How AI works
For those who want to know more about how AI works and the underlying technology, here is a brief overview.
Generative AI, such as ChatGPT, uses deep learning techniques to generate text based on prompts. Through a process called training or learning, AI models are exposed to large amounts of text data, allowing them to learn the patterns, grammar and context of the language. The models learn which words, phrases, sentences and pieces of information are most often connected.
When given a specific task or question, the AI model generates an answer by utilising its learned knowledge and generating coherent and contextually relevant text. Taking the example of literature searches, AI tools use natural language processing (NLP) techniques to analyse and 'understand' text. The models can scan many scientific articles and journals using algorithms that take into account factors such as keywords, topic relevance and citation networks. They learn the context in which the words belong. By applying machine learning (training) algorithms, these tools can identify relevant sources, saving valuable time and effort compared to manually sifting through literature.
Paraphrasing and summarising tools, powered by AI, use NLP techniques to 'understand' and even rephrase text. They can assess the meaning of sentences, understand the relationships between words, and generate alternative expressions while maintaining the original intent. These tools can help students avoid plagiarism by helping to paraphrase complex ideas or summarise long passages in a concise way.
Comparing scientific articles and journals can be facilitated by AI tools that use semantic analysis and text similarity algorithms. By evaluating the content and context of different sources, these tools can highlight similarities, differences and connections between different texts.
It is important to understand that while AI tools offer great benefits, their use should be accompanied by critical evaluation and human oversight. In particular, as teachers and trainers, we should be careful to ensure that the generated content complies with ethical guidelines, academic integrity and, for example, the European Commission's guidelines for the use of AI in education. The same approach is important for students to understand.