The year 2023 has witnessed increasingly polarised debates on artificial intelligence’s benefits and harms. The worldwide use of AI tools has become the metric for gauging the progress of the human race or its extinction.
AI cuts down employment opportunities for the huge workforce across the globe. This makes it seem like a threat to society because of its ability to influence human decisions, perform activities with calculated precision, and generate answers in a matter of seconds. However, AI has sparked tremendous arguments, and much has been written on the subject, including whether using AI tools such as ChatGPT and Bard can compromise academic integrity and ethics. This has turned out to be a rather grey zone. As artificial intelligence becomes a rather permanent resident in our houses, I find it important to know how they are constructed and what we can do to utilise them without letting AI subdue our individual, human voices.
AI generation tools, just as every other tool, are created by collecting data — a large quantity of language data. This is known as a corpus, and it is an important constituent in the study of both linguistics and computer science. Generating a computer language that would closely mimic natural language has been a long-desired objective of the AI development community, which has been facilitated to a great extent by natural language processing (NLP).
Keeping aside the complexities of the process, it is beneficial to learn that the better functioning of such tools depends on the amount of data they are trained on. This establishes a directly proportional relationship between the amount of data the model is fed and how well it will function in generating the most appropriate results, which is why ChatGPT cannot give us data that is not already present on the World Wide Web.
I like to think of these tools as a much more advanced and sophisticated search engine where you can edit the output data to your requirements. Therefore, I think that rather than accepting or dismissing AI, it is important to learn how to optimise the use of the tools. A complete dependence on them or an outright rejection would deprive us of the opportunity to learn, and in the current scenario of rapidly evolving data and information, we would not want to do that.