The Rise of AI-Assisted Advocacy and Its Consequences
In a startling revelation that exposes the vulnerabilities in modern policy-making processes, an Australian environmental organization has admitted to using artificial intelligence to generate over 100 submissions to government inquiries—many containing fabricated evidence, nonexistent research papers, and references to government agencies that haven’t existed for more than a decade. The case of Rainforest Reserves Australia (RRA) represents a watershed moment in understanding how AI tools can be misused to influence public policy debates, particularly around contentious issues like renewable energy development.
The organization, which has gained prominence in conservative circles for its opposition to renewable energy projects, made submissions to multiple state and federal inquiries containing citations to scientific papers that the supposed publishers confirm do not exist. This development comes amid broader industry developments in AI infrastructure that are making such tools increasingly accessible.
Fabricated Authorities and Nonexistent Research
Guardian Australia’s investigation uncovered multiple instances of fabricated content in RRA’s submissions. In documents opposing a Queensland windfarm development, the organization referenced the “Queensland Environmental Protection Agency”—an entity that was abolished in 2009. The submissions also cited nonexistent bodies like the “Australian Regional Planning Commission” and “Queensland Planning Authority,” raising serious questions about the verification processes used in preparing these influential documents.
Perhaps most concerning were the references to scientific research that simply doesn’t exist. RRA submissions cited two papers from the Journal of Cleaner Production as evidence that renewable energy infrastructure releases “forever chemicals” into the environment. A spokesperson for Elsevier, the journal’s publisher, confirmed: “These references appear to be hallucinated and do not exist—we have not found any articles with those titles published in Elsevier journals.” This pattern reflects growing concerns about how recent technology implementations sometimes outpace proper oversight mechanisms.
Misrepresentation of Legitimate Research
The investigation revealed that RRA’s submissions frequently misrepresented the work of established academics to support their arguments. The organization cited Harvard science historian Professor Naomi Oreskes’ 2010 book “Merchants of Doubt” to claim that net zero emissions policies relied on “incomplete science.” Oreskes told Guardian Australia: “Merchants of Doubt does not support that claim… the passage cites my work in a way that is 100% misleading.”
Similarly, submissions referenced Professor Bob Brulle of Brown University, an expert on climate change opposition networks, claiming his work supported arguments that renewable energy advocacy lacked context. Brulle responded: “The citations are totally misleading. I have never written on these topics in any of my papers. To say that these citations support [RRA’s] argument is absurd.” These incidents highlight the challenges facing market trends in technology verification and validation.
The AI Admission and Its Implications
Anne S Smith, identified as RRA’s submission writer, acknowledged using AI to help prepare more than 100 submissions to various government bodies since August 2024. She also admitted using AI to generate responses to Guardian Australia’s questions, producing a 1,500-word document that she later confirmed was AI-assisted. Smith defended the practice, stating she used “a range of analytical tools including AI-assisted literature searches, data synthesis, and document preparation” that were “entirely under my direction.”
Dr. Aaron Snoswell, a senior research fellow in AI accountability at Queensland University of Technology’s GenAI Lab, analyzed samples of the submissions using AI detection platforms. “Looking at some of these documents, there were large portions of text that the platforms were very confident were AI generated,” he said, noting that inconsistent references “is a classic mistake that’s made by AI systems.” This case demonstrates why related innovations in AI detection and verification are becoming increasingly important.
Corrupting the Evidence Base
Cam Walker, campaigns coordinator at Friends of the Earth Australia, reviewed the RRA submissions and expressed grave concerns about their impact on democratic processes. “We’ve found multiple submissions across different renewable energy projects, all authored by the same person from RRA… all showing the same pattern of fake citations,” he said. “When you cite a government department that was abolished 16 years ago, or reference reports that don’t exist, that’s not community representation. It’s a misrepresentation.”
Walker emphasized that while legitimate concerns about renewable energy planning exist, submissions containing fabricated evidence “poison the well for legitimate environmental concerns.” The situation described in this comprehensive coverage illustrates how technological tools can be weaponized in policy debates.
Broader Implications for Policy and Technology
This case emerges as governments worldwide grapple with establishing guardrails for AI use in various sectors. The incident demonstrates how AI tools can amplify misinformation in policy debates when used without proper oversight. Dr. Snoswell noted that while AI use itself isn’t problematic, “AI-generated work needed to be double checked”—a step that appears to have been overlooked in RRA’s submission process.
As AI systems become more sophisticated, the line between human-generated and AI-assisted content continues to blur, creating new challenges for verifying the authenticity and accuracy of information presented in official proceedings. The RRA case serves as a cautionary tale about the importance of maintaining human oversight and rigorous fact-checking when utilizing AI tools in advocacy and policy development.
The organization’s influence extends beyond submission writing—RRA has coordinated open letters signed by notable Australians including energy entrepreneur Trevor St Baker, Dick Smith, Indigenous advocate Warren Mundine, and several nuclear energy advocates. While there’s no suggestion these materials were AI-generated, the case raises broader questions about how AI might be influencing environmental policy debates behind the scenes.
As AI tools become increasingly integrated into research and advocacy work, this case highlights the urgent need for clear guidelines, verification protocols, and accountability measures to ensure the integrity of evidence presented in policy-making processes.
This article aggregates information from publicly available sources. All trademarks and copyrights belong to their respective owners.
