Many newspaper articles, TV shows and social media posts discuss climate change and promote many different views. In this talk, I will share some of the ways that AI can be used to understand this space and highlight ways we can recognise the effectiveness of these messages. We’ll see how social media interactions can shed light on levels of polarisation in engagement with climate change media, how the technology behind ChatGPT can reveal the ‘echo’ of attempts to influence attitudes and how we can efficiently identify patterns in thousands of images in media coverage.

The YouTube Regrets project is part of a pioneering set of projects designed to scrutinize and gather evidence on the harmful and dangerous nature of YouTube recommendations. Working with 37,380 users, the team analysed tens of thousands of videos which, though recommended by YouTube’s algorithm, viewers regretted watching or being recommended to. This crowdsourced investigation used Mozilla’s Regrets Reporter, a browser extension for volunteers to report their ‘regrets’. Chico Camargo coordinated a team of 41 student research assistants to analyse these regretted videos, finding that these YouTube recommendations are often harmful and misinformative, and that non-English speaking countries are hit the hardest.

How can we start to capture and communicate the local effects of climate change through images? Images are everywhere. Beyond capturing memories or helping to explain concepts, they can be vivid and emotive tools, drawing us in to a topic. They also play a key role in the media, communicating beyond words and shaping how we think. Yet, despite their importance, photography and imagery don’t tend to get the recognition they deserve. Often relegated to the role of background ‘wallpaper’, images are often given little thought by researchers or journalists. This was a session run by Saffron O’Neill, together with her PhD students Ned Westwood, Sylvia Hayes, Veronica White and Ollie Blewett; alongside WellImage CIC founder, photographer Vanessa Miles. Together, they ran a seminar and guided photography walk through Exeter city centre. They challenged participants to view the city in a new way, and to create photographs which might re-envision how we might cope, adapt, and flourish in a climate-changed future.

C3DS Director Travis Coan, with co-authors Constantine Boussalis (Trinty College Dublin), Mirya Holman (University of Houston), and Stefan Müller (University College Dublin), won Walter Lippmann Best Article of the Year Award for their article on “Gender, candidate emotional expression, and voter reactions during televised debates". This study explores how male and female politicians use their facial expressions and voices to convey emotions to voters during televised debates, while also being influenced by gender roles. They analysed videos of five German debates and found that female politicians, like Angela Merkel and women from minor parties, exhibit less anger but convey emotions similarly to male counterparts. Interestingly, voters reward women's expressions of happiness and penalize their displays of anger. The study was published in the American Political Science Review.