The Limits of Language and the Search for Understanding in Artificial Intelligence

Book excerpt: The development of artificial intelligence (AI) has led to remarkable advances in natural language processing (NLP), enabling machines to process human language with increasing sophistication. While this progress holds extraordinary promise for various applications, it also raises profound philosophical questions about meaning, sentience and understanding within language games. This paper delves into the intricate relationship and interplay between AI, language and meaning. It explores some of the philosophical underpinnings of language, examining how AI systems can extract, translate and manipulate semantically sensible content. It also investigates a few of the challenges of developing AI systems with the ability to ‘understand’ meaningful language that goes beyond surface semantic and syntactic proficiency, algorithmic intelligence and the probabilistic semantic route finding used by Large Language Models (LLMs) with their reliance on large data sets. The paper addresses the wider limits of logic and language for humans as well as for digital intelligence. By examining some of the philosophical and practical dimensions of meaning in AI NLP and LLM, this paper aims to foster a deeper understanding of the challenges and opportunities that lie ahead in this rapidly evolving field. It seeks to promote informed discussions about AI language models, ensuring that these powerful tools are used to improve human understanding and communication. The paper seeks to encourage greater humility in how Homo sapiens define and approach the concept of intelligence. The author deprecates our historic excessive interspecies exceptionalism. The author makes no claims of original thoughts or research in the fields of philosophy of language, linguistics or the development of more generally applicable AI. The paper is intended to help specify the key issues using ordinary human readable language and to understand some of the main conceptual issues involved in the development of artificial general intelligence (AGI). Images stated as being by the author have been created using generative AI image creation tools.