Science Feedback, QuotaClimat, and Data for Good Launch Climate Safeguard, an AI Tool to Detect Climate Misinformation on TV and Radio
Jan. 6, 2025 – Science Feedback, in partnership with QuotaClimat, and Data for Good, has been awarded a grant from JournalismAI Innovation Challenge, supported by the Google News Initiative (GNI), to develop Climate Safeguards—an AI-powered system that tracks and analyzes climate misinformation on TV and radio. This initiative will equip journalists, media organizations, and regulators with real-time alerts and analysis to monitor misleading climate narratives and respond with accurate, science-based analysis.
Despite TV’s vast audience and influence, there is currently no systematic way to monitor climate-related claims aired on broadcast media. Talk shows and political debates frequently feature misleading climate arguments with little or no correction. Several examples illustrate this problem:
- On August 8, 2023, the program Punchline été on CNews allowed a guest to claim, without contradiction, that “climate change is a lie, a scam.” This led to the first-ever financial penalty issued by ARCOM for climate misinformation—€20,000—on July 10, 2024, marking a global precedent.
- On December 7, 2023, March 25, 2024, and May 2, 2024, Bercoff dans tous ses états on Sud Radio aired multiple misleading claims contradicting or downplaying the scientific consensus on climate change. ARCOM issued two warnings to the broadcaster, ruling that these segments violated journalistic accuracy standards.
While efforts to counter misinformation have largely focused on social media, misleading claims on TV and radio have gone mostly unchecked. With this initiative, Climate Safeguards will bridge this gap by using AI-powered tools to scan broadcast content at scale, identify misleading claims, and provide data-driven insights to strengthen accurate reporting.
How Climate Safeguards Works
Led by Science Feedback in collaboration with QuotaClimat and Data for Good, Climate Safeguards will use AI technology to:
- Monitor TV and radio transcripts for possible cases of climate misinformation.
- Analyze how much misinformation exists across different programs.
- Provide fact-checkers and regulators with real-time alerts and reports.
- Publish findings to help inform the public and media organizations.
While AI plays a key role in detecting misinformation, interpreting, contextualizing, and analyzing the results will rely exclusively on human expertise.
The Climate Safeguard project will contribute to the research efforts of the Observatoire des Médias sur l’Écologie, specifically for its work on climate misinformation in French media.
This initiative serves two key audiences. TV audiences will gain access to fact-based reports to verify climate-related claims made on air. Meanwhile, regulators, TV networks, and policymakers will receive data-driven analyses to assess misinformation trends and guide media standards.
Running from January to September 2025, the project will focus on French TV, with test phases in Germany and Brazil to explore broader applications. Real-time monitoring will begin early in the year, with a comprehensive report on climate misinformation trends set for release in September 2025.
About Science Feedback
Science Feedback is a nonpartisan, independent organization that works to improve the credibility of science-related information online. With a network of scientists and experts, Science Feedback conducts thorough verifications and analysis to inform the public and counter misinformation. The organization is based in France and operates globally, offering resources in multiple languages to support informed decision-making.
This project is funded by JournalismAI Innovation Challenge, supported by the Google News Initiative. Science Feedback, QuotaClimat, and Data for Good are developing AI-powered tools to detect and analyze climate misinformation. This project empowers fact-checkers, media, and regulators with actionable insights to expose and combat misleading climate narratives.
