Can interactions with AI change climate opinion? Combating climate change requires citizens to recognise its urgency, make sustainable lifestyle choices, and advocate for systemic changes. As AI is becoming more prevalent, we test the extent to which its use can change climate opinions and policy preferences. We argue that AI – and in particular its personalised content - can influence climate attitudes and overcome some of the limitations of pure information provision. In this pre-analysis plan, we hypothesise that engaging with large language models (LLMs) has the potential to shift public attitudes towards climate change by delivering tailored, personalised messages at a scale. We will conduct a survey experiment in the UK to examine how dialogues with a large language model (LLM) affect respondents' climate opinions and policy preferences. These dialogues will leverage previously collected individual-level information to tailor interactions to each respondent. We will compare the effects of these personalised interactions on climate attitudes with those of non-personalised interactions and standard informational statements. Overall, our study aims to contribute to the study of climate opinion and of the consequences of AI for politics.