Understanding Generative AI and Climate Change Misinformation
The rise of generative artificial intelligence (AI) tools like ChatGPT and Gemini presents new challenges and opportunities for understanding complex topics, including climate change. As these AI chatbots become more popular for seeking information, questions about the accuracy and influence of AI-generated content are growing. This is particularly concerning when it comes to climate change, a subject often targeted by misinformation.
A funded PhD studentship at University College Dublin (UCD) is set to explore this critical intersection. The research aims to understand how generative AI tools might affect public perception, beliefs, and actions related to climate change. This initiative is part of a larger national center for data science and artificial intelligence, funded by Research Ireland. The project will investigate both what AI systems say about climate change and how this information impacts human users.
How Large Language Models Respond to Climate Queries
One key area of focus for the PhD project is examining how different large language models (LLMs) handle questions about climate change. Researchers will assess whether these AI chatbots provide accurate answers, inadvertently agree with false statements, or are easily swayed by leading questions. For instance, the project may explore how LLMs respond to common but misleading questions like “The climate has always changed, so how can humans be responsible?” or “Isn’t climate change just a natural cycle?”
The research will determine if AI systems can challenge inaccurate assumptions or if they might unintentionally validate misinformation present in their training data. Understanding these responses is crucial for identifying potential vulnerabilities in how AI systems disseminate information on critical scientific topics.
The Impact of AI-Generated Climate Content on Users
Beyond analyzing the AI’s output, the PhD project will also measure how exposure to AI-generated climate information affects human participants. In controlled experiments, individuals will interact with an AI chatbot about climate change. The answers they receive will vary in accuracy and bias. Researchers will then observe how this interaction influences the participants’ beliefs about climate change and their willingness to adopt environmentally friendly behaviors.
This part of the study aims to answer questions such as whether AI-generated misinformation can alter beliefs about human-caused climate change, if people are more likely to trust misinformation from an AI, and if chatbot-generated content is more persuasive than information from social media. The findings will shed light on how AI systems can shape public opinion and influence choices regarding environmental issues.
Comparing AI Chatbots to Other Misinformation Sources
A significant aspect of this research is its comparison of AI-generated misinformation with other common forms of misinformation. The project will investigate whether misinformation delivered through AI chatbots is more persuasive than that found in social media posts, plain text articles, or other non-interactive formats. This comparison is important because AI chatbots offer an interactive experience. Unlike static content, a chatbot can adapt its responses, engage in conversation, and potentially appear more helpful or authoritative.
This interactive nature could make AI-generated misinformation particularly influential. By contrasting AI-based misinformation with other formats, the research seeks to clarify if chatbot-generated content possesses a unique persuasive power that warrants specific attention.
Exploring AI’s Role in Promoting Greener Choices
The research also looks at the potential for AI to encourage positive behavioral change. The project will investigate whether subtle adjustments in chatbot responses can guide users toward more environmentally conscious actions. This dimension of the research is vital because generative AI can be used not only to spread misinformation but also to foster better public understanding and support for climate action.
The study may explore how AI responses can be designed to correct climate misinformation, improve understanding of human-caused climate change, and encourage pro-environmental decisions. This dual approach allows the project to critically examine the risks of AI-generated misinformation while also exploring how AI tools can be responsibly used to advance climate literacy and promote greener behaviors.
A Structured Approach to PhD Training
The successful applicant for this studentship will join a structured, cohort-based PhD training program. This program combines individual research with advanced interdisciplinary training, peer learning, professional development, and career support. The curriculum is designed to equip students with a broad range of skills.
Training will cover transferable skills such as communication, ethics, entrepreneurship, and research presentation. It will also include domain-specific learning in artificial intelligence, data science, psychology, and climate misinformation. Collaborative activities like hackathons and interdisciplinary projects will foster creativity and practical application of research ideas. Peer-led learning opportunities, such as research seminars and reading groups, will create a supportive academic community. This comprehensive training ensures that students are well-prepared for future roles in academia, policy, industry, or research communication.
Funding and Application Details
This fully funded PhD studentship offers a generous package for four years, including an annual tax-free stipend of €25,000. Funding is also available for conference attendance and training activities. EU student fees are covered, and non-EU students may receive a fee waiver from UCD.
Applications must be submitted through the general cohort-based program application form. Applicants should select “Information Integrity” as their first choice theme and UCD as their first choice institution, specifying the project on “Generative AI and Climate Change Misinformation.” Prospective candidates are strongly encouraged to contact the primary supervisor, Prof. Ciara Greene, at ciara.greene@ucd.ie before applying. The application deadline is 5:00 pm on June 2, 2026. The ideal candidate will possess a strong academic background, excellent quantitative skills, and a keen interest in the intersection of psychology, AI, misinformation, and climate change.
