We stand at the nexus of an information glut and an action deficit — particularly when it comes to defining the crisis of our time: climate change. Despite an excessive amount of data, a disconnect persists. According to the Yale Program on Climate Change Communication, 72% of Americans claim climate change stokes their anxieties, yet only 33% actually bring it up in conversations, even with those closest to them. Why this incongruity between apprehension and dialogue? It’s not just the message that counts, but how it’s delivered.
Just this year, the Intergovernmental Panel on Climate Change (IPCC) published its latest report — a document describing the state of scientific, technical, and socio-economic knowledge on climate change, its impacts and future risks. Crafted by hundreds of experts, and ratified by over 200 nations, the full report is sobering, unflinching, and, at 115 pages, overwhelming. Nonetheless, journalists can receive just a 24-hour embargo to distill this information into a comprehensive yet digestible read.
To directly tackle this communication challenge, we initiated a specialized test case ChatGPT which offers journalists facing tight deadlines the capability to rapidly analyze expansive data sets. For this exercise, we input a large volume of data extracted from the latest IPCC report. Then, we prompted the robot to “review and analyze the following data, then concisely give us the main points.” We repeated this process to ensure we gave the AI a broader contextual view of the entire report. To assess the effectiveness of the AI tool in data interpretation, we compared its output with the initial stories on the IPCC report published by national news sources including The Washington Post, New York Times, Los Angeles Times, and Wall Street Journal. Our objective was to gauge the analytical accuracy and narrative coherence that AI tools could potentially bring to journalistic coverage.
While journalistic renditions did an incredible job prioritizing urgency and accessibility — backing these sentiments with accurate data — they sometimes glossed over nuances, particularly the impacts of climate change on human health and social equity emphasized in the report, typically for the sake of simplicity and brevity. However, we found that AI could quickly link these technical details without omitting subtleties — like how climate change disproportionately impacts different communities and other specific health impacts.
For example, in their coverage of the recent IPCC report, prominent national publications like The New York Times, The Wall Street Journal, and Washington Post, were unequivocal in asserting that human activities have become the main driver of climate change, impacting everything from the ocean to the atmosphere. However, there was a divergence in depth. While some publications like The New York Times connected climate data to food and water security, others either briefly touched on this or omitted it altogether. In our parallel experiment, the AI tool demonstrated an ability to connect the dots on another level, effectively articulating how food and water security are not standalone issues; they’re components of a complicated web that has pervasive repercussions on our daily lives. It not only extended this analysis to encompass issues like urban infrastructure but also included both physical and mental health impacts, as emphasized in the full IPCC report and summary.
To effectively tell the story of climate change, we need to simply tell the comprehensive story of climate change-– with all of its possible implications for our futures. AI can help journalists cut through a dense fog of information and distill insights, enabling broader contextualization for storytellers to explore, ultimately magnifying the potential for storytelling.
It is critical to acknowledge that AI does not possess the ethical discernment seasoned journalists have honed after years of practice. Moreover, even though AI can sift and decipher enormous datasets leading to helpful analysis, the model cannot inherently understand the social, political, or cultural nuances that are often critical to great journalism. Consequently, misinformation or unverified AI-generated content could sow confusion, dilute the urgency of the issue, or even reassert harmful ideologies that have inhibited action toward climate change to begin with. While AI should be a tool used cautiously to enhance human storytelling and journalistic endeavors, we believe it holds great potential in supporting journalists navigating large data sets related to climate change and beyond. Ultimately, the true measure of its impacts on climate journalism will not be counted by clicks or shares, but in the depths of the conversations it ignites and the meaningful impacts those conversations inspire.
Allison Agsten leads USC Annenberg’s Center for Climate Journalism and Communication, leveraging her diverse experience from CNN and LACMA to shape the future of climate communication. She pioneers art-focused climate discussions as the first curator of the USC Wrigley Institute for Environmental Studies. A UCLA and Harvard graduate, Agsten masterfully merges journalism, arts, and public engagement to tackle climate discourse.
Michael Kittilson is a first-year graduate student at USC Annenberg studying public relations and advertising, Kittilson aspires to help solve the world’s toughest messaging and communication problems. His background spans 5 years in various roles that intersect strategic communications, tech, and policy, including work with the U.S. Department of State and national media organizations.