Artificial Intelligence and education: Opportunity or risk? Opinion from the conversational AI chatbots

Artificial Intelligence (AI) in education has undergone significant developments in recent years. Tools like ChatGPT are currently used in educational contexts, either as an assistant to teachers generating didactic material, or as a virtual tutor for students, to answer questions and facilitate collaboration. The purpose of this paper is to investigate benefits and risks of the use of AIs based on the opinion of conversational AI chatbots themselves. Three interviews were held with conversational AIs: ChatGPT, Gemini and LLaMA. Twenty categories were identified from the four dimensions developed in the interviews, with 572 total citations. The Learning Resources, Advantages and Learning Benefits categories stand out. ChatGPT is prominent in terms of resources available for learning and teaching, as well as the benefits that teachers and students find in its application, awareness of the risks and challenges they face in education to make appropriate use of large language models (LLMs). On the other hand, Gemini stands out in identifying the use and strategies of appropriate use and ethical considerations that educational agents must consider regarding the use of AI in education. Finally, LLaMA excels in identifying the resources, benefits, advantages and challenges for learning. In conclusion, conversational AIs are aware of both possibilities of their use in education and risks and challenges faced in this field. Therefore, conversational AIs have great potential and impact in education, but it must be up to teachers and students to promote ethical and appropriate use for the experience to be positive and beneficial in education.