.
AI as a Metaphorical Enterprise
What is supported here is that human intelligence, in both its cognitive and practical spheres, operates metaphorically. Given that, AI exhibits significant similarities to human intelligence; it can also be understood as metaphorical. This is exemplified by ChatGPT, a generative AI that demonstrates practical intelligence through the generation of linguistic representations and facilitation of communication. ChatGPT engages with an artificial cognitive system, evolving and learning through interactions with humans, and continuously updating its knowledge base. Its capacity to generate creative and novel responses by mapping across different domains underscores the metaphorical nature of AI and highlights its potential to enhance human cognition and problem-solving capabilities.
The Potential of AI in Enhancing Ethical Systems
Common-sense ethical principles play a vital role in maintaining social cohesion and guiding behavior in everyday interactions. Unlike formal ethical principles that require rigorous analysis and justification, they provide an instinctive framework for decision-making. This approach to ethics is particularly appreciated in diverse and multicultural societies, where individuals from different backgrounds must find common ground and navigate moral dilemmas together.
ChatGPT exemplifies how AI can enhance our understanding and application of CSE. As a type of intelligence that learns and develops language models through communication, primarily through linguistic interactions, ChatGPT stores the knowledge gained from these interactions in its memory and continuously modifies this knowledge based on new communications. It employs a method that not only stores and modifies knowledge but also generates new patterns and proposals by exerting what I refer to as metaphorical mappings between domains.
Its ability to learn from a vast array of human interactions allows it to develop a nuanced understanding of CSE. By analyzing the ethical judgments and preferences expressed by users from different cultural and social backgrounds, it can identify features that reflect a broad spectrum of common-sense moral intuitions. This capability renders AI a powerful tool, capable of more effectively exploring new potential moral principles, their detailed aspects, and innovative methods for modifying and enhancing them for greater utility on a global scale. To elucidate, ChatGPT is utilized by individuals from diverse cultures and societies, thereby accumulating a vast repository of knowledge concerning human preferences, likes, dislikes, and common ethical standards. It comprehends what is commonly recognized as ethical and unethical, which can be categorized as Common-Sense Ethical concepts. Additionally, it possesses the capability to compare and analyze the data and knowledge it has acquired.
Overall, ChatGPT and, by extension, AI, can be utilized in present-day scholarship to achieve several objectives:
- Identifying novel common-sense ethical principles on both global and local scales.
- Providing clearer and more comprehensible interpretations of these concepts.
- Modifying them based on new insights and evolving contexts.
- Exploring new possibilities for enhancing our ethical landscape.
- Facilitating the dissemination and education of ethical knowledge to a broader audience.
- Developing ethical guidelines and frameworks that are responsive to the changing needs and values of society.
Possible Risks and Hope for AI’s Ethical Future
While the potential of AI to enhance our ethical systems is promising, it is crucial to acknowledge the significant risks and limitations associated with this technology. The ambitious objectives of AI, such as identifying novel common-sense ethical principles, providing clearer interpretations, and developing ethical guidelines, may not always be fulfilled. AI systems can fail to consider the full diversity of ethical ideas due to algorithmic errors or biases inherent in the data they are trained on. There is also the risk that AI might misinterpret the nuances of human ethical judgments, leading to recommendations that are impractical or even detrimental in real-world contexts. Moreover, the normative principles suggested by AI might not be universally applicable or may lack the depth needed for complex ethical decision-making. A particularly concerning risk is the potential for AI to be manipulated by its designers or stakeholders, intentionally or unintentionally promoting a specific moral framework that does not align with the lived realities of diverse communities. In the worst-case scenario, AI-generated ethical principles could be fundamentally flawed or even immoral, causing harm rather than promoting social good. Given these risks, future research must focus on mitigating these dangers by developing new models of AI, especially generative models, that are more transparent, accountable, and aligned with the ethical complexities of human societies. Ensuring that AI contributes positively to our ethical landscape requires careful consideration and ongoing refinement of the technology to address these potential pitfalls.
In conclusion, while the comparative study of common-sense ethics (CSE) by AI holds the potential to yield significant benefits, it is essential to approach this technology with caution. AI can offer advancements in ethical understanding and application through its precision, efficiency, and scalability. However, the risks of algorithmic biases, misinterpretation, and potential manipulation must not be overlooked. To truly harness AI as a valuable opportunity for society, it is important to treat the results of using AI with great caution and a critical approach, ensuring that its contributions to ethics are carefully evaluated and responsibly applied.
References
“ChatGPT.” Accessed August 2, 2024. https://openai.com/chatgpt/.
“Common Sense Ethics.” Ethics Sage. Accessed August 2, 2024. https://www.ethicssage.com/2016/11/common-sense-ethics.html.
“Generative AI Examples.” Google Cloud. Accessed August 2, 2024. https://cloud.google.com/use-cases/generative-ai.
“Generative AI Is Changing Everything. But What’s Left When the Hype Is Gone?” MIT Technology Review. Accessed August 2, 2024. https://www.technologyreview.com/2022/12/16/1065005/generative-ai-revolution-art/.
Gibbs, Raymond W., and Markus Tendahl. 2011. “Coupling of Metaphoric Cognition and Communication: A Reply to Deirdre Wilson.” Intercultural Pragmatics 8 (4): 601–9. https://doi.org/10.1515/IPRG.2011.027.
“Grammar as a Tool for Overcoming Inter-Religious Misunderstanding: A Philosophical Approach.” 2023. University of Groningen. December 15, 2023. https://www.rug.nl/research/centre-for-religious-studies/research-centres/centre-religion-conflict-globalization/blog/grammar-as-a-tool-for-overcoming-inter-religious-misunderstanding-a-philosophicalapproach.
“How ChatGPT and Our Language Models Are Developed | OpenAI Help Center.” Accessed August 2, 2024. https://help.openai.com/en/articles/7842364-how-chatgpt-and-our-language-models-are-developed.
“How To Compare by Ralph Weber | PDF | Immanuel Kant | Reason.” Accessed August 2, 2024. https://www.scribd.com/document/555340683/How-To-Compare-by-Ralph-Weber.
Lakoff, George. 1980. Metaphors We Live By-Archive. Chicago: University of Chicago Press. http://archive.org/details/metaphorsweliveb0000lako.
“OpenAI.” Accessed August 2, 2024. https://openai.com/.
“The Evolution of AI: From Rule-Based Systems to Machine Learning | by Venny Turner | Medium.” Accessed August 2, 2024. https://medium.com/@vennyturner/the-evolution-of-ai-from-rule-based-systems-to-machine-learning-af66d6ce88eb.
“What Is Generative AI?” Caltech Science Exchange. Accessed August 2, 2024. http://scienceexchange.caltech.edu/topics/artificial-intelligence-research/generative-ai.
About the author
Dr. Javad Taheri is a philosopher of religion who specializes in comparative philosophy of religion with a particular focus on the philosophy of religious language. His scholarly investigations span topics such as interreligious understanding and conflicts, as well as the cognitive science of religion, with a particular focus on cognitive semantics and philosophy, with a keen interest in the contemporary theory of metaphor and ethics. Beyond his research activities, Dr. Taheri also lectures on philosophy, anthropology, and the intellectual history of religions at the University of Groningen, the Netherlands.