AbstractArtificial Intelligence (AI) has a significant impact on society today and it is likely to have an even greater influence on education in the future. The introduction of new technology in schools affects both students and teachers, as they must handle and consider yet another aspect of their profession. Therefore, it is valuable for teachers to gain more understanding andknowledge about AI technology. The study aims to highlights large language models such as ChatGPT and Co-pilot, to use them as pedagogical tools in teaching to develop student´s criticalthinking related to generative AI in lower secondary school. The study´s ontological approach combines Lesson Study (LS) and Learnings Study, which have their origins in phenomenography. Variation theory within Learning Study is used to interpret the study´s phenomena. The study´s epistemological approach describes the process of knowledge acquisition. The study process has been abductive, starting in theory, approaching empirical data through observations and tests, and returning to theory to find explanations. The results are based on LS in classes 9A and 9B, filming, observations, recorded conversations with teachers and pre- and post-test on students. Based on variation theory´s learning object, critical aspect, variation, and generalization the empirical data is interpreted with support from previous research. The purpose is to use LS to highlight teachers´ didactic choices, how their teachingdevelops and is affected, and the impact of LS on students´ critical thinking when chatbots areused in teaching. In current LS teachers gain insight into students´ critical thinking and understanding of generative AI, including both students´(un)critical attitudes and theirincreasedverification ofsource awareness regarding chatbots, which also was shown by the pre- and posttests. Some students demonstrated examples of critical thinking, and many of the students reflected on what a source- critical approach meant when using chatbots. However, it also emerged that some students struggle to understand the concepts of objectivity and emotionally charged words. Several students in each class did not question the chatbot´s answers and found it difficult to maintain a critical stance towards its responses. Many students understood that words can be emotionally charged, which creates difficulties when the chatbot´s text contains many complex words. Therefore, it is advantageous to explore chatbots critically together,teacher and students. By developing knowledge and understanding of generative AI together and being aware of potential risks, teachers with different subject expertise can develop chatbots into a complement in their teaching.